Drivers for Serverworks chipset
I have on order a Dell server, which is based on the ServerWorks Grand Champion SL chipset. I was wondering if anyone knows what drivers should be used with this chipset, particularly the on-board 10/100/1000 ethernet? I expect there is Linux support, because Dell offers RedHat installed (I didn't order that, figuring I would just load up Woody once the machien arrives). Thanks. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: kernel 2.6.0-test4 and ASUS P4P800
Don't forget that for hyperthreading support to appear like 2 cpus, you also need to enable ACPI Processor Enumeration for HT (ACPI_HT) and enable HT in the bios. Then again, I have yet to get a bootable 2.6.0-testx on my P4P800. If you have a working config file, please forward if you would! (I suspect my use of gcc 3.3 is part of the problem, so I have to figure out how to make gcc use my old 2.95). nl On Fri, 2003-10-17 at 00:40, Jim McCloskey wrote: > > |> In all of the 2.6.0 series (not just -ac), you can enable SATA with > |> your 865/875 chipset by selecting the ATA | IDE block devices | > |> Intel PIIXn chipsets support option. > > Thank you! I didn't know this. So the crucial kernel-options, then, > would be: > > CONFIG_SMP=y > CONFIG_NR_CPUS=2 > > CONFIG_IDE=y > CONFIG_BLK_DEV_IDE=y > > CONFIG_BLK_DEV_PIIX=y > > Do I have that right? > > Thanks, > > Jim > > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
NFS has a file size overflow problem
I don't know if this is known or not known, so I am throwing it out here first to see if anyone has heard of this one before. My guess is this is a known problem... I use nfs to share my /home and a /shared directory on my fileserver to workstations. The fileserver runs testing and is up to date. The workstation was just upgraded to unstable. As a backup mechanism, on the server, I tar up my whole directory tree, split it into 600 MB files and burn to cds. I figure in the event of catastrophe, just copy the cd files back to hard disk, cat them together, and untar. So, I backup via: tar zcf /shared/backup/hometar; split -b 600m hometar then take each split file for burning to cd. Anyway, that's all beside the point except to explain how I wound up with a file that's about 5.5 GB. So, I do ls -l in the backup directory and in one xterm the file size is 5.5 GB, and in the other it's around 1.3 GB. So, what gives? Further, the split which divides up into 600MB files generates the expected number (around 9) of subfiles, so I know the actual size must be around 5.5 GB. Then, I realize that if NFS reports the file size in a 32bit entity, then subtracting the 4 GB or so you can get in 32 bits, the overflow is around the difference, around 1.3 GB or so. AND, I see the correct 5.5 GB in the xterm that is ssh'd over to the server and therefore NOT getting info via nfs, while the incorrectly small amount is in an xterm logged in to the workstation. So I'm thinking maybe nfs is stuffing a 64 bit file size on the server into a 32 bit filesize in the info it sends to the workstation. Here's the weird thing, though: I tried wc hometar in both the xterm on the server and the one one on the workstation - and they both give the same results that ls -l showed, eg running wc on the workstation gives the same incorrect 1.3 GB character count. Now, I would have guessed that wc worked by reading chars until EOF - so you'd figure it would get the correct 5.5GB answer. I'm further guessing, however, that the filesystem, after stat'ing a file, knows how large it is and doesn't let you read off the end of the file and generates the EOF without having to ask the server for data that exists but which the NFS client thinks is beyond the end of the file. So: presumably NFS is known to use 32 bit file sizes, but ls understands a 64 bit file size? I also note that the fstat, stat etc system calls return the file size as an off_t, which in the i86 header files for 2.4.22 is a long - which should be 32 bits on x86- so how does ls -l get the correct >32bit file size? nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
OT: Rio SonicBlue S10
I am trying to figure out if this device is supported under linux-usb. I've found no direct references to it (the Rio-500 seems supported by a specific driver, but that's the closest I've come). It seems that this device might be supported as a usb mass storage device - usb-storage driver. Does anyone know if that is true? If so, am I correct in assuming that the proper filesystem to mount is FAT? Thanks. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Sort of OT: network logins
The subject doesn't really explain what I am looking for, but I couldn't think of a better two-word summary. I know this is a bit off topic, but I was hoping the collective expertise would provide me with some ideas. I am looking for an approach to the problem of having multiple installations of debian on each computer on my lan that I use. While it is certainly reasonable to have a minimal install on each system, consisting of a basic debian system, it seems counterproductive to have to install each program that I use on each workstation, rather than having such software "served" by a central applications server. My present setup consists of a fileserver which exports various directories via nfs, including both a network-wide data store (called /share, for lack of a better idea), and /home. /home of course contains all my home directories (for myself and everyone else using our systems), and on each local workstation I have a full debian install with all software, and in the /home directory the actual files are symlinks to the appropriate nfs share. By way of example, the workstation mounts server:/home onto /nfs; my home directory on the workstation (/home/nl) is a symlink to /nfs/nl. This way, no matter which workstation I log into, I have my global /home/nl directory. Network-wide logins are handled by nis. Currently, I install all software onto each workstation. It would be far easier, however, to install it once onto an application server and have it available to each workstation. I've thought of two possible solutions: 1) Somehow get apt to install the software to /opt on the server, and nfs mount /opt to each workstation; 2) Install as usual to the server, and have each workstaiton mount /usr via nfs from the server; Since /etc would be local to each workstation, the same install could conceivably be used by each system with it operating differently because of different config files (X comes to mind here, since hardware may differ). Another issue is that I use KDE. On the plus side, if I edit my kde menus on one system to point to the appropriate places in /usr or /opt, then since menus are stored in my home directory, I'll have the right stuff whereever I log in. A problem, however, is that (as far as I can tell) KDE does not understand multiple simultaneous logins, and therefore I risk file corruption (or worse?) if I log in twice to my account at the same time. I had thought of solving this latter problem by implementing a login script to copy /home//.kde from the server to local storage, and then a logout script to sync it back onto the server at logout. Theoretically, I would need to do this for any porgrams that cannot sucessfully sync shared storage (like evolution), however - so this isn't really a good overall solution. I am also unsure how to make kde run a script at session start and end (or if there is even a way to make this happen under KDE). Any advice, pointers to references, etc, thoughts greatly appreciated. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Photo archiving software?
Wondering if anyone has any suggestions for photo indexing and archiving software - something like (what I have been told) iPhoto does on the Mac. I have boxes full of old photos, and would like to scan them all in for storage; I need something that will handle that sort of task, including indexing, retrieval, viewing, etc. Ideally it would allow archving to CD as well, but if not I can figure out another solution to that problem. Any suggestions welcome. Thanks. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: List of users from a certain group
On Sun, 2003-02-02 at 14:18, Jason Lunz wrote: > [EMAIL PROTECTED] said: > > I have a group of users in my system just for mail, the name of that > > group is "correo". > > > > My question is, how do I obtain a list of users from this group? > > $ apt-get install members > $ members correo > > Jason > Why not just use $ grep correo /etc/group | cut -f 1 -d ":" | xargs echo -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
nfs is driving me crazy...
I could use some help from someone knowledgeable re: nfs. I have been using nfs for some time now, probably at least two years. My fileserver contains all user's files in /home, and /home is exported by nfs. This directory is mounted to /nfs on my workstation, and serves as my home directory on the workstation because /home/nl is a symlink to /nfs/nl. Similarly, I have a directory on the server called /shared, and this is also exported by nfs and mounted to /shared on the workstation. I was in the process of upgrading the workstation, which I did by putting in a new HD, partitioning, and copying (via tar) all data from the corresponding partitions on the old (smaller) hd. I then removed the old hd, swapped the new one in as primary, and rebooted. The problem was that I was unable to connect up to the nfs shares. Despite a great deal of hacking around with this tonight, including apt-get --purge remove'ing nfs-kernel-server nfs-common and reinstalling on the server (just in case), I couldn't get mounting the nfs shares from the workstation to happen. Finally, I rebooted both machines, and the /home export mounts, but the /shared one will not. I've been playing around with it for a while and I am just plain stuck. For reference, both systems run testing. However, it's been a long time since I did an apt-get dist-upgrade on the server (for obvious reasons), while I do the dist-upgrade regularly on the workstation. I cannot see how this would matter, though, and in any case apt-get -u dist-upgrade on the server shows all the packages that would be upgraded and added and there doesn't seem to be anything relevant to nfs in there. WHen I try (on workstation): mount -t nfs -o rw,hard,intr,rsize=8192,wsize=8192 nfs:/shared /shared (NOTE: nfs is assigned to the server) I get: "mount: RPC: Unable to receive; errno = Connection refused" Now, here's the output of showmount -a nfs issued from the workstation: rpc mount dump: RPC: Unable to receive; errno = Connection refused" and from rpcinfo -p nfs issued on the server: program vers proto port 102 tcp111 portmapper 102 udp111 portmapper 191 udp804 yppasswdd 142 udp802 ypserv 141 udp802 ypserv 142 tcp806 ypserv 141 tcp806 ypserv 6001000691 udp807 6001000691 tcp809 172 udp814 ypbind 171 udp814 ypbind 172 tcp817 ypbind 171 tcp817 ypbind 132 udp 2049 nfs 1000211 udp 32770 nlockmgr 1000213 udp 32770 nlockmgr 1000214 udp 32770 nlockmgr 151 udp868 mountd 151 tcp871 mountd 152 udp868 mountd 152 tcp871 mountd 1000241 udp891 status 1000241 tcp894 status So at least the portmapper seems to be working. Because rpcinfo seems to work, I assume the port-mapper is working ok, and based on these error messages it seems that rpc.mountd on the server is refusing connections, but I don't know how to go further or figure this out any further. Any help would be greatly greatly appreciated, as I am STUCK. Oh, yeah, I should also mention that both hosts.allow and hosts.deny on the server contain onlycomments, so if I understand tcpwrappers correctly (which I understand remote mounting to use) emtpy hosts.allow and hosts.deny should result in everything being accepted, so that shouldn't be the problem. Thanks to any help or suggestions on what to look at further. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
A little more on my nfs problem...
Also, some more stuff I've been noticing: 1) My understanding was that /var/lib/nfs/xtab was the file actually consulted by mountd to respond to an nfs mount request. Interestingly, that file only contains the info on /home, but not on the /shared export, even though /etc/exports includes both. The same directory has a file called etab, and that file does contain both exports. Issuing exportfs -a -v prints out two lines, indicating both exports are being processed, but xtab doesn't change. I don't understand this. exportfs -r gives the same behavoir. I tried exporting another directory (/tmp) by adding it to /etc/exports and reissuing exportfs -a -> the new export got added to etab but not xtab, and in any case I could not mount the new share either - same connection refused error from the workstation. 2) I thought maybe the problem was that the /shared mount point is Reiserfs, thinking (without good reason) that maybe there is a problem with nfs exporting reiser filesystems, but the /tmp export I experimented with is ext2 (as is /home, the export that seems to work, at least for now). I suspect there is something wrong with mountd, and that it died after letting me mount the one export that works, but I cannot figure out what it is. There are no error messages in syslog to help me. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Last nfs related post tonight...
OK, one final bit of info. I tried showmount -a on the server and it too failed witha connection refused error. ps aux didn't show mountd running! So, /etc/init.d/nfs-kernel-server restart, and now mountd is running. Immediately thereafter I could mount the shared partition. Based on this, it seems that the problem is that mountd is exiting unexpectedly...and I have no idea why this would occur, I cannot find any error messages in the logs to explain this. Any thoughts> nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Do all debian kernel packages require an initrd?
Tonight, I tried to apt-get install kernel-image-2.4.21-4-k7 (AMD Athlon system). During the installation phase, I was given a warning that this image requires an initial ramdisk - which I am not using, as far as I am aware, with the stock 2.4.18-bf2.4 kernel that originally came with woody (now sarge). Are all debian kernel packages created with initrd? If so, is justing proceeding with the install sufficient, or is there some special way to modify my menu.lst for grub to handle this (the install gives instructions for lilo, but I don't use lilo). Thanks. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Last nfs related post tonight...
On Mon, 2003-09-08 at 02:55, Erik Rask wrote: > On 7 Sep 2003, Neal Lippman wrote: > > > OK, one final bit of info. I tried showmount -a on the server and it too > > failed witha connection refused error. > > > > ps aux didn't show mountd running! So, /etc/init.d/nfs-kernel-server > > restart, and now mountd is running. > > > > Immediately thereafter I could mount the shared partition. > > > > Based on this, it seems that the problem is that mountd is exiting > > unexpectedly...and I have no idea why this would occur, I cannot find > > any error messages in the logs to explain this. > > > > Any thoughts> > > > > I have the same experience as Marc Wilson described earlier, although the > nfs-kernel-server/testing package doesn't work regardless of client > version. There are bug reports to this effect. My solution was to > downgrade to stable and hold it as long as testing is messed. > To see if your problem is the same as mine, try restarting NFS on the > server, mount one share and then one more. If the first succeeds and any > subsequent mounts fail, that's what I had. > Back to work. > This is the same sort of behavoir that I sawit's starting to sound like mountd is broken in testing, although I don't recall any recent upgrades to it and I've been running this setup for some time now without this kind of problem. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Boot time module configuration - stored in initrd.img?
I noted strange behavoir after installing linux-kernel-2.4.21-4.On reboot, a host of modules that I had never selected with modconf were installed. They were not in /etc/modules, however, so the kernel upgrade had clearly not modified this file (although they did appear selected in modconf - perhaps modconf determines what is installed by looking at /proc/modules rather than /etc/modules?). Anyway, "strings /boot/initrd.img-2.4.21" did show all of these modules listing, making me think that these modules are hardcoded into the initrd image. If this is the case, then of course removing them with modconf will have no effect on reboot time - so if I don't want all these modules, there is no alternative but to compile my own kernel - or add a boot time script to remove the unwanted modules, which is far less elegant. Is this an accurate understanding of what is going on here? Thanks. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
kernel-source-2.6.0-test-2
Has anyone been able to successfully use this package? I installed it but make xconfig just dies with a million errors, all coming out of scripts/kconfig/qconf.o. For reference, I have installed libqt-dev, so I thought I would be good to go Debian version is testing, btw. Thanks. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: .config file for kernel-2.6.0-test4
>From what I have read (see my later post re inability to get 2.6.0 working, btw), there are enough incompatibilities between 2.4.x and 2.6.0-test-n that just copying the .config will not work. You could use that as a starting point, but you shoudl go through the config line by line and modify as needed. Personally, I prefer to build a .config from scratch in this instance, since enough seems to be new that i like to read the help for each item and decide. nl On Tue, 2003-09-16 at 12:00, Brian P. Flaherty wrote: > mess-mate <[EMAIL PROTECTED]> writes: > > > > So, can I use my .config file from the 2.4.20 kernel ?? > > Does 'make oldconfig' work? It should read through an old config, > taking your past answers and ask about new options. Note that between > 2.4.20 and 2.6.0-test4, there will be a lot of questions. > > Brian > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Not impressed with 2.6.0-test2 so far...
Well, so far I'm less than impressed with 2.6.0-test2. Specifically, although I have installed both the source and run several compiles, as well as the pre-compiled kernel-image version, I have yet to get a running and working 2.6.0. The precompiled package in testing doesn't work at all; after numerous error messages during the boot (which go by so fast I cannot see them all), followed by a bunch of failures trying to load lvm_mod (don't know why), the system panics after trying to kill the init process. Not impressive. My own compiled version gets farther, but after a mess of dma errors on /dev/hda6 (don't get that running 2.4.22), staring KDE fails and I am dumped back to a login prompt in the console. From what I can see, there are a lot of usb errors. Again, don't know why, becaue the keyboard doesn't work so I cannot log in and check dmesg. I'm wondering if anyone has gotten 2.6.0-test2 compiled and runnign at all and has any pointers? I've read what I can online, fixed the pitfalls noted (for instance, since CONFIG_VT, CONFIG_VGA_CONSOLE, etc must be compiled in and not modules, you'd think the config scripts would tell you that) but not yet making progress. Oh well, another config and another try tonight... nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Not impressed with 2.6.0-test2 so far...
On Tue, 2003-09-16 at 22:11, J. Bruce Fields wrote: > On Tue, Sep 16, 2003 at 09:07:38PM -0400, Neal Lippman wrote: > > I'm wondering if anyone has gotten 2.6.0-test2 compiled and runnign at > > all > > Of course; tons of people are running it. It's what I use for most of > my daily work. My laptop, alas, is a little unstable under 2.6, but two > other computers at work are running fine. > > > and has any pointers? > > It probably won't be possible to do any troubleshooting without more > detail about your particular situation: exact error messages, hardware > involved, etc. Unfortunately, it's a bit hard to provide full details because, as I outlined above, I cannot even get a system working well enough to log in and review dmesg. For the record: Hardware is: AMD Athlon XP 1700+ on MSI K7T-266Pro motherboard, 512 MB Mushkin PC2100 DRAM. 1 WD 80 GB HD, IDE, as hda 1 Pioneer 106S DVD reader as hdc 1 HP CDRW as hdd (ide-scsi under 2.4.x) Linksys 10/100 ethernet card (tulip driver) ATI Radeon 7500 video, AGP Creative SB Value sound card (PCI) Using 2.6.0-test2, I find that with the kernel_image in sarge, I get a series of "dma errors" on hdc, followed by errors loading module lvm_mod, followed by a panic after attempting to kill /sbin/init. With my own complied 2.6.0-test2, I am finally able to boot, but kdm cannot start. At the login prompt, I cannot log in as the keyboard does not work. I will review the .config and maybe I made the keyboard a module and it is not getting loaded. I d/l'd 2.6.0-test5 from www.kernel.org tonight. That doesn't even compile, with error: drivers/built-in.o(.text+0x1d553): In function `register_serial_portandirq`:: undefined reference to `register_serial` drivers/built-in.o(.text+0x1d6ad): In function `wmave_exit':: undefined reference to `unregister_serial' I haven't even begun to look at the source yet to see if I can figure that one out. I'd be happy to post my .config files if anyone has the time to look through them. Oh yes, I am running sarge. gcc version is 3.3.1 20030626 Debian prerelease. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Question on make-kpkg
I've been using make-kpkg to build my new kernel,playing around with 2.6.0-test2. (aside: Thanks to all who responded to my earlier request for help.I've got a bootable 2.6.0-test built, but now I need to play with the config a bit and figure out what's causing some of the errors during boot; at least it hasn't crashed in about 24 hrs so far, though). Anyway, as I play with the config parameters, I'd like to be able to create kernel packages with different names to help keep track of what's what. This is helpful by labelling packages with the build date, and also with some info, esp as I will eventually build a package for a different architecture as well (I'm going to swap my MSI motherboard/Athlon process for a new ASUS MB and P4, and I'd like to prebuild the kernel and have it ready to boot). Anyway, I started out by using: fakeroot make-kpkg append-to-version=.20030915 kernel_image because I first did the build on 9/15. Now, if I try to issue fakeroot make-kpkg append-to-version=.20030917 kernel_image, I get an error message telling me that the changelog wants to build vmlinuz-2.6.0-test2.20030915 "but I thought I was building vmlinuz-2.6.0-test2.20030917" I was expecting that I could keep using make-kpkg in a given linux source tree and change the append-to-version on each build, but obviously I don't understand something important about make-kpkg. The man page isn't really helping me out here. Can anyone shed some light? THanks. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
modconf under 2.6.0-test?
Thanks again to those who have offered great advice on getting 2.6.0-test2 up and running. I have a bootable system, got sound working with relatively little trouble,so I'm continuing to work down myproblem list. Would appreciate thoughts on two more problems: 1) modconf doesn't find any modules. It just comes up with an empty menu,with the only choice being "exit." I understood that modconf used the output of uname to find the appropriate tree in /lib/modules, and since that tree is where it should be, I am stumped. I do have the new module-init-toosl installed, and I can install modules just fine. 2) Shutdown: as far as I can tell, the USB layer is trouble. In booting up, I get a ton of usb related errors - no specifics yet, I'm trying to narrow down where these are coming from first. However, on shutdown, while dconfiguring the usb devices, the shutdown hangs and never completes, so I just have to hard-reset the machine. I ran into this a long time ago on an early 2.4 series kernel on an ASUS CUSL-2/Pentium III system, and never figured out the problem. Wondered if anyone else is seeing somehting similar. For reference, the only USB device hooked up is an external Belkin hub and a generic USB memory card reader (that handles CF, SM, etc). The reader always worked fine under 2.4.x with the usb-storage module. I actually haven't testing it yet to see if it reads cards under 2.6.0. Thanks. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
modconf under 2.6.0-testn
I have been unable to get modconf to work under 2.6.0-test2 or -test4. The program runs, but shows only the "exit" option, with no modules displayed for selecting. Has anyone else seen this sort of problem,and if, any suggestions? nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
is the tty serial layer working in 2.6.0-testx?
I have run into another problem with 2.6.0-test2 and -test4. I use a Pilot M505 which I have always synced using the serial port (I never got around to getting it connected via USB). This has worked fine under the 2.4.x series, but does not work now that I have started using 2.6.0-test2 and -test4. I think the problem lies with the serial tty layer and not the pilot-link interface, because "setserial -g /dev/ttyS0" returns "/dev/ttyS0: No such device." The output of "ls -l /dev/ttyS0" is: crw-rw 1 root pilot 4, 64/dev/ttyS0 which looks just fine and reviewing devices.txt in the kernel docs indicate that the major/minor for this device hasn't changed. I am pretty sure the drive is installed; here's the output of "cat /proc/devices": Character devices: 1 mem 2 pty 3 ttyp 4 /dev/vc/0 4 tty 5 /dev/tty 5 /dev/console 5 /dev/ptmx 6 lp 7 vcs 10 misc 13 input 14 sound 21 sg 29 fb 116 alsa 128 ptm 136 pts 171 ieee1394 180 usb 202 cpu/msr 226 drm Block devices: 2 fd 3 ide0 22 ide1 so it sure looks like the devices for tty are there. I'm not sure how else to troubleshoot - does anyone have any ideas? nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Very confusing network problem!
I have encountered a networking problem that has me just stumped, so I wonder if anyone could lend some thoughts. My post is long to try to give enough detail to help figure out what is going on; I appreciate anyone who can read this and lend some thoughts. Here's the background: The system is debian testing. Previously, it used an MSI K7T-266 MB / AMD Athlon XP processor, with a Linksys 10/100 NIC. THe kernel was a self-compiled 2.4.21 from the debian kernel-sources, and I was also playing with a 2.6.0-test4 for about a week now. Everything has worked fine. However, in a fit up upgrade-frenzy, last night I swapped out that MB/processor and installed a new ASUS P4P800 (intel 865 chipset, onboad 3com 3c940 10/100/1000MB NIC) with a P4-2.4GHz. Strangely (or not so strangely, I don't know yet) the 2.6.0-test4 kernel that I had precompiled (before swapping the MB) ready to go for a P4 processor doesn't boot with so many error messages, going by so fast, that I cannot even start to work on that. THe stock installed debian 2.4.18-bf24 also doesn't boot (which I find strange), getting through with the loading the kernel line and then just rebooting. Odd. Anyway, my precompiled 2.4.21 kernel for P4 does boot, both with HT enabled and disabled (the kernel is compiled for SMP support) and I am up and running KDE. I downloaded the driver for the onboard NIC from ASUS's website (yep, they had a Linux driver!) and compiled and installed it. THe module loads just fine, and I can bring up the interface. Now, here's the problem: I can ping basically anything - my Linksys cable modem, my in-house file server, the (outside) DNS servers for my ISP, whatever - as long as I enter the IP address. DNS resolution does not work (and yes, resolv.conf, etc are properly configured - it's the same setup that worked fine before I swapped the motherboard). Further, I cannot make any sort of UDP or TCP connection that works. For instance, I usually connect to my file server via SSH, but now, when I SSH to the server (using its IP address) the connection just hangs. THe Linksys cable modem is configurable via a http interface, but if I try to get the index page: wget -v http://192.168.1.1/index.html the -v flag shows me that a connection IS made to the host on port 80, then wget reports that it is downloading the requested web page, then: nothing, it just hangs forever. Since it's new hardware and a new driver, the temptation is to assume that either a) the driver is borked (Jim McCloskey has told me of a patch to 2.4.22 he's using from the 3com website that I can try), or the hardware is borked (maybe a bad motherboard?). However, since I CAN ping any host internal or external to my network, that leads me to believe that the hardware must be working. Doesn't that seem right? If so, why am I unable to get TCP/UDP connections to fly? (Yes, I checked my .config and I did include TCP/IP networking, and anyway I shouldn't be able to ping if I didn't). Any advice/thoughts would be greatly appreciated, because I am totally stumped on this one. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
NIS updating question
I use nis on my LAN for storing passwd, group, hosts, services, etc, information. I need to update the hosts file with new ip addresses. Is there an command or easy way to propagate updated host information to the NIS database, or do I just use ypinit and rebuild the whoe database? Thanks. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
kernel 2.6.0-test4 and ASUS P4P800
I'm having a heck of a time getting 2.6.0-test4 to boot on my ASUS P$P800-based system. I didn't have any real trouble getting a config working on my Athlon based system, but I cannot figure out where i am going wrong... If anyone has this MB and has successfully compiled a 2.6.0-test kernel, I would GREATLY appreciate it if you would be able to send me a copy of your .confg file for comparison with mine. Thanks. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Web-based e-mail system?
You may wish to check out "horde". My ISP uses this for webmail, and it is very nice. I understand it to be open source. I do not know the status of it in debian, or what backend mail stores it can access, soyou will need to research it. nl On Tue, 2003-10-07 at 09:49, Aaron wrote: > Hey, Debian users! > > I currently use a fetchmail / procmail / mbox / mutt e-mail setup, > with ssmtp (properly linked through `sendmail` of course) for sending. > I would really like to have a web mail system set up so that I can at > least read, if not send, e-mail from my website as well. > > Does anyone know of a package that can put mbox mail on the web? It > sounds kind of silly, given the inefficiency of mbox, so I'm not > holding high hopes, but if anyone has info. about it, that'd be great. > > Thanks, > > -- > Aaron Bieber > - > Graphic Design // Web Design > http://www.fisheyemultimedia.com/ > [EMAIL PROTECTED] > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Is anyone else having trouble isntalling/upgrading mplayer?
I'm just wondering if anyone here is experiencing apt-get problems with mplayer. I'm running testing. I was using mplayer and the mplayer-plugin for mozilla just fine on my prior (athlon) system - I had mplayer-k7 installed. I upgraded that system to a new mb and processor, so now it's a P4 system, and the installed mplayer was balking, so I removed the existing mplayer. Now, when I apt-get install mplayer-686, I get the following: output from apt-get follows Reading Package Lists... Building Dependency Tree... Some packages could not be installed. This may mean that you have requested an impossible situation or if you are using the unstable distribution that some required packages have not yet been created or been moved out of Incoming. Since you only requested a single operation it is extremely likely that the package is simply not installable and a bug report against that package should be filed. The following information may help to resolve the situation: The following packages have unmet dependencies: kde: Depends: kdebase-audiolibs but it is not going to be installed or kdebase3-audiolibs but it is not installable mplayer-686: Depends: libavcodec1 (>= 1:0.4.8) but it is not going to be installed Depends: libvorbis0a (>= 1.0.0) but it is not going to be installed *end output The only thing is, here's the output from: apt-show-versions | grep "audiolibs\|avcodec\|vorbis" **output follows: libvorbisenc2/testing uptodate 1.0.0-3 libvorbis0 1.0.0-1 installed: No available version in archive kdebase-audiolibs/testing uptodate 4:2.2.2-14 libvorbisfile3/testing uptodate 1.0.0-3 libavcodec0 1:0.4.6-sarge0.1 installed: No available version in archive vorbis-tools/testing upgradeable from 1.0rc3-1 to 1.0.0-2 end output so it looks to be like audiolibs IS installed, while strangely it thinks that libavcodec0 is installed while having "no version in archive". I should also note that "apt-get check" does not report any dependency problems. The whole problem seems to be the kdebase-audiolibs, because apt-get install libavcodec1 fails because libvorbis0a is not isntalled, AND apt-get install libvorbis0a fails because kde has an unmet dependence of kdebase-audiolibsand I cannot apt-get remove kdebase-audiolibs becuase it gives the same unmet dependency for kde error as trying the install, while apt-get install kdebase-audiolibs tells me it is already installed. I tried apt-get --reinstall for both kdebase-audiolibs and kde, but that didn't fix the problem. I'm stuck. Any suggestions would be mighty appreciated. Thanks. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Weird dependency problems installing mplayer...
I posted this once before, but figured it was worth another shot to see if anyone could help, as I am still stuck... I'm running testing. I was using mplayer and the mplayer-plugin for mozilla just fine on my prior (athlon) system - I had mplayer-k7 installed. I upgraded that system to a new mb and processor, so now it's a P4 system, and the installed mplayer was balking, so I removed the existing mplayer. Now, when I apt-get install mplayer-686, I get the following: output from apt-get follows Reading Package Lists... Building Dependency Tree... Some packages could not be installed. This may mean that you have requested an impossible situation or if you are using the unstable distribution that some required packages have not yet been created or been moved out of Incoming. Since you only requested a single operation it is extremely likely that the package is simply not installable and a bug report against that package should be filed. The following information may help to resolve the situation: The following packages have unmet dependencies: kde: Depends: kdebase-audiolibs but it is not going to be installed or kdebase3-audiolibs but it is not installable mplayer-686: Depends: libavcodec1 (>= 1:0.4.8) but it is not going to be installed Depends: libvorbis0a (>= 1.0.0) but it is not going to be installed *end output The only thing is, here's the output from: apt-show-versions | grep "audiolibs\|avcodec\|vorbis" **output follows: libvorbisenc2/testing uptodate 1.0.0-3 libvorbis0 1.0.0-1 installed: No available version in archive kdebase-audiolibs/testing uptodate 4:2.2.2-14 libvorbisfile3/testing uptodate 1.0.0-3 libavcodec0 1:0.4.6-sarge0.1 installed: No available version in archive vorbis-tools/testing upgradeable from 1.0rc3-1 to 1.0.0-2 end output so it looks to be like audiolibs IS installed, while strangely it thinks that libavcodec0 is installed while having "no version in archive". I should also note that "apt-get check" does not report any dependency problems. The whole problem seems to be the kdebase-audiolibs, because apt-get install libavcodec1 fails because libvorbis0a is not isntalled, AND apt-get install libvorbis0a fails because kde has an unmet dependence of kdebase-audiolibsand I cannot apt-get remove kdebase-audiolibs becuase it gives the same unmet dependency for kde error as trying the install, while apt-get install kdebase-audiolibs tells me it is already installed. I tried apt-get --reinstall for both kdebase-audiolibs and kde, but that didn't fix the problem. I'm stuck. Any suggestions would be mighty appreciated. Thanks. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Totally confused about what apt-get is doing...
OK, I've posted here before with my problem installing mplayer, and I've finally tracked things down to a missing libvorbis0a package. Here's what I cannot (now) figure out. 1) apt-show-versions | grep "vorbis" - shows me that I have no installed packages with vorbis in it, so I don't have libvorbis0a installed. 2) apt-get install libvorbis0a tells me: Reading Package Lists... Building Dependency Tree... Some packages could not be installed. This may mean that you have requested an impossible situation or if you are using the unstable distribution that some required packages have not yet been created or been moved out of Incoming. Since you only requested a single operation it is extremely likely that the package is simply not installable and a bug report against that package should be filed. The following information may help to resolve the situation: The following packages have unmet dependencies: kde: Depends: kdebase-audiolibs but it is not going to be installed or kdebase3-audiolibs but it is not installable E: Broken packages so clearly I cannot install libvorbis0a. But why? 3) apt-get -d install libvorbis0a says the same thing, so I cannot seem to even just download the deb. 4) I know the deb is there, because on packages.debian.org, I can find the libvorbis0a package under testing (yes, I have testing properly configured in /etc/apt/sources.list), and I can manually retrieve the deb from the same server I have listed in my sources.list file. 5) dpkg --info on the package thusly retrieved seems to show the correct information in it. So, does this mean something is broken with apt, the packages lists (I did do an apt-get update first), this packages, or my system? (I should mention that apt-get check does not show any errors). Thanks for any help! nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Totally confused about what apt-get is doing...
On Sun, 2003-10-12 at 15:07, Naitik Shah wrote: > I dont know if you've already dont this, but could you post > your sources.list? > Yes, here it is: deb http://http.us.debian.org/debian testing main contrib non-free deb http://non-us.debian.org/debian-non-US testing/non-US main contrib non-free deb http://security.debian.org/ testing/updates main contrib non-free #deb http://mplayer.nmeos.net testing/ deb http://marillat.free.fr/ testing main -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Totally confused about what apt-get is doing...
On Sun, 2003-10-12 at 15:08, Shri Shrikumar wrote: > On Sun, 2003-10-12 at 19:30, Neal Lippman wrote: > > 2) apt-get install libvorbis0a tells me: > > > > Reading Package Lists... > > Building Dependency Tree... > > Some packages could not be installed. This may mean that you have > > requested an impossible situation or if you are using the unstable > > distribution that some required packages have not yet been created > > or been moved out of Incoming. > > > > Since you only requested a single operation it is extremely likely that > > the package is simply not installable and a bug report against > > that package should be filed. > > The following information may help to resolve the situation: > > > > The following packages have unmet dependencies: > > kde: Depends: kdebase-audiolibs but it is not going to be installed or > > kdebase3-audiolibs but it is not installable > > E: Broken packages > > > > so clearly I cannot install libvorbis0a. But why? > > Do you have testing / unstable system running with KDE 3 from unstable? > If so, try running > > $apt-get -t unstable install libvorbis0a. > No, I am running testing with version of KDE standard in testing (2.2.2, I think). -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Creating a new machine with the same set of packages as an existing one?
On Sun, 2003-10-12 at 18:45, stan wrote: > How cna I do this? > > The machine I want to match is a "testting" machine, that I quit updating > about a month agao. Still has Gnome 1.4 for instance. > > Will this ne a problem? > One approach is to create a list of all of the packages currently installed on your system, for instance with dpkg --get-selections or with apt-show-packages. You can parse the output and save into a file to get the names of all of the installed packages, for instance: dpkg --get-selections | cut -f 1 >package-listing-from-dpkg or apt-show-versions | cut -f 1 -d / >package-listing-from-apt Now, install a base woody system on the new machine. Modify your /etc/apt/sources.list for testing, and do apt-get update; apt-get dist-upgrade to bring you up to testing since that is what your current machine is. Now, using either of the above package lists, you can do: for pkg in $(cat package-listing-from-xxx); do apt-get install $pkg; done which will work although not be the most efficient way (because you will have to run apt-get for each package, and hence dependences are re-resolved over and over, or: cat package-listing-from-xxx | xargs apt-get install will also work and let apt-get do all the resolving at one time, PROVIDED the total listing of your packages does not exceed the maximum line length of bash. Since it is likely that it WILL do so, you might want to stick to the first approach of installing each package one by one. Slower, but of course later packages that have already been installed won't need to happen again. I haven't testing this in practice, and note that there isn't any error handling here, but you can always give it a shot. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Creating a new machine with the same set of packages as an existing one?
On Sun, 2003-10-12 at 21:14, Roberto Sanchez wrote: > Neal Lippman wrote: > > > Now, using either of the above package lists, you can do: > > for pkg in $(cat package-listing-from-xxx); > > do > > apt-get install $pkg; > > done > > which will work although not be the most efficient way (because you will > > have to run apt-get for each package, and hence dependences are > > re-resolved over and over, or: > > Better way: > > (after adjusting sources.list) > > apt-get update > dpkg --set-selections apt-get upgrade > > Yes, of course. Dumb of me to use --get-selections without also using --set-selections. Doh. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Strange permissions behavoir on NFS mount...
Wondering if anyone can explain sort of odd behavoir I noted tonight on an nfs share. The setup: I have mounted an exported nfs share from my server onto my /home directory on my workstation (mount -t nfs nfs:/home /home). On the server machine (nfs), /home is exported with root_squash, so I would expect that as root on the workstation, I would NOT have access to ANY of the contents of this share. Here's what is actually happening, though. Assume there are two users, userA and userB, both with directories in /home on the nfs server: permissions directory drwx-- userA drwx-- userB When you look at the /home from the workstation, you see exactly the same directories and permissions, as expected. Now, suppose I log in as userA, and attempt to look in /home/userA and /home/userB. As expected, I can look into /home/userA, but NOT into /home/userB. Now, I su to root. As root, I can STILL look into /home/userA but NOT into /home/userB. Strange. Interestingly, if I look at two subdirs in /user/userA at this point, dir A and dir B, with the following permissions: permissions directory drwxr-xr-x A drwx-- B after su'ing from userA to root I can look into directory A but NOT into B. Now, I su to userB, and things reverse: I can see into /home/userB but not /home/userA, again as expected. If I again su to root starting as userB, I can still see into userB but not userA. I find this hard to sort out. It's as though, despite su'ing to root, the NFS server is still getting the original uid, not root. But, somehow, once I am into the dir (eg looking inside /home/userA), trying to access a subdir no longer has that uid (because I can get into /home/userA/A but NOT /home/userA/B after the su to root). Does anyone understand this? What uid/gid is sent to the nfs server after an su to root? Note, that I have checked the results of getuid(), geteuid(), getgid(), getegid(), getresuid() and getresgid() as userA, after su to root, and with sudo to run a program to output these values. As userA, uid and gid are all userA as expected. As root, whether after su or with sudo, uid/gid are all 0, so these values are what would be expected...and not consistent with teh above results. Any thoughts? nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
More info...? NFS bug
I am convinced something is wrong with nfs. This is a f/u to my earlier post on nfs permissiosn problems with root_squash. Having played with this a bit more, I find that as userA -> su to root, sometimes I can see userB's files and sometimes not on the nfs share...even though with root_squash and permissions being drwx-- on all directories, I should not be able to see any of these files after su to root. The nfs server runs testing and should be up to date, unless there has beena new nfs release in the last few days. Haev I found a bug somewhere in nfs? nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Linksys PCMCIA network card - what module to use?
I decided to have a go at getting debian installed on my Thinkpad 770ED; mostly I was able to get the initial module configuration to work (a few glitches, but I can work those out later). However, until I can get networking up and running, I cannot continue the install. I have a Linksys 10/100 PC card (PCMCIA), but I cannot figure out what the correct driver is for this. Anyone have any information? The card is a mode PCMPC100, for what it's worth. Thanks. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Network startup on IBM Thinkpad
, ds, iThanks to all who responded with info to my question regarding the correct drive for a Linksys PC-Card 10/100 NIC for IBM Thinkpad. As best as I have figured out from the various responses coupled with online sources, the correct driver is the axnet_cs driver. Interestingly, this driver does NOT appear under the selection choices from modconf during woody install. When I retried the install, I simply did not select any NIC driver, which caused the install to skip configuring the network entirely (bad). Once I was installed, however,I just manually created an entry in /etc/network/interfaces for eth0, restarted the network (/etc/init.d/networking start), and somehow the system found the network card entirely on its own. I do seem to have the necessary modules running: pcnet_cs, 8390, ds, i82365, pcmcia_core - all of which seem somehow related to the pcmcia subsystem, and I am guessing that pcnet_cs is the pcmcia network driver, which somehow seems to work. I don't fully understand what happened, but so far so good. Here's the problem: when I reboot, the system doesn't seem to see the nic unless, after the boot is completed, I remove the card and reinsert it. I assume the problem is that the bootup sequence is not probing the card correctly, but when the card is reinserted, cardmgr is then correclty configuring it. So, my question: how can I get the pccard to be probed at startup and properly configured? This isn't a trivial point for me, since my user accounts are actually served on my home lan via nis, but nis cannot start up because it canot bind to the ypserver because the network isn't fully up... Any help apprecaited. Thanks. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
X 4.1.0 - is there gatos code?
I was wondering if any knows of the "official" XF 4.1.0 debs in woody include any gatos code? I am planning to hook up an ati radeon 7500 dual port (DVI-I and VGA) card, and have been advised that the drivers modified with "gatos" project code do not work properly. Thanks. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
ATI Radeon 7500 / X4.1.0 - need help!
I just got an ATI Radeon 7500 card with dual outputs (DVI-I and VGA). My search of the Xfree86 site, various newgroups, etc, turns up conflicting information regarding the status of this card. The only driver referenced on the XFree site is the ATI site, which does not list any radeon cards as supported. My install (Woody) does include a radeon_drv.o module (as well as the ati_drv.o module) in the X drivers tree, but I cannot find a reference to it on the XFree site. Any info on the current support state of this card, and which driver is correct appreciated. If anyone has a sample XFConfig-4 file that supports this card, that would be appreciate as well. Finally, I would like to use the DVI-I port, and maybe dual-head a second VGA monitor off of it as well - but maybe that's asking for too much. Thanks in advance. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: XF4.1 on ATI Mobility Radeon 7500
On Tue, 2002-10-22 at 08:07, Florian Thiel wrote: > On Tue, Oct 22, 2002 at 01:35:47PM +0200, Juergen Stuber wrote: > > Oleg <[EMAIL PROTECTED]> writes: > > > > > > I've installed Debian on half a dozen different machines, and configuring X > > > has always been the greatest pain. Currently, I'm trying to do this on a > > > Fujitsu Lifebook E-7110 laptop that has an ATI Mobility Radeon 7500 card > > > (32MB) and an LCD that can do up to 1400x1050 @ 60 Hz. > > > > > > When I tried installing Mandrake 9.0 just for kicks, it configured X without > > > even asking me any questions (I should have probably saved its XF86Config-4 > > > for reference, but I didn't). Now with Debian on the same machine instead, I > > > spent all night trying to get ANY graphics to work. WTF? > > > > AFAIK you need XFree 4.2, which is in unstable. > > Yes. The Radeon drivers are only in 4.2. You could use apt's pinning > feature (worked for me; I use testing). > > Is there a difference between this and the "ati" driver that I am told will work with ATI radeon 7500 cards under X4.1.0? How does this differ from the radeon_drv.o driver I also have on my system (woody with X4.1.0)? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Strange nis / services problem - any thoughts?
I just recently converted over to using nis to deliver my /etc/passwd and /etc/group data on my home lan. All has been working fine with that for a few weeks, so I decided last night to convert over to having my workstation obtain /etc/services, /etc/protocols, /etc/ethers, and /etc/rpc data from the nis server as well. I modified /etc/nsswitch accordingling (by changing the entires for services, protocols, ether, and rpc from "files" to "nis files"). Tonight, I booted up, and issued "sudo apt-get update" as I usually do - and instead of retrieving the appropriate info, apt spit out a slew of error messages being unable to connect to any of the deb mirrors I usually use. At first, I thought the mirrors were broken, but realized that several different mirrors were unlikely to _all_ be broken the same way at the same time. Thinking my mod to nsswitch.conf could have broken something, I edited it and removed "nis" from the above databases, and retried the apt-get update, and it worked fine. Further, I found that one of my standard tricks to check network connectivity, telnetting to an smtp port on an outside server, failed - when I tried "telnet smtp.foo.com smtp", I couldn't get connected when nsswitch used nis for services, but could connect when it only used files. Because the line reads 'services nis files', I assume that nis must be returning bogus info for the service, because if it just failed to return anything, then presumable the lookup would go on to my local file which would give the right data. I then wrote a quick program to call getservbyname with a specified service parameter and print out the results, and found that I got back the same data regardless of whether nssswitch specifies nis is to be used or not. Finally, I can specify nis for ethers, rpc, and protocols without breaking anything, but if I do for services, then apt, telnet, etc are all broken. I'm not sure what's going on here to cause this to happen. Any thoughts appreciated! Thanks. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Need help with NIS
I just recently converted over to using nis to deliver my /etc/passwd and /etc/group data on my home lan. All has been working fine with that for a few weeks, so I decided last night to convert over to having my workstation obtain /etc/services, /etc/protocols, /etc/ethers, and /etc/rpc data from the nis server as well. I modified /etc/nsswitch accordingling (by changing the entires for services, protocols, ether, and rpc from "files" to "nis files"). Tonight, I booted up, and issued "sudo apt-get update" as I usually do - and instead of retrieving the appropriate info, apt spit out a slew of error messages being unable to connect to any of the deb mirrors I usually use. At first, I thought the mirrors were broken, but realized that several different mirrors were unlikely to _all_ be broken the same way at the same time. Thinking my mod to nsswitch.conf could have broken something, I edited it and removed "nis" from the above databases, and retried the apt-get update, and it worked fine. Further, I found that one of my standard tricks to check network connectivity, telnetting to an smtp port on an outside server, failed - when I tried "telnet smtp.foo.com smtp", I couldn't get connected when nsswitch used nis for services, but could connect when it only used files. Because the line reads 'services nis files', I assume that nis must be returning bogus info for the service, because if it just failed to return anything, then presumable the lookup would go on to my local file which would give the right data. I then wrote a quick program to call getservbyname with a specified service parameter and print out the results, and found that I got back the same data regardless of whether nssswitch specifies nis is to be used or not. Finally, I can specify nis for ethers, rpc, and protocols without breaking anything, but if I do for services, then apt, telnet, etc are all broken. I'm not sure what's going on here to cause this to happen. Any thoughts appreciated! Thanks. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
bash scripting question
I am trying to solve a bash scripting problem, but I cannot figure it
out.
I frequently need to execute a command of the form:
for x in {A,B,C,D,E,F,G,H,I,J,K,L,M,N,O,P,Q,R,S,T,U,V,W,X,Y,Z); do
;
done
This works fine if I actually type out the entire alphabet list on the
command line as above, but that's sort of a pain. So, I tried setting a
shell variable to the alphabet string (export alpha="A,B,C,...,Z"), but
then the command:
for x in {$alpha} ;
do
echo $x;
done
winds up printing the string "{A,B,C,...,Z}" rather than each letter on
a separate line as expected.
I've tried various versions, including escaping the {} characters, etc,
using xargs, etc, but I cannot hit upon a sequence that works.
I also tried writing a program that printed the alphabet string to
stdout, but same results.
Can anyone suggest a syntax that would do the trick here?
Thanks.
nl
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Mounting a partition with access allowed for only one user
I am wresting with a mount question but getting nowhere. Suppose that I want to mount a disk partition onto a subdirectory of /home, but want it to be available only to one particular user. Is there any way to effect this? eg, I create a directory called /home/restricted, and then want to mount a local disk partition containing an ext2fs file system: mount /dev/hda8 /home/restricted When I do the mount, the ownership and permissions of /home/restricted are now changed to root.root and rwxrwxrwx, regardless of what I have set for the ownership / permissions of /home/restricted prior to the mount. How can I make it so that the partition mounted is accessible only to a selected user? Thanks. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
What's the best way to contact someone knowledgeable on nis?
I'm trying to sort out a strange problem with nis - or at least, I think it's with nis. I haven't been able to find answers, including with an earlier post hear, so I was wondering if I could locate, perhaps, someone knowledgeable in the internal workings of nis, but I don't know how to go about that. Anyone with suggestions? The problem I have is strange. I use nis for lookups for passwd, groups, hosts, etc, so that I can keep a central database on one server in my home LAN and not have to duplicate accounts, etc. The problem is with /etc/services. When I set up nsswitch.conf to use nis for servers, some commands no longer work properly. For instance, apt-get. If I try apt-get update or apt-get upgrade, they fail, stating that connections to my various apt sources are refused. However, when I edit nsswitch.conf so that servers uses the local file instead of nis, all works fine. I've tried ltrace'ing and strace'ing apt-get to see if I can spot where the problem lies, unsuccessfully. The same thing happens if I try to telnet to a specific port (eg telnet mail.foo.bar pop3), but only when services is being delivered via nis. I'm at a loss to figure out why this is the case. A test program to use getservicebyname works identically regardless of whether services is handled by nis or not. Any thoughts? NL -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
OT: awk problem
Sorry for another OT post here, but this is the best place I've found
for linux knowledge even when not debian specific.
I am trying to write a short awk script to process my /etc/services file
into a format that the utilities supplied with yp/nis can correctly
handle. What I need to do is to parse off the service name and
portnumber/protocol, and the following comments, and rewrite these
fields out in a different format. My script does the parsing correctly,
but after processing about 991 lines, it fails saying:
'awk: program limit exceeded: maximum number of fields size=32767
FILENAME="-" FNR=991 NR=991'
I don't know why this would occur - it would seem that the field count
should reset with each line read, and the line that follows the last
processed line is not different in format that any preceeding line, so I
cannot figure out why this fails. I'm obviously not an awk expert, but
still... Note that the script is written to exclude lines with an email
address in them (finds "@") and lines with the string "--none---"
because these are comment lines which I don't need and nis will not need
anyway...but other comment lines (starting with "#") are kept for
readability and just output without the parsing.
Here's the command I am running:
cat /etc/services | \
awk '{ if (index($0,"@") == 0 && index($0,"--none---") == 0) { \
if substr($0,1,1) == "#") print $0; \
else { \
TMP = $1"\t\t"$2"\t\t#";\
for (i = 3; i <= NF; i++) TMP = $TMP $i; \
print $TMP; \
} \
} \
}'
Thanks for any suggestions.
nl
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: OT: awk problem
No, it doesn't help...I tried putting it in a file - same exact problem.
I assume you ran it with mawk? (/usr/bin/awk links to
/etc/alternatives/awk which links to /usr/bin/mawk on my system).
nl
On Sun, 2002-11-10 at 22:53, Elizabeth Barham wrote:
> It works for me although I did place your script into a file and
> ran it as:
>
> awk -f neals-script.awk /etc/services
>
> {
> if (index($0,"@") == 0 && index($0,"--none---") == 0) {
> if(substr($0,1,1) == "#") {
> print $0;
> } else {
> TMP = $1"\t\t"$2"\t\t#";
> for (i = 3; i <= NF; i++)
> TMP = $TMP $i;
> print $TMP;
> }
> }
> }
> }
>
> hth, Elizabeth
>
>
> --
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
>
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: NFS Question
On Sat, 2003-08-09 at 19:19, Nyc0n wrote:
> Im going to attempt to set up an NFS, I have done so in the past and
> failed, ive read all the howtos and everything, but I had one question,
> I already have the drives I want to make into the NFS, they are
> currently reiserfs and already have stuff on them, do they have to be
> blank? Or can I just tell it which drives I want to use and it take over
> the reiserfs ?
NFS isn't really a filesystem in the same way that ReiserFS or ext{2,3},
are. NFS is a protocol for making your existing file systems accessible
to other systems over a network connection. The windows world equivalent
is that you might format your hard drive as FAT{12,16,32}, NTFS, etc and
then share the drive over the network via SMB. You can mix and match
on-disk filesystems and network sharing protocols, too - for instance,
you could have an ext2 filesystem on a hard drive partition on your
linux system, shared via SAMBA to other systems that understand the SMB
protocol (like a windows system).
You do not need to do anything specific to your current file systems,
whether they are ReiserFS, ext2, etc. YOu just need to install NFS, run
the NFS daemons on your server, and export the filesystems through NFS,
and then NFS clients can connect to them. You specify the exported file
systems on the server via /etc/exports, and on the client mount them as
nfs type file systems.
nl
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
How to "fool" apt-get into believing a package is installed?
I am looking to install squid onto a system which runs my imap mail store, using cyrus-imap. As part of the cyrus installation, I have also installed cyrus-sasl. I compiled both myself, because at the time (over 1 year ago) that I did this, the version of cyrus-sasl and cyrus-imap running in stable and testing was quite old and outdated. My problem now is that squid depends on libsasl2, which I already have installed via my self-compiled cyrus-sasl. Rather than risk losing or breaking my cyrus-imap installation by installing the debian libsasl2 package, it would be better to make apt- realize that I already have a sasl2 library installed. Is there a reasonable way to do this? Thanks. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: How do I install a second harddrive
On Fri, 2003-08-29 at 14:38, Wathen, Metherion wrote: > Hi all, > I know that sounds stoopid, but I'm having trouble finding info > that tells me what I need to do to get a second harddrive on my system > recognized by linux. > When i boot the system dmesg clearly shows hdb as found. > however I don't know what to do to see the contents (if any) of hdb in a program > like gmc. > So my question is > 1) what do I use to 'see' the 2nd harddrive > 2) how do I format it for use w/o messing up my 1st drive > 3) how do I make it part of my existing filesystem > I'll take a stab at this, although there are, I'm sure, others who could be of greater help. Your system appears to already "see" the second HD since dmesg shows that hdb is being found; I assume you already had one harddrive called hda, and you installed the second hard drive as the slave drive on IDE channel 0, hence it comes up as hdb. (presumably you also have a cd rom or something on IDE channel 1, coming up as hdc). Anyway, assuming this is a new harddrive fresh out of the box, you need to partition it first, and then format it. DO NOT, repeat, DO NOT do this if this is a previously used HD from which you intend to reclaim data - partitioning and formatting will lose all existing data on the drive. In any case: to partition the drive, you use an fdisk tool - fdisk, cfdisk, etc. I usually use fdisk. As root, you would do "fdisk /dev/hdb" and then using the "m" command in fdisk to see a list of allowed commands. Basically you will create as many partitions as you like, either primary or logical, and of course you would (in general) set the type of partition to linux, assuming you are going to use a standard linux file system like ext2, ext3, or reiserfs. If you are planning any windows type file systems like FAT, NTFS, etc, you would set the appropriate partition types. Once you have the partitions the way you want them, write the partition table to the disk and exit fdisk. I usually reboot after repartitioning the HD, as the program advises, although I'm not sure (under Linux) that this step is absoluately necessary. You want to be really careful that you don't mistype and fdisk /dev/hda by mistake, because then youcould wipe out the data on your existing drive (you have to write the partition table to the drive using the proper fdisk command; no changes are actually written to the drive until you do this, so if you realize you mistyped the initial command to get into fdisk, just exit before doing the write.) "man fdisk" for more info. Once the new HD is partitioned, you can address the individual "drives" just created using /dev/hdbn where n is a number representing the partition on the drive. Primary partitions are numbers 1-4; logical partitions start at 5. Note, however, that you can have a max of 4 primary partitions, and if you want logical partitions, one of those must be an "extended" partition which can contain other (logical partitions). When I need a multipartition drive, I usually made three primary partitions (here called /dev/hdb1 /dev/hdb2 and /dev/hdb3) and one extended partition in which I place the logical partitions (/dev/hdb5 /dev/hdb6 /dev/hdb7 and so on) so there is not /dev/hda4 because the container extended partition doesn't actually "show up" as a usable device. Now you need to format each partition as whatever filesystem you want to use - ext2, ext3, reiser, etc. Use "man mkfs" to learn more. Once the partitions are formatted, you need to mount them. Decide on your mount points, and use the mount command (man mount); also add the appropriate lines to /ext/fstab so the partitions are automounted at reboot. (man fstab). That about does it. Let's assume, for instance, you wanted to make the new drive contain all the data in your current /home tree, and needed to move all the current data in /home to the new partition as well. Once you've partitioned the drive(so you have one big partition, /dev/hdb1), format it and mount it to a temporary location: mkfs -t ext2 /dev/hdb1 #format /dev/hdb1 as ext2 file system mount -t ext2 /dev/hdb1 /mnt# mount to temporary location now, copy old data from /home to "new" location cp -a /home /mnt OR, tar up /home and untar into /mnt with cd /mnt; tar -cl /home | tar -xv#you may want to change options #to suit your needs then replace old /home with new /home umount /mnt mount -t ext2 /dev/hdb1 /home (note all of the above requires that you are root.) Once you are sure everything made it over to the new drive, you may wish to delete the contents of the old /home tree. MAKE SURE you umount the /dev/hdb1 partition from /home to "uncover" the old contents in /home before you do this!!! Remember that in linux, you mount a drive (/dev/hdb1) to a "mount point" (here, /home) and when you do that, the newly mounted drive hides the original contents at that mount point, but it is still there, accessible after you umount the drive f
Re: sony digital handycam (DCR-TRV60E) in linux
I also have a Sony digicam, which is connected to my system via a firewire card. To use it, via firewire, I modprobe'd video1394, ohci1394, raw1394 and ieee1394. All were with standard woody install (later dist-upgraded to testing) and no kernel recompile was needed. I have used both dvgrab and kino to grab video from the camera. They work "ok" but not great; in fact, I noticed an increase in frame-droppage for unclear reasons when I went from stable to testing, but I'm not sure why. I'm not sure how you would do the same via a usb connection if that is what you are using as I've never looked into that. nl On Sat, 2003-08-30 at 12:19, Benedict Verheyen wrote: > Hi, > > > i bought a Sony DCR-TRV60E camera. I can also take pictures as well as > record videos. I can mount the camera as /dev/sda1 to look at the > pictures i've taken but i don't know how i can download the film i've > recorded. > Do i need to compile additional stuff in the kernel? > What programs can i use to playback the film and maybe edit it? > > Thanks for any pointers or documentation. > > Benedict > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Packages: required vs recommended vs suggested
I was wondering if someone could clarify how apt-get handles the various categories of related packages. I did in install last pm of a package which "recommends" other packages, which it turned out I needed in order to make things work. However, apt-get install did not automagically install the recommended packages (it did get required dependencies) and I spent quite a while before I figured that out. How do you cause apt-get to at least ask you about installing recommended, or suggested, package dependencies? nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
[OT] Why does X need so much CPU power?
I'm just wondering if anyone has any info on why X seems to need so much CPU power? Way back when, probably around 1996 or 1997, I first tried to install Linux. Back then, I tried distro's from Corel and Redhat. My system was a Pentium 133 with 48 (and then 96) MB Ram. This system ran both Win 95 and Win NT 4.0 reasonably well, but when I made the switch and installed Linux, any sort of desktop - eg Gnome or KDE, not a vanilla WM) was just so slow as to be unusable. Eventually I gave up for a while and went back to WinNT for some time. For the past 3 years or so, my workstation has been exclusively Linux, first Mandrake on a PIII-800, and for the last year, I've been hooked on Debian on an Athlon XP 1700+, and on both of those systems performance has been just fine, so I didn't really think about the troubles I originally had, and when I did, I figured I must have done something wrong on my first install attempts on the Pentium system. A few months ago, I decided to put debian on my old Laptop, an IBM Thinkpad 770ED (PII-266, 64MB Ram). Once again, with KDE running, the desktop was so slow and unresponsive as to be really unusable (except in an xterm window). This is a system that has run Win95, Win98, and WinNT just fine over the years. So, my question is: Why does X seem to need so much more CPU power than windows - such that systems I have tried to use that worked fine with various windows flavors just were unusable with KDE loaded? I assume the problem isn't in Linux itself, since my old Pentium 133 was just fine with X not running, and enough people have attested to the ability of systems with Pentium processors running Linux without X being able to handle massive firewall, router, web server duties, etc. Maybe the problem is KDE and not X - but I had similar trouble with Gnome, so it isn't just a KDE issue. I'm just curious and wonder if anyone has any thoughts. Thanks. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: [OT] Why does X need so much CPU power?
On Mon, 2003-09-01 at 18:20, Nicos Gollan wrote: > On Tuesday 02 September 2003 00:02, Neal Lippman wrote: > > I'm just wondering if anyone has any info on why X seems to need so much > > CPU power? > > It's not X eating resources like mad, it's the way desktop environments > forcing it to do things that it was never meant to do.It was never meant to > display eye candy like KDE and Gnome feature. You'll find that it's doing > just fine with a "lighter" window manager that doesn't use transparencies and > tons of bitmaps for window decorations (FVWM2, OLWM, WindowMaker, etc.). > WindowMaker should run OK on a Pentium 266 measured on its performance on my > 150MHz laptop w/32MB RAM. After some time you won't miss too many things. > > IMO the whole X(free) system needs a healthy kick in the butt. It's one of the > main factors in keeping Linux away from the desktop, not just lacking in > performance and features, but also a royal PITA to configure with new > problems cropping up every five minutes. > > I'm going to bed now. But perhaps this one will keep people away from the > "Quoting" and C popularity threads which are scrolling off to the right; > reading them is like coding python with a tabwidth of 8. (xinerama is another > thing in X that's FUBAR while we're at it, I literally *lost my mouse > pointer* while trying to set it up.) > Well, most replies to my posting have pinned the "blame" on KDE and Gnome rather than X per se. I'll have to reinstall on the laptop and see how it looks with a more minimal WM. This does still beg the question of how Win95/98/Me/NT, etc, managed to provide a reasonable "desktop" when KDE/Gnome could not, however. It really doesn't seem to me that either KDE or Gnome provide a more complex desktop environment than Windows, at least not from the end-user perspective, even if the underlying OS (eg Linux vs Windows) is more robust and possible more feature-full. From what little I know of X, I'd tend to agree that X is being overtaxed supporting a desktop environment that it was never designed to do. Aside from the present market penetration of X (which could also be used to argue to stick with Windows instead of ever having adopted Linux), what would be the obstacle (other than, of course, the time/effort for development) for a new graphics paradigm to sit atop Linux? [Yes, I know there'd be a lot of apps to redo and so forth as well, although if there were a Gtk+ compatibility layer...) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Packages: required vs recommended vs suggested
On Mon, 2003-09-01 at 02:05, Bob Proulx wrote: > Marc Wilson wrote: > > Neal Lippman wrote: > > > I did in install last pm of a package which "recommends" other packages, > > > which it turned out I needed in order to make things work. > > > > Then perhaps they're actually dependencies? Did you file a bug? > > For discussion purposes what package are you talking about? > The specific problem was with libsasl2 which lists libsasl2-modules as recommended, but in fact, without these (or other) plugins, the package is unusable (see the output of apt-cache search libsasl2). I suppose it is possible that libsasl2 makes libsasl2-modules a recommended but not required dependency in case someone has other sasl2-pam plugins available, but I'm not sure where this would be the case (the modules contained in sasl2-modules are part of the source that comes with cyrus-sasl on which libsasl2 is built). Since I'm far from a sasl expert, I emailed this info to the cyrus-imap maintainer (who was very helpful in getting my install of cyrus21-imapd going) to see what he thought about this before I file a bug. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
i865 based systems
Has anyone installed a 2.4 series kernel on a system based on the intel I865 chipset? If so, I am wondering if there are any recommendations re: motherboards that seem to work well. >From a quick grep through the 2.4.21 source (latest kernel available for testing) is appears that both i865 support and sATA support are in the kernel, which are the two major issues with this chipset (since I think the onboard gigabit LAN became supported in 2.4.19). As far as I can tell, there is sATA support in 2.4.21, although I cannot figure out what the device major number or /dev entries for that would be. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: [OT] Why does X need so much CPU power?
On Wed, 2003-09-03 at 18:01, Micha Feigin wrote: > The main problem I see with linux is the lack of commercial programs. > Unfortunately for some stuff there is no way around it. For commercial > quality image/video processing for example there is no alternative at > the moment, or places where you need to be able to show reliability > certificates, which cost quite a lot with free software. Sometimes you > also need some to be legally liable if something goes wrong when running > critical systems, and that costs money. > I think that whats holding linux back from the home market is mostly > that people tend to stay with the preinstalled os they get with the > computer, since replacing it is usually to daunting for most users. Also > linux currently has a name as a hard to install/configure/maintain for > geeks only os. To get it into the home market, it needs to change its > market image and arrive preinstalled to change the market share. > I think this is well said. The fact is that, in my opinion, most people would be able to work quite satisfactorily with a good installation of KDE or Gnome, at least as well as with Win98. Where Windows has Linux beat is a) the OS comes pre-installed which just plain makes it easier, and b) there really is good consistently (not perfect, but good) between the way the start menu is configured from install to install. This is in contrast to both KDE and Gnome, where the K or G menu comes up with a mess of programs, not well organized into logical categories and submenus, and often with menu items created without the programs installed (for instance, a "Games" submenu even though I never install the games programs). The lack of good commercial apps really is a problem that we open source zealots don't want to acknowledge, and the reason it's a problem is very straightforward. While I think we would all agree that the quality of the Linux kernel, X, KDE, Gnome, etc is at least as good and often better (like the kernel) than the equivalent components in Windows-land, the fact is that many of the apps that we use regularly are not as slick, polished, or feature-rich as similar programs in windows land. I think that this is largely because while some of he large projects (kernel for instance) have many developers and a a fair number of those developers are working full time on Linux under the auspices of whatever Linux or non-Linux company sponsors them, a large number of the other programs that would be mighty useful for the Joe-Desktop-Windows-95 user are written by one guy in his spare time trying to hold down a day job - and it's just plain hard to get a lot of quality programming down in the odd hours between when the kids are in bed and when I need to go to bed myself to be up the next AM for my real (non-computer) job. Here's an example: digital photography. Kudos to the gphoto team aside, so far I've only been able to find ONE application that handles the highly useful task of importing digital photos either from a digicam or from jpg files on disk and displaying them in a photoalbum kind of interface, and that program is unfinished and sparse compared to similar programs for Mac or Windows systems. [Aside: This is not intended to be a poke at the guy writing this program, far from it - I wish he could work on it full time so that I could get it to use!] For the Mac you get iPhoto, the "definitive app" in this category; for windows there are many options including the highly rated Adobe Photoshop Album. But nothing at that level for Linux. The difference: the programs for Windows and Mac are developed by companies devoting teams to this full-time, so no wonder they make faster programs. Heck, if we could get as many people working on "lPhoto" for Linux as there are on the kernel... Another problem Linux faces is that frankly too much choice is as bad as too little. Having competing desktops, while often put forward as a advantage (choice is good), is fine if you are an enthusiast who likes experimenting with KDE, Gnome, Windowmaker, Blackbox, etc until you get just what you want, customized the way you want it. But for Joe-Desktop buying a computer, even having to choose between K and G during his standard Dedhat install is just a decision he cannot make - so he goes to Windows, where no choice really is a better choice. Having development efforts all focussed on ONE really good, fast, well-written desktop (with an advanced config mode for those who really really really want to customize the appearance and function of every last pixel) would, I think, really help Linux move onto the desktop. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
"Restoring" system after MB/processor upgrade
I'm thinking about upgrading my system from its current AMD Athlon XP based system to a P4 (actually, I need to put together a new system for my son, so I'm thinking of giving him my current MB/processor and putting a new MB/processor into my system). Obviously, I don't want to lose my current testing installation in the process. I'm hoping that an MB/processor swap into the case and a reboot will leave me just with a rerun of modconf to change installed modules to match the new hardward's ethernet and sound cards (I'm goign to reuse my current video card) and I'll be back in business. In case things go awry, however, and I have to do a new fresh install, I was thinking that I could use apt-show-versions to good effect. My thinking was to use apt-show-versions -b to get a listing of all installed packages, run the output through sed to remove the "/testing" part of each package so I am left with just a listing of installed packages, and save the output to a file. apt-show-verions -b | sed -e "s/\/testing/" >saved-package-list Then, if the whole thing goes kablooey and I need to reinstall from scratch, I just throw in my woody CD, redo the install and configure modules. Then, I can re-update to testing with dist-upgrade, and then use apt-get install $(cat saved-package-list) to suck back in all the packages and stuff presently on my system. Since all of my data and my /home are on an nfs share from my server, including all my kde menus and stuff, this should (hopefully) give me a full system restore. Comments - will this work? nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: "Restoring" system after MB/processor upgrade
On Thu, 2003-09-04 at 16:22, Karsten M. Self wrote: > on Thu, Sep 04, 2003 at 12:07:04AM -0400, Neal Lippman ([EMAIL PROTECTED]) wrote: > > I'm thinking about upgrading my system from its current AMD Athlon XP > > based system to a P4 (actually, I need to put together a new system for > > my son, so I'm thinking of giving him my current MB/processor and > > putting a new MB/processor into my system). > > > > Obviously, I don't want to lose my current testing installation in the > > process. I'm hoping that an MB/processor swap into the case and a reboot > > will leave me just with a rerun of modconf to change installed modules > > to match the new hardward's ethernet and sound cards (I'm goign to reuse > > my current video card) and I'll be back in business. > > Most likely, nothing's going to change. That was my assumption / hope...but you never know. > > There are no drivers specific to the CPU (mostly), though there are > kernel flavors specific to specific kernel architectures. If you're > upgrading CPUs within the P4 familiy, it's a transparent change. Well, I'm going from an Athlon XP to a P4 system; the new system will probably by i865 based - as far as I can tell, 2.4.21 supports the 865 chipset. With sATA, on board lan, etc, there will be some modconf'ing to do. The only package I am aware of that is specifically compiled for the K7 platform is mplayer, so I'll obviously need to remove and reinstall the 586 platform version. > > Best bet is to have bootable media (boot floppy, Tom's Root Boot, > LNX-BBC, Knoppix) handy. I always keep a tom's root boot handy just in case (it's also helpful at work when I need to reboot a windows workstation into Linux too ). -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
autofs permissions
I am wondering how I can control the permissions of a directory created dynamicall using autofs. I have autofs4 enabled (sarge). I have a map for the /mount directory (called /etc/auto.mount), which has an entry for my floppy drive (/dev/fd0). The problem is that ordinary (non root) users cannot write to a floppy. If I put in a blank floppy and try to copy a file to it (cp /mount/floppy), the autofs created directory /mount/floppy has owner.group root.root (the owner/group of /mount),and permissions rwxr--r-- (the permissions of the /mount directory). Presumably autofs is creating the /mount/floppy directory with the propagated permissions of /mount. I don't want to chmod /mount to rwxrw-rw-, because that would allow ordinary users to potentially drop files, etc, into /mount itself. Does anyone know of a way to affect the behavoir of autofs to create the subdirectory with the needed permissions? I though of changing the map to run a script that does the mount and chmod's is; I don't know if that is possible or not. I could chgrp /mount to some useful group (like floppy), set its permissions to rwxrw-r-- and let anyone who needs to write to the floppy be in the floppy group, but that would still allow those users the ability to write into /mount, and that's not what is required - they just need to be able to write into /mount/floppy. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
CD label maker - what's your preference?
I am looking for a utility for making cd labels (round ones to stick onto the CD itself) and jewel case covers, and need to do this both for audio cds (I like to copy my CD's so I can have a copy at home and a copy in the car) and for data cd's. A scan through apt-cache shows (at least) the following offerings: cdlabelgen cd-circleprint disc-cover kover Another choice, kcdlabel, which I used a while back when I ran Mandrake, doesn't seem to be available, at least in the official debian packages. Any comments on relative merits / drawbacks for these packages greatly appreciated. Thanks. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Security Questions
A few answers, but first a question: How do you know that your network has definitely been compromised? If the only evidence you have is that your daughter received returned emails she didn't send, how to you know that someone didn't generate those emails elsewhere, spoofing her email address and reply-to header? You could gain some insight by looking at the emails more carefully; if the emails were bounced by a mail relay along the way and returned, then you can see by the headers if the originated at your normal smtp (outgoing) server or not, which would give you some information. Just a thought. Assuming your network IS compromised (and even if it isn't), remember that a wireless network cannot be 100% secured with current implementations, but you can do a lot to limit breaches: employ WEP encryption, disable SSID broadcasting, set up your router to allow access only by specified MAC addresses (this is the hardware address encoding into every ethernet interface, not the IP address used by the corresponding computer), and so on. There was an excellent article on wireless LAN security on arstechnica.com a few months ago that you should read. My personal preference for dealing with a compromised system is to fully wipe the HD (repartition and reformat) and reinstall everything. There are, I suppose, other approaches you can use - for instance, if you compared the size, timestamp, and contents of every "system" file on a suspected system to those on a clean distribution and found no mismatches, it would be unlikely at a root kit was installed. I have seen occasional references to scripts that can do this here and there, but don't recall any of them offhand. Another approach is to build a "snapshot" of the entire system (file name, timestamp, permissions, size, uid.gid, etc) into a file. YOu can then periodically compare the state of the current system to the saved database (stored, of course, offline to avoid being changed by a malicious intruder) and that would tell you if the system has been changed or not. That doesn't help you here, but does help to detect a later breach. (Remember, if you do something like this, that every time you apt-get install/remove/update/upgrade/dist-upgrade you change the system, so you need to run the scanner to verify system intregity before any apt-get - to ensure the system is ok before you start -, and again after the apt-get in order to create a new snapshot). Lastly, bear in mind that as long as you allow an "unsecured" system behind your firewall (your grandson's computer) there is no way to ensure the security of your network. Good luck. On Sun, 2003-04-06 at 10:09, Thomas H. George,,, wrote: > I have read Security-Quickstart-HOWTO. > > I believe my home network has been compromised (my daughter received > returned emails she neversent) and plan to take drastic action. The > network consists of DSL modem, a wireless router and four computers. I > have no concerns about the family members and the houses in the > neighborhood are widely separated so it is very unlikely that the > wireless connection has been used by outsiders. The DSL link to the > internet is my concern. Here are my quesions: > > 1. How to erase hard drives? I plan to pull one computer off line and > reinstall Debian Woody and Windows from CD's (Regretably I still need > Windows for a few applications). Is reinstallation enough or must, and > can, the hard drives be wiped clean of any residual programs? > > 2. What is the best Firewall? I have an old Compaq 486 machine with no > math coprocessor. I assume I can install two ethernet cards (I believe > it has two PCI slots, must look though), load Woody, set up iptables and > a sniffer and place it between the DSL modem and the wireless router. > When I am ready to put this firewall in place I have all the computers > off line. I will bring up the one that has its operating systems and > applications reinstsalled from CD's and download all the security > updates from Debian and Microsoft. The procedure can then be repeated > for the other computers. > > 3. DHCP or static addresses? I have been using static addresses. I > believe I have seen in the references that it is possible to set the > wireless router to receive and transmit to these addresses only? If so, > is this the best approach? > > 4. How to deal with a rogue computer? The fly in this ointment is my > grandson's laptop, a gift from his father (my daughter's ex-husband). > It came with XP Professional and I don't have the CD's to reistall it. > My grandson likes to go on the internet and also use our wireless > network to print his homework on one of the printers attached to the > fixed computers. Would it work and not compromise the system if I give > it a static address and instruct the other computer's on the network to > refuse any transmissions from this address? And could I then attach one > of the printers to the computer serv
Re: Cheap high resolution (1600x1200) video card for debian
I am using an ATI Radeon 7500 on the DVI output at its maximum res of 1600x1200, and it works fine with X under testing. This card can do higher resolutions on analogue as well. I would suggest avoiding an OEM card based on the Radeon chips. I have read that the OEM cards, for unclear (to me) reasons to not work properly with standard XFree drivers, although perhaps ATI's own drivers would work better. I had an OEM version initially, and there was a line of incorrect pixels runnign down the side of the screen. This problem went away when I got a new ATI-branded 7500; of course, I don't know if this was just single defective card or a problem with teh OEM version. I have seen the ATI-branded 7500 cards with dual VGA/DVI output for around $75, so it's not that much more expensive. nl On Wed, 2003-05-28 at 10:01, J F wrote: > Is there a cheap high resolution (1600x1200) video card for debian. > > My old system is running an ATI 3D Pro Turbo PC2TV installed > 4 years ago. > > The thing I liked about it is I have a Virtual desktop > of approximately 2000 x 1800 > with a panning window of 1600x1200. > > Fonts were readable even if medium size. > > Does the Radeon 7000 have or the Nvdida have > this capability? > > I would even go for a virtual desktop of 3200x2400 > if the card and driver would let me without a > lot of hassle. > > 3D is not important, but 2D and crispt text > is very, very important. > > I see some ATI OEM radeon 7000 for $32 to $45 > and wonder if they would work ok? > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Compile options for standard debian kernel?
Is there a way to find out the config options that were used for compiling the standard 2.4.19 kernel that comes with stable or testing? I need to compile a new kernel as the stock kernel does not seem to have the udf file system support enabled, but it would be handy to know what was compiled in the distributed kernel vs modules, etc, so that I don't mess up my system configuration when I configure the new kernel. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Compile options for standard debian kernel?
Yes. Thanks much! On Sun, 2003-06-01 at 21:18, Kenton Brede wrote: > On Sun, Jun 01, 2003 at 09:07:08PM -0400, Neal Lippman wrote: > > Is there a way to find out the config options that were used for > > compiling the standard 2.4.19 kernel that comes with stable or testing? > > > > I need to compile a new kernel as the stock kernel does not seem to have > > the udf file system support enabled, but it would be handy to know what > > was compiled in the distributed kernel vs modules, etc, so that I don't > > mess up my system configuration when I configure the new kernel. > > > > Should be in your /boot directory. Probably called config-2.4.19. > hth, > kent > > -- > "I am always doing that which I can not do, >in order that I may learn how to do it." --Pablo Picasso > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Compile options for standard debian kernel?
Thanks. I actually have 2.4.18-bf24 currently, and UDF isn't mentioned at all in the config file. I thought it was included in the sources for 2.4.18, but that may be wrong. In any case the module does not appear on my system and modconf does not list it, so I need a recompile or update to 2.4.19. On Sun, 2003-06-01 at 21:37, Donald Spoon wrote: > Neal Lippman wrote: > > Is there a way to find out the config options that were used for > > compiling the standard 2.4.19 kernel that comes with stable or testing? > > > > I need to compile a new kernel as the stock kernel does not seem to have > > the udf file system support enabled, but it would be handy to know what > > was compiled in the distributed kernel vs modules, etc, so that I don't > > mess up my system configuration when I configure the new kernel. > > > > nl > > If you installed your current 2.4.19 kernel from a Debian > "kernel-image" deb, then you will have a copy of the ".config" file used > to create it in your /boot/ directory. It will probably be named > something like: "config-2.4.19-k7". > > BTW, here is what my 2.4.19-k7 kernel config file says about UDF: > > CONFIG_UDF_FS=m > # CONFIG_UDF_RW is not set > > Looks like it is already setup as a loadable module for read-only > support. You might try modconf and see it it is listed there... it is > listed on my system. If it is, try loading it and see what happens. It > might save you the trouble of a re-compile... > > Cheers, > -Don Spoon- > > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Modules weirdness after upgrade to testing...
Tonight, I took the plunge and upgraded woody->testing, so that I could get X4.2.1 and try out my new ATI 7500. Anyway, there are a few weird things that I could use some help on; this post relates to modules. I did the upgrade by changing all occurrences of "stable" to "testing" in /etc/apt/sources.list, then: apt-get update apt-get dist-upgrade On bootup, I am getting error messages that state: Dec 21 22:12:00 gandalf insmod: Note: /etc/modules.conf is more recent than /lib/modules/2.4.18-bf2.4/modules.dep I'm not quite sure why this is occurring. I tried "sudo depmod -a" to update modules.dep, but on reboot the error message is unchanged. I was assuming I would not need to do "update-modules" under the assumption that update-modules was already part of the dist-upgrade process... cat /proc/modules shows the right stuff installed. So, do I need to be concerned about this, and if so, what do I do to fix it? Thanks. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
After upgrade to testing, ?zinf problem
I just dist-upgraded from woody->testing tonight. After the upgrade, I tried to test things out to ensure nothing was broken, and when I got to testing sound, tried to play an ogg file with freeamp - which just didn't work right. I also could not get ogg123 to work, since it demanded libvorbisfile.so.0, and after the upgrade I only seem to have libvorbisfile.so.3. (Note: I couldn't guarantee ogg123 worked before the upgrade, since I never tried it as freeamp always worked so well). A bit of looking around let me to conclude that freeamp is superceded by zinf in testing, so I tried: apt-get remove freeamp apt-get install zinf zinf works, in the sense of playing the files. However, when I open the equalizer and try the balance control to shift from centerline to playing only on the left or right speaker, only the left plays - and after using the balance control, there is no sound at all until I adjust the sound level control. I don't remember seeing this problem at all with freeamp, and I also note that sound is fine with KDE generated sounds (such as when you log in from kdm) - so the sound stuff seems to work ok (as it should, really, since I did not change the kernel or modules). Is this a zinf bug - or is something misconfigured? And, is ogg123 still supposed to work,and if so, why didn't it upgrade to understand the appropriate libraries that I have? (oh, btw, /usr/lib/libvorbisfile.so.3 is a symlink created tonight at the time I did the dist-upgrade, so I assume this is an upgraded version or a new file...) Thanks. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
A little dpkg wierdness after upgrade to testing
I noted a little weirdness with apt-get and dpkg after my dist-upgrade from woody->testing tonight. My practice is to run in my usual user-level account (nl) and use sudo for all root-like operations. When I install, for install, I use 'sudo apt-get install ' and so on. That has always worked just fine, even when my current directory is /home/nl. Of note, however, /home/nl is just a symlink to /nfs/nl, and /nfs is (surprise) and nfs mounted share from my fileserver. The relevance of this is that when running as root on my workstation, I cannot create/access files in my usual home directory (/home/nl) because that maps to the nfs share which as root_squash for security reasons. Now, to the point. After the dist-upgrade to testing, when I do sudo apt-get install , it fails because it is trying to create a file called .dpkg (or something similar) which it cannot do, because my current working dir is /home/nl. Of course, it's easy to fix this by just cd'ing to /tmp which is mounted on a local partition. However, I bring this up because this did NOT ever happen under woody. This makes me think something has changed in apt or (more likely) dpkg - and apparently dpkg is trying to create a file in teh home dir of my personal account even though, via sudo, it should be finding the home dir for root. I don't know if this is a bug, feature, or "don't care" - but I thought it was worth asking about. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: network print server
On Sat, 2003-01-11 at 21:21, Michael West wrote: > I have been using samba as a print server at home, but > I am often mucking around on my debian box, and sometimes this > results in my wife not being able to print. > > To increase user satisfaction I am considering purchasing one of > these dedicated print server appliances which attach directly to > the network. It must work for win98, XP pro and Debian. Ideally > it would have parallel and usb support. > > I am looking for recommendation as to where to research what would > work best. I couldn't find anything at linuxprinting.org on this. > I would guess that just about any printserver will work fine for you. I have been using a Linksys 3-port printserver (has three parallel ports) for several years without any problems. While I don't think this particular model is still available, I would imagine that the current models work quite similarly; in fact, there was info on the Linksys website when I first bought it that told me how to make it work with linux. The printserver that I have has a built-in lpd server, so you just talk to it using lpd. It may also support IPP, I'm not totally sure of this. It was easier to configure under Linux - I just told CUPS to use it as an lpd server at DeviceURI lpd://192.168.1.10/L1. This particular device labels its three ports L1, L2, and L3. There are instructions for setting the IP address for the device here: http://www.linksys.com/support/support.asp?spid=18#tcpip and you can do this using telnet or tftp pretty easily. Under W95/98, it was actually a bit harder to set up because I had to install supplied "TCP redirector" software which basically allowed me to create "printer ports" that redirect to the printserver. I think this is much easier under XP as my understanding is that XP supports IPP directly. I'm not totally sure about this as I pretty much boot my windows 98 machine only for Quicken, and even that's going away now with Quicken working pretty well under CrossoverOffice/Wine. Good luck. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
email-fax gateway - need suggestions
I was hoping some folks here could give me some suggestions on setting up an email to fax gateway. My project is basically to provide a mechanism so someone can email to a an address in my office (eg "[EMAIL PROTECTED]") and have that email automatically faxed out. This idea is to eliminate the manpower requirement for faxing out reports, which is overwhelming our staff who presently have to print the email and manually fax each report, sometimes to multiple recipients. I am aware that this problem has been addressed in various ways using efax or hylafax, but my requirements are a bit unique compared to the solutions I've seen: 1. Most commonly, the document that needs to be faxed will be an MS Word document, send as an attachment to the email message. As a result, simply trying to fax the email body itself won't work. 2. The people generating the email that needs to be faxed will generally know the name of a person (or people) to recieve the fax, but not their fax numbers. We will have an LDAP directory online which will contain fax numbers, so the handler needs to look up the recipient(s) and translate names to fax numbers automatically. 3. The sender should get back an email acknowledgement so he knows if the fax went through or not. 4. After the fax is sent, the file itself (word document) needs to be saved on our fileserver for later reference as well. Our present email implementation uses cyrus-imapd as a mail store. Our email is received at an external pop3 server. We use fetchmail running as a daemon to periodically retrieve email for all accounts and forward the email to cyrus (via an lmtp connection). I was therefore thinking of the following solution: 1. Write a demon that listens for lmtp connections from fetchmail which will forward fax related emails to this demon via lmtp (I am in the process of writing something similar to handle emails and forward them to the printer automatically, so I can reuse this code anyway). 2. The demon can process the email body based on mime-time. For straight text, should assume a first line of the format "FAX:[,recipient...]>". For MSWord attachments, the process is more complicated: The filename will consist of the local patch to store the file, followed by a list of recipients. Via this mechanism, the demon knows both where to store the file on the file server and to whom the file should be faxed. 3. Using wv, the file is converted from MS Word to postscript, and can then be fed into efax for transmission. The exit status of efax indicates whether the fax went through, which can be emailed back to the sender. I would appreciate anyone who knows of a better / simpler solution, or comments on what I have proposed I would very much appreciate it. nl PS: I am aware that there isn't any security in the above system, and I recognize the opening this gives for someone to use my fax as a forwarding station. I haven't yet decided how to handle security in a meaningful way, but I'm also open to suggestions on this score. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Mplayer help
Looking for some help. I've installed mplayer from the .debs available at deb http://mplayer.nmeos.net stable/ deb http://marillat.free.fr/ stable main and it seems to work ok (a few errors on startup about missing fonts and all, but I can live with that for now). I would like to be able to view realmedia files via mplayer as well. I obtained the RealVideo for Linux from RealMedia, and tried installed the libraries that came with it. I copied all the library files in the "Codecs" directory from the RealVideo install into /usr/local/lib, added /usr/local/lib to /etc/ld.so.conf, and ran ldconfig. However, mplayer still cannot play realvideo files. Does anyone know if there is something else I need to install/configure for this to work? Is the version available from the above source NOT compiled to support realmedia files, and so I should uninstall and compile one for myself? N -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Mplayer help
On Tue, 2002-09-03 at 07:33, Udo Schlaepfer wrote: > Burkhard Ritter <[EMAIL PROTECTED]> writes: > > > Date: 22.06.2002 18:33:20 (CEST) > > ?? > > > does mplayer support realmedia? if that is the case you have probably to > > compile it from source and configure it in the right way. so yes, give > > compiling a try. > > Yes mplayer does. One should read the docu (§2.1.1.6, §2.2.1.8, §2.2.2.5). > > Tschoe Udo. > I did indeed read the directions - which state that mplayer "should detect the RealPlayer libraries in the standard locations of a full installation." My problem is that I don't know what that standard location should be, and that's what I am looking for help with. I tried copying the realplayer libraries (obtained as suggested in the docs from RealMedia) to /usr/local/lib, adding said dir to /etc/ld.so.conf, and running ldconfig, but that didn't cause mplayer to be able to recognize .rm video files. I'm wondering what other location I should be trying. Thanks. N -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
NIS setup question
I don't think I understand the NIS Howto. I was left with the impression that with libc6 installed, I would not need ypbind because libc6 knows about NIS all by itself. I have set up another system as an NIS server (running ypserv), but my client system cannot seem to see the information from that server. It is the case that ypbind must be running on any NIS client system regardless of whether it is using libc6 or not? Thanks. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
How "stable" is "testing"?
I am wondering how stable people are finding testing for use on their workstations. I am running woody, and very happy with it. However, I would like to be a bit more up to date with some of my software - for instance, I'd like to be using KDE 3 instead of 2.2, and the newest evolution, so I was thinking about doing a dist-upgrade to sarge. I don't, however, look forward to severe breakage now that I finally have my system configured and working. Any advice appreciated. Thanks. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: HTML editing
On Tuesday 11 December 2001 01:25, Alec wrote: > Hi > > I just discovered that Mozilla has a gread WYSIWYG HTML editor. Up until > now my favorite choice was VIM. I'm still trying to figure out how to start > that editor without starting Mozilla itself, but does anyone know if KDE > offers any WYSIWYG HTML editors of similar excellence? > apt-cache search kde | grep -i html > produced nothing :( > You are right! I never tried editing HTML with Mozilla composer before, but I just gave it a quick spin and it looks excellent. I am unaware of any specific KDE-based HTML editor. I have used Quanta Plus and Bluefish, but unfortunately have found problems and buggy behavoir in both, which were disappointing.
Help getting KDE!
I could use some pointers; I've gone back a few months in the mailing list archives, and just cannot find an answer to this. I'm a debian newbie, having been running Mandrake for the past year. I just got a second computer to use for debian, figuring I'd work out all the kinks on the install there before trashing my "main" machine with a from-scratch debian install. Anyway, I got potato CDs (burned them from downloaded isos) and installed potato on the new system, without installing X. I then upgraded to woody (changed stable to testing in sources.list, and did apt-get update and apt-get upgrade-dist) and this seemed to go ok (actually I lost the "man" command, but apt-get install man-db seemed to fix that). My problem now is: How do I get kde? apt-get install kde doesn't work, telling me I need to specify a package name. Looking at the listing of packages in /var/apt/cache, there are a whole bunch of kde-related packages, but it isn't clear what to select to just get the whole kde installed (I am assuming that when I apt-get kde, X will come along automatically due to dependencies). Do I need to install each kde package individually? That doesn't seem right, since I thought the whole point of debian (well, one of the whole points, anyway) was the power and simplicity of apt-get. I also tried tasksel, but while it does give me gnome as a choice, I prefer kde and that doesn't seem to be one of the pre-configured tasks. Help appreciated, especially if there is some location I should be looking in that gives this sort of information in general for various packages. I have a bunch of stuff on my MDK system that was installed by the MDK installer automagically (like cdrecord and cups). I don't mind getting the specific things I want via apt-get (in fact, I prefer it so my harddrive doesn't resemble a junkyard filled with stuff I don't want) but it's a bummer if I have to ask for help to identify the packages needed for each thing. Neal
How to read CD label?
Is there any utility that can give me the lablel on a CD (mouinted or unmounted - doesn't matter to me). I generally burn CD's with a volume label (with cdrecord), and I use them to keep track of what is on the CD, especially for backups which I encode with the date of the backup for future reference. However, when I mount the CD (I actually use autofs) I don't see any way to easily tell what CD is in the drive. 'man -k cd 'didn't turn up anything that seems to do the trick, and if it is in the CD howto, I missed it. Thanks. Neal
OT: Need help from bash experts...
Sorry for OT posting; I am not sure if there is a newsgroups for bash experts, so I figured I'd try here. I need to be able to write a bash script that can copy files from a directory who's name includes a space (long story, but it's a windows directory under Win98 on a machine on my network). Anyway, supposing that the directory is "/mount/windows/spaced name" and I need to copy all of the files in the directory to a target. At a bash prompt, I can issue either: cp "/mount/windows/spaced name/*" target OR cp /mount/windows/spaced\ name/* target and all works fine. However, from within a bash script, something like: #!/bin/sh sourcedir=/mount/windows/spaced\ name cp $sourcedir/* target fails, because the space isn't properly passed to cp, AND further the shell doesn't do expansion on the wild card in the file name. I have tried all sorts of variants: sourcedir="/mount/windows/spaced\ name" sourcedir="/mount/windows/spaced\\ name" sourcedir=/mount/windows/spaced\\ name" and the cp command with various quotes and not quotes, and cannot seem to get this to work. Within the script, if I try: sourcedir="mount/windows/spaced\ name" anotherdir="mount/windows" echo $sourcedir/* echo $anotherdir/* the first echo shows that there is no expansion of the wild card, while the second echo works as expected, echoing the name of every file in /mount/windows. Clearly the problem is that dreadful space in the directory name. Any help on how to syntax this greatly appreciated. Thanks. N
Re: autofs
On Tuesday 12 February 2002 14:40, Mirek Dobsicek wrote: > Hi all, > I feel a bit lost in running autofs. > > I've installed autofs package, compilled autofs4 support into kernel > and hoped that now my /floppy becomes automounted, but no :-(( > > I read man pages from autofs and auto.master, but i'm too lame to > understand it. Please help me. > Perhaps you could post the following data, which might help in troubleshooting: 1) Contents of /proc/modules 2) Contents of /etc/auto.master 3) Contents of specific config files pointed to in /etc/auto.master (etc /etc/auto.mount, /etc/auto.mnt, whatever) N
Re: newbie .. how to start with autofs
On Thursday 08 November 2001 04:25, Mirek Dobsicek wrote: > Hi, > I'd like to have my cdrom and floppy, automatically mounted and umounted. > A friend told me "use autofs" . But this advice is not any help to me :-( > > Mirek Dobsicek I've been using autofs under MDK 7.2 and 8.1 for some time now (yeah, I know - but I haven't had time to face a new install and upgrade process to change to Debian, although I've been planning it for a while). Anyway, autofs works extrememly well (for me). I use it to automatically mount my DVD, CDRW, floppy, a windows share over the network, etc. Basically, there are two components: the autofs code in the kernel, and the automount daemon which is used to do the actual mounting on demand. You need to ensure that your kernel is compliled with autofs support in it. There are two versions, autofs (which is the version 3 form) and autofs4 (version 4). I use the latter, although I am unsure what the differences are. If you choose to build autofs4 as a module, you need the line alias autofs autofs4 in /etc/modules.conf so that modutils will know to look for autofs4.o (the autofs4 module) when the kernel needs autofs. If you are using autofs you obviously don't need that line. You also will need a startup script to launch the automount daemon on startup; presumably the .deb package that automount comes in for debian will include such a script. The system is configured with several files. /etc/auto.master is used when the daemon is started / restarted and contains the "mount points" which autofs controls. Each mount point is the root of a directory tree, under which automounting can occur. For instance, here's my /etc/auto.master: # $Id: auto.master,v 1.2 1997/10/06 21:52:03 hpa Exp $ # Sample auto.master file # Format of this file: # mountpoint map options # For details of the format look at autofs(8). #/misc /etc/auto.misc --timeout=60 #/net /etc/auto.net --timeout=60 /mount /etc/auto.mount --timeout=15 You can see that I have commented out the ones that came with the MDK install, and I have a mount point called /mount, under which all my automounting takes place. You can have as many such mount points are you want. This file specifies that the details of how to handle /mount is containined in the configuration file /etc/auto.mount, which looks like this: # $Id: auto.master,v 1.2 1997/10/06 21:52:04 hpa Exp $ # Created 10-21-2001 NL # This is an automounter map and it has the following format # key [ -mount-options-separated-by-comma ] location # Details may be found in the autofs(5) manpage dvd -fstype=iso9660,ro,nosuid,nodev,exec,user :/dev/dvd cdrw-fstype=iso9660,ro,nosuid,nodev,noexec,user :/dev/cdrw floppy -fstype=auto,nosuid,noauto,user,nodev,unhide:/dev/fd0 windows -fstype=vfat,user,exec,umask=0 :/dev/hda2 kirk-fstype=smbfs,user,nosuid,noauto,noexec,rw,password=frobozz ://kirk/ddrive So, if I cd to /mount/dvd, autofs mounts /dev/dvd onto the mount point /mount/dvd, as an iso9660 filesystem. Similarly for cd to /mount/cdrw; cd'ing to /mount/floppy mounts /dev/fd0 and autodetects the file system on the floppy in the drive. /mount/windows is a windows partition on my computer in /dev/hda2, which /mount/kirk automounts the windows share //kirk/ddrive onto /mount/kirk as an smb filesystem. All of this works magically and transparently once configured. Once you have autofs / automount installed, man autofs, man automount, man auto.master will tell you all about how to handle this stuff. You can automount NFS file shares, or even have scripts executed (so that if you set things up correctly, for instance, you can have autofs handle things so that if you cd to /mount/xxx where xxx is the name of any system on your LAN, the /home directory on that system is automatically mounted for you - and you can set this up to work as new systems are added/removed from the network without having to change your configuration files once the initial scripts are set up. It's very powerful and I have not had a single glitch (or at least, the only glitch I have is because of the way windows does (or does not) wake up properly from a suspend on my laptop). Good luck. Neal
Re: persistent storage hardware: recommendations, comments, and opinions please
On Monday 21 January 2002 22:26, [EMAIL PROTECTED] wrote: > > "dman" == dman <[EMAIL PROTECTED]> writes: > > dman> First my current setup: > dman> 10GB Maxtor IDE disk > dman> Size Used Avail Use% Mounted on > dman> 5.6G 5.1G 271M 96% / > dman> 4.0G 2.5G 1.3G 64% /home > dman> 125M 40k 124M 1% /tmp (tmpfs, not on-disk) > > dman> I want to buy another disk to augment this. There's not enough > room dman> on this one for my music _and_ an OS (and working files -- > homework, dman> etc). I would with briand's comments. IDE drives should be plenty fast enough for most desktop/home use systems, and even for a small server (eg something for a home LAN and/or small office setup) and will save you plenty of $ over a SCSI setup. In terms of specific drives: I am presently running an IBM Deskstart XGP 45MB (7200 rpm/ATA-100) and a WD Caviar 40GB 7200 rpm ATA-100 in my main machine. I got the IBM for about $140 online (mwave.com) and have had no problems with it, but would probably not recommend it as a first choice right now due to the presence of drives of this model that can suddenly trash all your data (a lot of postings on line about this a few months ago). Mine works fine, but maybe I've been lucky so far. The WD (which houses my /home partition) has also been without problems, and both drives installed without any special effort under Linux. Actually, I never got around to switching them to ata-100, so they both have been running ata-66. Speed is excellent even without ata-100. Other drivers to consider are the WD Caviar 80GB 7200/ata-100, which seems to go through various specials from time to time - last week it was at OfficeMax for $199 with an $80 rebate. Sadly, I didn't get there until the end of the week and all were sold out. Another drive, which I considering for my house server (which I am just building) is the Maxtor Diamondmax series, which seems to be well liked at places like tomshardware.com. My best recommendation is to stick with a brand like the Maxtor Diamondmax or WD Caviar, and watch the flyers in the weekend papers from CompUSA, Best Buy, Office Max, Staples, etc. You are bound to find a drive with a significant discount or rebate if you can afford to wait a few weeks for one to come up. N
Re: Where can I get JetAdmin for Linux?
On Tuesday 22 January 2002 10:29, Brian P. Flaherty wrote: > "Stan Brown" <[EMAIL PROTECTED]> writes: > > I'm replacing a HP-UX workstation with a woody box. > > > > One of the functions of this HP box is to rotuinely check that status of > > a HP jetAdmin printer. > > > > Can I get JetAdmin software for Linux? If so has anyone got a pointer to > > a location? > > I don't know if you can get JetAdmin for Linux, but I suggest looking > at CUPS. I have JetAdmin at work on Solaris and CUPS at home and they > seem pretty similar to me. When you say check the status of an HP > printer, what do you mean? Is it something more than a cron job > grepping the result of 'lpstat -p printer'? > > Brian http://www.hp.com/cposupport/networking/software/hpwebjet_linux.selfx.html Good luck.
Woody / SMP - how to set up
I am wondering if a stock woody install will support a Pentium 4 system with Hyperthreading to appear as a two processor system. Will the standard woody 2.4.18 kernel do so, or is a recompile required? nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
How "unstable" is unstable?
I'm wondering, from those running sid, just how "unstable" is it at the present time? The reason I am asking is that I would like to move on to KDE 3 and am feeling behind the times, still using KDE 2.1 in woody. I've been reluctant to track sid since I do need my workstation to be up and working pretty well, so I'd be interested in hearing from some who are using unstable regularly. Thanks. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Suggestions on choosing an IMAP server
On Wed, 2003-07-16 at 19:07, Mike Fedyk wrote: > Hi D-U :) > > I'm going to switch from pop3 to imap, and I'd like to know what you guys > think is the best debian packaged IMAP server, and why. > > I'd really like to use MailDir mailboxes. > > Thanks. Although it seems that courier-imap is the most popular imap server for debian users, I have been using cyrus imap for about 1.5 yrs now and have been quite satisfied with it. The reason that I selected cyrus over courier was that it provided me with the ability to use fetchmail to retrieve my email from all of my pop3 mailboxes and forward the mail via lmtp into the imap store, because cyrus provides an lmtp daemon; this means that I don't have to run a full smtp daemon. I could not figure out whether courier does or does not include the capability to recieve incoming messages via lmtp. A drawback to cyrus is that, last time I checked, both stable and testing included only an older version of cyrus; the most recent (2.x) series is only in unstable. I d/l'd the latest sources and compiled them myself, which turns out to be somewhat difficult - there were a lot of glitches and gotchas which held me up for some time before I got everything sorted out. I did this for my home server; when I tried to install a later cyrus version in setting up my office's server (both, at the time, woody) I ran into a _different_ set of gotchas. Once I had them up and running, however, they have worked perfectly. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
automount permissions problem
Here's my problem, maybe someone can help. I have a usb media reader than handles smartmedia, compact flash, etc. It works fine in the sense of being able to put media into it and mount the media, using the usb-storage module, and copy files off the media. The devices look like scsi disk drives, as expected with the use of usb-storage. It would be ideal to be able to use autofs to mount these drives, as I do for my dvd and cdrw drives, so I added the appropriate entries into auto.mount, the control file for my /mount automount directory. It looks like this: # Automount map file # 6/4/02 nl # #format: # key [ -mount-options-separated-by-comma ] location dvd -fstype=iso9660,ro,nosuid,nodev,exec,user :/dev/dvd cdrw-fstype=iso9660,ro,nosuid,nodev,exec,user :/dev/cdrw floppy -fstype=vfat,nosuid,noauto,user,nodev,unhide:/dev/floppy cf -fstype=vfat,ro,nosuid,noauto,user :/dev/cf sm -fstype=vfat,ro,nosuid,noauto,user :/dev/sm /dev/cf and /dev/sm are symlinks to /dev/sda1 and /dev/sdb1. The media cards automount all right, but the protections on the directories created by autofs is rwxr--r--, which means I cannot cd into or list the files in subdirectories on the media unless I become root, which is not ideal. And that's the problem. I don't have this problem with automounted cd's or dvd's - the protection mask for those is rwxr-xr-x, so all works fine. Is there something I should have done in the auto.mount file? Is this something about the way the cf and sm cards work (they are formatted in a digital camera, so I don't really have any control over how various bits are set in the filesystem. Any thoughts/help appreciated. Thanks. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: automount permissions problem
On Sat, 2003-07-26 at 18:06, Shawn Lamson wrote: > On Sat, July 26 at 5:49 PM EDT > Neal Lippman <[EMAIL PROTECTED]> wrote: > > > > > I don't use automount but to get proper permissions for my vfat fs's i > use this type or entry in fstab > /dev/hdb1 /mnt/win98 vfatdefaults,gid=6,umask=002 > > then as long as I am in group 6 ( disk on my box ) I have read access... > The key is the umask portion and equates to permissions of 775 or > rwxrwxr-x in case you aren't familiar with umask. > Ah...that makes sense. I will review mount options; I should be able to include any options mount accepts in the auto.mount file...hopefully this will work. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Sandisk sddr-75 CF/SM usb reader problems
On Sun, 2003-07-27 at 16:38, Michael Waters wrote: > Hi list, > > I can't get this usb reader to work at all with a smartmedia card (i > don't have a compact flash card to test) although searching on the web > suggests that the sddr-75 is a standard USB Mass Storage device and > there are reports that it works in linux. > > I've followed the suggestions here: > http://vic.dyndns.org/linux-UsbMassStorage/ > http://www.qbik.ch/usb/devices/showdev.php?id=1227 > > I'm using the debian supplied kernel-image 2.4.21-3-k7 with sid as well > as a self-complied 2.4.21 and also tried versions of 2.4.20 & 2.4.19. > > To load the usb modules, in a boot-up script, I have: > > modprobe usbcore > modprobe usb-uhci > modprobe usb-storage > mount -t usbdevfs none /proc/bus/usb > > Instead of doing this, I've also tried using hotplug and usbmgr as well > as loading the modules in /etc/modules. > > > `mount -t vfat /dev/sd[a,b] /mnt/flash/` gives: > mount: /dev/sd[a,b] is not a valid block device > > I think the problem may be that sg is not mapping the reader slots to > /dev/sda and /dev/sdb. When I do `sg_map`, I get: (scd0 is a cdrw) > > /dev/sg0 /dev/scd0 > /dev/sg1 > /dev/sg2 > > Does anyone know how to get /dev/sg2 mapped to /dev/sdb? > > I bought the reader because I was having similar problems with my > camera. It's an Olympus D-510. I didn't use it for a while but I had > it working fine in linux last year with an earlier 2.4 kernel. > The output for sg_map indicates that the sg1 and sg2 device are not being detected as disk drivers, tape drives, or cdroms. I'm not sure why that is the case at this point. Two questions: 1. Are you certain that the USB ports on your computer work - since you indicate problems accessing your camera directly as a usb device as well. 2. It would be helpful to unplug the media reader, plug it back in, and see what messages, if any, the usb-storage module generates in the system message buffer (use dmesg to see this or cat /var/log/syslog). nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Sandisk sddr-75 CF/SM usb reader problems
OK, here's what I notice: in /dev, your /dev/sd* device entries are correctly set up with block major device number 8, as is the case for the first (I think) 33 scsi disk devices. The output of /proc/devices, however, does not show that you have a block major device 8 entry, which should show the "sd" driver. My system's output for the same two commands does indeed show block major 8 as a device in /proc/devices, with sd as the driver - so I would surmise that this is your problem and why the devices don't seem to exist. When you try to mount /dev/sda1, the kernel should be looking for driver registered to handled block major 8, but there isn't one. A harder question is WHY you don't have one. On my system, it appears that scsi disk support is built into the kernel - because I haven't loaded sd.o as a module and no such module appears in /lib/modules/2.4.18.../kernel/drivers/scsi. You need to verify that on your system, with whatever kernel you are running, that either a) scsi disk support is compiled into the kernel, or b) if it is built as a module, you are loading the module (eg by putting this entry into /etc/modules or by whatever other mechanism you prefer. While the kernel does dynamically locate and load some modules as needed based on system calls, I don't know if the scsi subsystem or mount subsystem does that for the scsi disk drivers; maybe someone on the list has delved into that aspect of the kernel code and can help here. My guess would be "no". Hope this helps. nl On Wed, 2003-07-30 at 02:56, Michael Waters wrote: > On Tue, Jul 29, 2003, Neal Lippman wrote: > > can you post the output of: > > > > 1) ls -l /dev/sd* > > > > and > > > > 2) cat /proc/devices > > Hi, thank you again for trying to help me. I appreciate it. Below is > the output. I apologize to the list for the size of this email but I > hope someone can see something amiss... > > # cat /proc/devices > Character devices: > 1 mem > 2 pty/m%d > 3 pty/s%d > 4 tts/%d > 5 cua/%d > 6 lp > 7 vcs > 10 misc > 14 sound > 29 fb > 128 ptm > 136 pts/%d > 162 raw > 180 usb > 226 drm > > Block devices: > 1 ramdisk > 3 ide0 > 22 ide1 > > > # ls -l /dev/sd* > brw-rw1 root disk 8, 0 Jul 27 02:51 /dev/sda > brw-rw1 root disk 8, 1 Jul 27 02:51 /dev/sda1 > brw-rw1 root disk 8, 10 Jul 27 02:51 /dev/sda10 > brw-rw1 root disk 8, 11 Jul 27 02:51 /dev/sda11 > brw-rw1 root disk 8, 12 Jul 27 02:51 /dev/sda12 > brw-rw1 root disk 8, 13 Jul 27 02:51 /dev/sda13 > brw-rw1 root disk 8, 14 Jul 27 02:51 /dev/sda14 > brw-rw1 root disk 8, 15 Jul 27 02:51 /dev/sda15 > brw-rw1 root disk 8, 2 Jul 27 02:51 /dev/sda2 > brw-rw1 root disk 8, 3 Jul 27 02:51 /dev/sda3 > brw-rw1 root disk 8, 4 Jul 27 02:51 /dev/sda4 > brw-rw1 root disk 8, 5 Jul 27 02:51 /dev/sda5 > brw-rw1 root disk 8, 6 Jul 27 02:51 /dev/sda6 > brw-rw1 root disk 8, 7 Jul 27 02:51 /dev/sda7 > brw-rw1 root disk 8, 8 Jul 27 02:51 /dev/sda8 > brw-rw1 root disk 8, 9 Jul 27 02:51 /dev/sda9 > brw-rw1 root disk 8, 16 Jul 27 02:51 /dev/sdb > brw-rw1 root disk 8, 17 Jul 27 02:51 /dev/sdb1 > brw-rw1 root disk 8, 26 Jul 27 02:51 /dev/sdb10 > brw-rw1 root disk 8, 27 Jul 27 02:51 /dev/sdb11 > brw-rw1 root disk 8, 28 Jul 27 02:51 /dev/sdb12 > brw-rw1 root disk 8, 29 Jul 27 02:51 /dev/sdb13 > brw-rw1 root disk 8, 30 Jul 27 02:51 /dev/sdb14 > brw-rw1 root disk 8, 31 Jul 27 02:51 /dev/sdb15 > brw-rw1 root disk 8, 18 Jul 27 02:51 /dev/sdb2 > brw-rw1 root disk 8, 19 Jul 27 02:51 /dev/sdb3 > brw-rw1 root disk 8, 20 Jul 27 02:51 /dev/sdb4 > brw-rw1 root disk 8, 21 Jul 27 02:51 /dev/sdb5 > brw-rw1 root disk 8, 22 Jul 27 02:51 /dev/sdb6 > brw-rw1 root disk 8, 23 Jul 27 02:51 /dev/sdb7 > brw-rw1 root disk 8, 24 Jul 27 02:51 /dev/sdb8 > brw-rw1 root disk 8, 25 Jul 27 02:51 /dev/sdb9 > brw-rw1 root disk 8, 32 Jul 27 02:51 /dev/sdc > brw-rw1 root disk 8, 33 Jul 27 02:51 /dev/sdc1 > brw-rw1 root disk 8, 42 Jul 27 02:51 /dev/sdc10 > brw-rw1 root disk 8, 43 Jul 27 02:51 /dev/sdc11 > brw-rw1 root disk 8,
Re: Sandisk sddr-75 CF/SM usb reader problems [solved]
Congrats. Glad to hear you are up and running! On Thu, 2003-07-31 at 03:43, Michael Waters wrote: > > it works!!! Thank you so much Neal, thank you, thank you, thank you. I > had given up hope. :) > > On Wed, Jul 30, 2003, Neal Lippman wrote: > > OK, here's what I notice: in /dev, your /dev/sd* device entries are > > correctly set up with block major device number 8, as is the case for > > the first (I think) 33 scsi disk devices. > > > > The output of /proc/devices, however, does not show that you have a > > block major device 8 entry, which should show the "sd" driver. > > > > My system's output for the same two commands does indeed show block > > major 8 as a device in /proc/devices, with sd as the driver - so I would > > surmise that this is your problem and why the devices don't seem to > > exist. When you try to mount /dev/sda1, the kernel should be looking > > for driver registered to handled block major 8, but there isn't one. > > > > A harder question is WHY you don't have one. On my system, it appears > > that scsi disk support is built into the kernel - because I haven't > > loaded sd.o as a module and no such module appears in > > /lib/modules/2.4.18.../kernel/drivers/scsi. > > I couldn't find an sd.o anywhere but browsing around in > /lib/modules/2.4.21.../kernel/drivers/scsi , I saw sd_mod.o . I tried > `modprobe sd_mod`, now I can use the reader. :) I've added sd_mod to > /etc/modules. I guess the kernel I was using last year when I had no > troubles with my camera must have had this compiled in; I don't > remember changing /etc/modules at all. > > > You need to verify that on your system, with whatever kernel you are > > running, that either a) scsi disk support is compiled into the kernel, > > or b) if it is built as a module, you are loading the module (eg by > > putting this entry into /etc/modules or by whatever other mechanism you > > prefer. > > > > While the kernel does dynamically locate and load some modules as needed > > based on system calls, I don't know if the scsi subsystem or mount > > subsystem does that for the scsi disk drivers; maybe someone on the list > > has delved into that aspect of the kernel code and can help here. My > > guess would be "no". > > Thank you again, it's really made my day to get this finally working. > > Michael > > > Just in case someone with similar trouble reads this in the future, the > necessary modules appear to be: > > scsi_mod > sd_mod > vfat > sg > usb-storage > usb-uhci (or I guess usb-ohci or usb_ehci depending on the motherboard) > usbcore > > the output of `mount` should show: > none on /proc/bus/usb type usbdevfs (rw) > > #sg_map: > /dev/sg0 /dev/scd0 > /dev/sg1 /dev/sda (this particular reader has 2 slots) > /dev/sg2 /dev/sdb > > # cat /proc/devices > Character devices: > 1 mem > 2 pty/m%d > 3 pty/s%d > 4 tts/%d > 5 cua/%d > 6 lp > 7 vcs > 10 misc > 14 sound > 21 sg > 29 fb > 128 ptm > 136 pts/%d > 162 raw > 180 usb > 226 drm > > Block devices: > 1 ramdisk > 3 ide0 > 8 sd > 11 sr > 22 ide1 > 65 sd > 66 sd > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Linux logical volume manager - a few questions
I am looking at installing the lvm layer on my file server, which is presently running woody. I have two 80GB hd's, one of which presently stores my /home partition (exported via both nfs and samba to the other systems on my home lan), and another which I just installed. (Actually, there's a third 80GB which stores /, /boot, /usr, /tmp, and /var, but I won't be including that in the lvm). Anyway, I figured the best approach was to turn both 80GB data drives into one 160gb logical volume. Since this is the main file store of everyone at home, stability is a very very high priority. My questions are: 1) Which lvm package to install? There are two obvious choices, lvm10 and lvm2. While lvm2 is the new rewrite, which is supposedly "stable", it apparently lacks some features and according to the debian.org description of the package is not yet ready for production use. So, I assume I am correct in going for lvm10 at the present time? 2) If I do go with lvm10, will upgrading to lvm2 once it is ready for production use just be a matter of apt-get install'ing lvm2 and removing lvm10, or are there incompatibilities in the on-disk structure that would mean starting over from scratch? That would be a major problem once I have stuff scattered across 160GB of logical space on two physical drives. 3) lvm10 recommends kernel version 2.4.20; I am running the standard 2.4.18 on the server. It is crucial to do this upgrade? (I suppose it wouldn't hurt since 2.4.20 contains the driver for my server's onboard gigabit ethernet chip, which I am not presently using as 2.4.18 did not support it, but still, I like to do as little as possible to the fileserver.) 4) Is anyone using lvm on their system who can comment on success, failure, pitfalls, etc? Thanks for any input. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Need help re LVM setup...
I am looking at installing the lvm layer on my file server, which is presently running woody. I have two 80GB hd's, one of which presently stores my /home partition (exported via both nfs and samba to the other systems on my home lan), and another which I just installed. (Actually, there's a third 80GB which stores /, /boot, /usr, /tmp, and /var, but I won't be including that in the lvm). Anyway, I figured the best approach was to turn both 80GB data drives into one 160gb logical volume. Since this is the main file store of everyone at home, stability is a very very high priority. My questions are: 1) Which lvm package to install? There are two obvious choices, lvm10 and lvm2. While lvm2 is the new rewrite, which is supposedly "stable", it apparently lacks some features and according to the debian.org description of the package is not yet ready for production use. So, I assume I am correct in going for lvm10 at the present time? 2) If I do go with lvm10, will upgrading to lvm2 once it is ready for production use just be a matter of apt-get install'ing lvm2 and removing lvm10, or are there incompatibilities in the on-disk structure that would mean starting over from scratch? That would be a major problem once I have stuff scattered across 160GB of logical space on two physical drives. 3) lvm10 recommends kernel version 2.4.20; I am running the standard 2.4.18 on the server. It is crucial to do this upgrade? (I suppose it wouldn't hurt since 2.4.20 contains the driver for my server's onboard gigabit ethernet chip, which I am not presently using as 2.4.18 did not support it, but still, I like to do as little as possible to the fileserver.) 4) Is anyone using lvm on their system who can comment on success, failure, pitfalls, etc? 5) I was debating using ReiserFS instead of ext2 as I have been - any thoughts on whether that would make sense? Thanks for any input. nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
What happened to Zinf in testing?
Can someone enlighten me on what happened to zinf in testing? I used to use freeamp. Around 6 months or so ago, freeamp seemed to disappear and was replaced in testing with zinf. Tonight I was looking to reorganize my mp3 and ogg files, and went looking to see what my program choices were, and foudn that zinf seems to no longer be available in testing (but is in unstable). Can anyone enlighten me? What are people using for mp3/ogg organization and playback under testing? (Ideally, I'd like something that even understood how to download into an mp3 player, but that might be asking for too much...but something that working like itunes would surely be a great thing) nl -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Moving /var to another drive
> George Karaolides <[EMAIL PROTECTED]> wrote: > > Hi Andrew, > > > > > > > > And transfer the data using tar: > > > > tar cplf - -C / var | tar xvf - -C /mnt > > I recently did something similar, moving my personal files to a new HD so that I can mount the entire drive (partitioned as one big drive) to /home/nl for my personal files. Interestingly, because I now have an entire ext2fs file system mounted as my home directory, I now have a lost+found directory in my home directory. Not, obviously, a big problem, just something to keep in mind. One thing I wondered about, though. You are using the technique of taring up the entire directory tree and then untaring it into the target partition. Why not just use cp -a instead? N
Re: GnuCash vs MoneyDance vs Kapitol
On Saturday 16 March 2002 10:15, hanasaki wrote: > Yup a new one for $35 that is Linux only. Thoughts? > > www.thekompany.com Well, I decided to buy Kapitol about 1.5 years ago, when it was first announced. I did this because Quicken is basically the only application that I use at home that requires me to boot Windows98, and is hence the only reason I ever boot windows now. After paying (I think it was $25 for the downloadable version), I had a tremendously difficult time actually doing the download, because the links I had never seemed to work. On the good side, however, the email response from the company was prompt and helpful, and I eventually got it downloaded. Unfortunately, it would not install due to dependences (Mandrake 7.1 at the time) that I couldn't get fixed. I was told by the company that those dependencies would be eliminated by the time the final product was released, slated for September 2001, so I decided to wait. I have checked their website intermittently, and every time I look it seems that the final release has been delayed...last I looked was back in January and I think they said there would be a final release in February. I guess I'll look again now and see there is a final release, and I'll give it a try and let you know... N
Re: OT: Wireless NIC to NIC; need WAP?
On Saturday 16 March 2002 17:34, Kent West wrote: > I've got two Debian (Sid) boxes in different parts of the house. Can I > put in a wireless NIC into each one of them, and them talk to each > other, or must I have a Wireless Access Point as an intermediary? Wireless NIC's can, as far as I am aware, operate in two different modes, one of which allows them to speak directly to each other, and the other which requires a separate wireless access point. So, you should be able to do what you want to do. I don't know if something specific at the driver level is required for this, however. > > Also, as long as I'm on the subject. Typically I'd buy some name brand > I'm more familiar with, like D-Link or Netgear, but the Siemens box > specifically mentions that Linux is a compatible OS. For that reason, > I'd like to throw my money toward Siemens. Any technical reasons not to? Don't know the answer to this. I have been contemplating adding some wireless to my network, and have been looking at Linksys for a very practical reason - my LAN currently employs Linksys cable-ethernet router and two Linksys hubs, and I have a Linksys adapter in my Linux Desktop. My experience has been that mxiing vendors can sometimes cause the network to be unreliable, but I don't quite understand why that has occurred. I don't know details about the Siemens products, but will surf their site, because if they are supporting Linux drivers (esp if GPL'd) that would be a real plus. One thing to keep in mind: It is not at all uncommon to find that wireless hardware from one vendor will not work with wireless hardware from other vendors, so it makes sense to select a vendor for, at least, all of your wireless needs and stick with that vendor. This is esp. true if software encryption is employed. > > And one more: why can't I find a PCI wireless NIC, instead of a PCI > wireless NIC adapter plus a wireless PCMCIA NIC? To me, that seems > stupid. (But then, someone mentioned that the Apple wireless Airport > basestation does the same thing; puts a wireless PCMCIA nic in an adapter.) Linksys now makes a PCI NIC that is wireless in addition to their wireless PCMCIA card and their PCI adapter for the PCMCIA card. Check their website. I do not know if the wavelan driver used for the PCMCIA card works for the PCI card, however - it may not, since this seems to be a very new product. Good luck. N
Video card support
I would appreciate input as to what video card to get for a new system I am about to assemble. My requirements: i don't game, so superfast 3d support isn't needed, but I do watch an occasional movie, and want good support for 2d graphics. I don't want to break the bank on this purchase either, so pricing is an important factor as well. My present system has a GeForce 2MX, which I am happy enough with. Since that system is a PIII with the i815 chipset, I could (in a pinch) just pull the GeForce out and reuse it. The old system is going to become a fileserver, and so could make do with the i815's onboard video. My present monitor is a 19" CRT, but I am strongly considering a move to one, maybe two LCD panels - so a card that supports dual monitors would be great, but at the least whatever i use needs to work well with a second card if need be to add the second monitor. I will likely install woody on this system, but with components of sid as needed. Thanks for any help. n -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Woody install infinite loop
I haven't been able to find anything posted on this; wondered if anyone has heard about this. I am attempting a woody install, having d/l'd the disk images over the weekend. This is going onto a new system; because it may be relevant, here are the specs that I think are needed: MSI K7T-266A Pro2A MB with Athlon XP1700+; 512MB ram GeForce2MX video card The install proceeds as expected until I get to the point of specifying the system time. The HW clock is NOT set to GMT, so I select NO for that option. I give my locality as US; Eastern. I then create a root account/password and a user account for myself. After I do that, the install loops back to ask me again if the HW clock is set to GMT. From that point on, nothing I do breaks this loop, and all of the above steps just keep on cycling. I have tried the install twice, restarting from the very beginning and reformatting all drive partitions to ensure there isn't any residual data from a prior install that could mess things up; same result each time. Interestingly, the FIRST time that the question is asked about the HW clock and GMT, the correct time is displayed, but each time thereafter the displayed time is off by two hours. I verified on reboot that the BIOS retains the correct time. Any thoughts? Is this a known bug that I just cannot find? I thought about not creating a user account until later, but didn't see how that would be the problem...but I could try it if no one has any better ideas. Thanks. Neal Lippman -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Woody install infinite loop seems fixed...
For those who like me were getting stuck on the woody install infinite loop problem (at the point where you specify your timezone, then create a root password and user accounts): I d/l'd a new woody CD #1 iso (via jigdo) over the w/e, and just tried it and the bug is now fixed. I guess the borked package is now replaced with teh working version. I've gotten to the point of scanning CD's for packages. Since I was planning more than one install, I d/l'd all 8 CDs (yeah, I know, I should have gone for the network install, but I was clearly being brain-damaged), so now to burn the remaining CD's so I can load them up. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Interesting question on nfs permissions
I have come across an interesting ( :confused: ) and difficult problem with an nfs mount; maybe someone has an idea. THe basic situation is that I have one linux box (running MDK 8.2, but that's probably not relevant) that I have just set up as a file server, and a second linux box (running Debian Woody, but again prob. not relevant) that I will use as my workstation. The server exports its /home partition via both samba and nfs, so that my home directories are accessible across the network. To access one's files from the workstation, I use symlinks from /home. Thus, on the debian workstation, /home/nl is a symlink to /nfs/nl, and obviously the nfs share from the server is mounted onto /nfs. Here's the problem: I used mkisofs to make an iso file with some things I wanted to burn to a cd, and then issued the appropriate cdrecord command, which failed due to a permission error on the iso file to burn. The reason is, obviously, that cdrecord is setuid root, and so the attempt to open the iso file on the nfs share appeared to come from root, and with root_squash on the nfs export, it couldn't get at the file. My thoughts as solutions are: 1. disable root_squash for any workstation that has a cdburner and thus might be subject to this problem (not a great idea, although since it's my home LAN and only I have root passwords, I could do this). 2. make cdrecord NOT setuid root. Since I have the cd burner owned by root, group cdrom, and my user account has the cdrom group, I should be able to write to it, but I don't know if cdrecord needs setuid root for other purposes. 3. Keep a local tmp folder, and whenever I make an iso file, just do so to that tmp directory, and burn the cd from there - or even just get in the habit of piping the output of mkisofs onto cdrecord directly. Any thoughts or suggestions? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
