AC97 on ICH4
Hi all, I have a new MSI 845G Mobo which is based on the i845G chipset and has the ICH4 bridge on it. The built in LAN works fine with the standard eepro driver - there's no issue there, but I can't get the sound to work. From what I can find, it should work with either the kernel 810 ac97 driver or the alsa i8x0 ac97 driver. However, when I try and load these modules I get: Intel 810 + AC97 Audio, version 0.21, 05:20:10 Sep 7 2002 PCI: Found IRQ 5 for device 00:1f.5 PCI: Sharing IRQ 5 with 00:1f.3 PCI: Setting latency timer of device 00:1f.5 to 64 i810: Intel ICH4 found at IO 0xdc00 and 0xe000, IRQ 5 i810_audio: Audio Controller supports 6 channels. i810_audio: Primary codec not ready. /lib/modules/2.4.19/kernel/drivers/sound/i810_audio.o: init_module: No such device Hint: insmod errors can be caused by incorrect module parameters, including invalid IO or IRQ parameters /lib/modules/2.4.19/kernel/drivers/sound/i810_audio.o: insmod /lib/modules/2.4.19/kernel/drivers/sound/i810_audio.o failed /lib/modules/2.4.19/kernel/drivers/sound/i810_audio.o: insmod i810_audio failed Installation failed. (that's for the kernel module. The Alsa modules give similar errors). I do notice that when the kernel loads I get the following message: Real Time Clock Driver v1.10e Non-volatile memory driver v1.1 Uniform Multi-Platform E-IDE driver Revision: 6.31 ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx ICH4: IDE controller on PCI bus 00 dev f9 PCI: Device 00:1f.1 not available because of resource collisions ICH4: (ide_setup_pci_device:) Could not enable device. Now I know that is saying conflicts with 1f.1 where as the sound seems to start at 1f.5. I have tried most things in the BIOS and the only PCI card I now have in there is my SCSI card which I really can't remove! Does anyone have any pointers on this one at all? -- Matthew Sackman Nottingham England BOFH Excuse Board: Someone was smoking in the computer room and set off the halon systems. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
nvidia and console freeze
Hi all, I know this has popped up before, but I have found a solution that works for me (P4, i845G, nVidia Ti 4200). Turn off APM in the BIOS. That's it folks! Matthew -- Matthew Sackman Nottingham England BOFH Excuse Board: Someone was smoking in the computer room and set off the halon systems. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Linux Mail Client
> I build computers as a side line, and I usually install Windows 98. Have > you tried setting Windows up in an easy-to-use form which allows a > family to have an email address each from the same ISP and only > see/download their own emails, making sure the correct address is shown > in the 'from' header? Outlook allows the use of 'accounts' but it is > still a nightmare to set up. If any user downloads mail it ends up in > the Inbox (or other folder) visible to the user who instigated the > download. Maybe I have not understood correctly the complexities of > setting up Outlook . . . . I know this is off topic but... In Outlook can be made to do things properly - in a multi-profiled system, you have to alter the location of the local mail folders in every profile to point to the same place. Then, it doesn't matter who downloads the mail, so long as you have the same filters under each profile then the mail will be sorted appropriately. However, this does mean that everyone has access to everyone's mail which is not always desirable. Back to Linux... If (to the person who started this thread) does not want to mix his personal and professional emails then he should have two accounts on his Linux machine. Then, as the mail comes in and is handed over to the MTA, the MTA should distribute the mail correctly to the correct account and hence the correct mail box. This would also have the possible advantage of preventing the mixing up of professional and personal work files, as the user would have a separate account for each instant. If you then also modify the group ids for the mail boxes, and assign that group id to your two accounts, then you /should/ be able to persuade a mail client to read mail under both accounts, possibly at the same time (you'd obviously have to specify the location of the mail box manually). I hope this helps. Matthew
Re: Linux Mail Client
> Monday, August 21, 2000, 11:35:05 AM, Mark wrote: > > I am somewhat tempted to ask why if you want to keep two sets of mail > > separate sets of mail you find it imperative to handle them both with > > one instance of a program. > > Convenience. There is no good reason not to, really. Why should I have > to use two when one is perfectly acceptable? The distinction is that the > program itself is able to handle the split whereas most cannot. I would be rather surprised if the new KMail couldn't handle things as you want (though I haven't fully tried it), however, this will not help you if you do not use X. I think that it is slightly unreasonable to expect to be able to keep two email accounts separate on your local machine and yet demand to be able to access both through a single instance of your MUA. To me, it would seem far more logical to access both local mail boxes through, if not separate accounts, then through separate instances of your MUA. Matthew
Re: Linux Mail Client
> > Monday, August 21, 2000, 11:35:05 AM, Mark wrote: > > > I am somewhat tempted to ask why if you want to keep two sets of mail > > > separate sets of mail you find it imperative to handle them both with > > > one instance of a program. > > > > Convenience. There is no good reason not to, really. Why should I > have > > to use two when one is perfectly acceptable? The distinction is that the > > program itself is able to handle the split whereas most cannot. > > I would be rather surprised if the new KMail couldn't handle things as you > want (though I haven't fully tried it), however, this will not help you if > you do not use X. I think that it is slightly unreasonable to expect to be > able to keep two email accounts separate on your local machine and yet > demand to be able to access both through a single instance of your MUA. To > me, it would seem far more logical to access both local mail boxes through, > if not separate accounts, then through separate instances of your MUA. Sorry, I'm pretty certain I mean mail client instead of MUA in the above paragraph. Matthew
Re: Linux Mail Client (was: Re: Web browsers for Linux (was: Re: Netscape Bus Error))
> No, I mean exactly what an MUA says it is. Mutt is an MUA but, to me, it > is not a mail client. A mail client is able to transfer and manipulate the > required data without need of other programs. A constant example I give, > which is flawed as all are, is web browsing. A web browser is, for the most > part, an HTTP client. We have the HTTP server and the HTTP client talking to > one another directly. We don't have an HTTP transport agent to get the data > to the HTTP user agent. Again, example, it is flawed, but it gets the basic > point across. It may interest you to know that there are many different ways to skin a cat. Clearly none of the ways currently available suit you 100% (or even 60%). However, it may interest you to know that in general, the modularization and breakdown of processes into many separate methods is generally thought to be A Good Thing. It is because of this that we have (for example) Micro Kernels. You may be further interested to know that under RISC OS, the entire web-browsing mechanics are as broken down as email is under Linux - you literally do have to have around 3-4 different processes running, which all communicate with each other to get the job done. This level of modularization offers far more power and flexibility, as it becomes easier to implement new features and capabilities (as the amount of code that has to be re-implemented from application to application is greatly reduced). I am far happier using a console mode MUA under Linux than I am using Outlook Express because I have far more 'nitty-gritty' control over what is going on. I may remind you that Linux is first and foremost a server OS. It is also a programmer's OS. As such, people who are not prepared to while away hundreds of hours reading man pages and docs and do not have an almost fundamental understanding of the OS are not going to find Linux a rewarding experience. Therefore, the attitude is, and will remain to be for some time, 'if it doesn't do what you want, make it do it yourself'. Matthew
Fw: Modem Troubles
Just to add to the details, I've just noticed that when dialling, the noises are louder under windows than under Linux: dunno if this will make a difference though. Matthew > Hay all. > > I'm having much trouble with connections. > Using: Potato, USR 3Com 56K FaxModem (external) > > I've got it set up right - the port I/Os and the IRQs are > right (using ttyS0), the initialisation strings are correct, > I'm using hardware control, loggin on to freeserve is OK > (using PAP), the static DNS resolvers are correct, > and things behave as normal... > > ...for about 10 seconds. Then the transfer drops off and I > get time-outs. However, plog reports a connection, as > does ifconfig, and the modem Carrier Detect light is on. > I really don't know what is going wrong - > under Windows I can get a constant 4000 bps, and I get this > too under Linux, but then the transfer just seems to slow > and then stop and I really don't know why. > > This is the (censored) contents of the relevant files: > > /etc/chatscripts/provider > # ispauth PAP > # abortstring > ABORT BUSY ABORT 'NO CARRIER' ABORT VOICE ABORT 'NO DIALTONE' > ABORT 'NO DIAL TONE' ABORT 'NO ANSWER' > # modeminit > '' AT&F1 > OK-AT-OK AT&F&B1&A3E0Q0V1&C1&D2&P1S0=0 > OK-AT&F&B1&A3E0Q0V1&C1&D2&P1S0=0-OK ATS7=60S19=0M1&M4&K1&H1&R2&I0B0X4 > OK-ATS7=60S19=0M1&M4&K1&H1&R2&I0B0X4-OK AT > # ispnumber > OK-AT-OK ATDT08440402001 > # ispconnect > CONNECT \d\c > # prelogin > # ispname > # isppassword > # postlogin > # end of pppconfig stuff > > Yes, the mad initialisation strings really are correct > (nicked from the windows connection log). > > /etc/ppp/peers/provider > # > hide-password > noauth > connect "/usr/sbin/chat -v -f /etc/chatscripts/provider" > debug > /dev/ttyS0 > 115200 > defaultroute > crtscts > noipdefault > user [censored].fsnet.co.uk > remotename provider > ipparam provider > > I really hope that someone knows what on earth is going wrong > here as I've no idea right now! > > Thanks, > > Matthew > > > -- > Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] < /dev/null > >
fetchmail problems
Hay all again! I have a POP3 mail box, into which goes all the mail for the domain sackman.co.uk. Emails that are for me ([EMAIL PROTECTED]) I want to download using fetchmail (the mail box is being used as a multi-drop mailbox). The problem is that using fetchmail's envelope "Received" option does not work for mailing lists, as it is looking at the last Received field in the headers which is showing the address that the email was sent to (the mailing-list server that received the email and then duplicated it). By examining the headers, I notice that my POP3 mailbox server inserts another "Received" header at the start of the email, quoting my email address. I have read all the documentation I can find on fetchmail, but can not find anyway of getting fetchmail to look at the first Received header instead of the last (note: on most emails coming through mailing lists, there are 3 or 4 received headers). As a result, all email is being download and most of it is being forwarded to 'postmaster' (which does eventually get to me), which is not how it should happen. I would love to know of the details in the .fetchmailrc file of anyone who is successfully selecting emails for themselves that are part of a mailing list (or any other information that could help!). Yours, Matthew
Re: firewall (fwd)
On Sun, Oct 01, 2000 at 11:51:50AM -0700, George Bonser wrote: > > This isn't necessarily the case. It certainly appears to vary by > > region. They don't do it here (Denver, Colorado). Perhaps this is > > because DSL is so easily available :} > > One interesting thing that many providers are doing is not allowing any > VPN traffic. If you want to "telecommute" and work from home, your company > is going to have to buy you a commercial VPN capable account. The > reasoning from the ISP standpoint is that the pricing on home accounts is > very low. They are designed for private personal use. If you want to put > these accounts into commercial service (they view a company offering > employees a VPN connection into the company net for purposes of performing > work to be commercial use) then you are going to need to buy a commercial > account (or, rather, your EMPLOYER will need to purchase the account). > > Individual home internet accounts are a "loss leader" for most ISP's. They > don't make beans from them and make their real money offering services to > business. In that light, I really can't blame them. > Over here in the UK, the ISP is free and the telecommunications network is what charges. ISPs here make money from the cost of phone calls to them (flat rate options are only just becomming available) or from advertising (they do nasty things like try to set the homepage of your browser to their own site). Neither the communication people nor the ISPs want you to use up too much of their bandwidth and so ADSL or cable modems or ISDN in reality are not feasible options for us. Thus we're limited to 56K modems. Apart from echo requests, I don't think the ISPs do anything in terms of scanning here: we get lines dropped after 2 hours, but they never complain about running servers: I run web servers and games servers on my box, and have never had a complaint. Examining the syslog shows nothing weird either. > It is going to get much more difficult as time goes by to find a basic > home account that will let you do much more than act as a basic client. > As always, you get what you pay for. However, I do think it is a bit much for an ISP to scan their clients - it's using up bandwidth, and afterall, most windows users don't even realise that their NetBIOS port (139) is open, and in most cases their personal pooy web server is also running in the background. Are non windows users being penalised just because they are more likely to know what they are doing and to make use of these technologies? Are the ISPs really trying to lobby microsoft into having printer and network sharing turned off by default and to remove the pws? Matthew > > > -- > Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] < /dev/null > >
Re: Fetchmail & Sendmail
On Fri, Feb 16, 2001 at 10:46:10AM -0800, Rob Hudson wrote: > I'm using the fetchmail-procmail-sendmail combo to get mail off of a > mailserver, filter it, and read it locally. I'm using that same mail > server to pass sent mail to. > > In this case, I don't know if sendmail is overkill and if I should be > using something different. > > What I'd like to fix is this... > > When fetchmail gets mail, I think it passes it to sendmail which > checks to make sure the domain exists. AFAIK, it is actually trying to work out where to deliver email: it is using the DNS servers to try to work out the final destination for the email. This behaviour is likely to be because of the usage of sendmail as a mail gateway, though of course you yourself are not using it as such. All you need to do (I think) is to make sure that in your /etc/nsswitch.conf file you have the hosts line as dns first and then files. Hence: hosts: dns files networks: files Thus if your dns servers are down then your computer will fall back on the files, in which case all you probably need is an entry or two in /etc/hosts with a localhost entry and your hostname entry. Personally I use postfix but the configuration is very similar to sendmail. I could of course be very wrong here! Matthew > Today, my DNS servers listed > in /etc/resolv.conf were down, and all the mail I fetched got bounced > and lost. I'd like for this not to happen and get any and all mail > that is in my box no matter if it's spam and has an unresolvable > domain or not. How can I set this up? > > Thanks, > Rob > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > > >
boot-floppies mirroring problems
Dear all. I'm trying to get the boot-floppies generation to work because I need to build a custom installation procedure for a machine I want to install debian on (the default kernels do not have drivers for the raid controller on this machine, and there is no ide hard disc, so a special kernel must be used to allow installation onto the raid array). I've installed the boot-floppies package, ncftpd, and ftpgrab. I've set ftpgrab up to copy ftp://ftp.debian.org/debian/dists/woody/main/binary-i386/base into /home/ftp/archive/debian I've been through the scripts in the boot-floppies package, and made a list of all the other packages that boot-floppies wants, and have setup ftpgrab to get all of those aswell (and dump them in /home/ftp/archive/debian). Thus everything is going into /home/ftp/archive/debian. Inside /usr/src/boot-floppies/config I have the following: ftp_archive := /home/ftp/archive/debian archive := $(ftp_archive)/ but this is not working as doing a make check seemingly lists every single package as being missing. What am I doing wrong here? Is there a file heirarchy that I must follow? Do packages need to be renamed? Can someone help please? Frustratedly, Matthew Sackman
apt-get, dpkg, sources lists and other exciting entities
Hi, About a week ago one of the reiserfs filing systems on my computer decided to go pop, and wiped off /var. No backups. So I was pretty much stumped and have spent the last week reinstalling from my potato disks. Last night I finally got my internet connection up and running properly, and as I'd managed to tar up and save (into a windows partition) a lot of other stuff (inc /home, /root, /etc), I wanted to get back up to potato (which I'd been previously tracking). So my /etc/apt/sources.list file looks like this: namkas:/home/matthew# cat /etc/apt/sources.list # See sources.list(5) for more information, especialy # Remember that you can only use http, ftp or file URIs # CDROMs are managed through the apt-cdrom tool. deb cdrom:[Debian GNU/Linux 2.2 _Potato_ - Unofficial i386 Binary-3 (2816)]/ unstable contrib main non-US/contrib non-US/main non-US/non-free non-free deb cdrom:[Debian GNU/Linux 2.2 _Potato_ - Unofficial i386 Binary-4 (2816)]/ unstable contrib main non-US/contrib non-US/main non-US/non-free non-free deb cdrom:[Debian GNU/Linux 2.2 _Potato_ - Unofficial i386 Binary-2 (2816)]/ unstable contrib main non-US/contrib non-US/main non-US/non-free non-free deb cdrom:[Debian GNU/Linux 2.2 _Potato_ - Unofficial i386 Binary-1 (2816)]/ unstable contrib main non-US/contrib non-US/main non-US/non-free non-free #deb http://http.us.debian.org/debian stable main contrib non-free #deb http://non-us.debian.org/debian-non-US stable/non-US main contrib non-free #deb http://security.debian.org stable/updates main contrib non-free # Uncomment if you want the apt-get source function to work #deb-src http://http.us.debian.org/debian stable main contrib non-free #deb-src http://non-us.debian.org/debian-non-US stable non-US deb ftp://non-us.debian.org/debian-non-US testing/non-US main contrib non-free deb ftp://ftp.uk.debian.org/debian testing contrib main non-free deb-src http://ftp.uk.debian.org/debian/testing main non-free contrib deb-src http://non-us.debian.org/debian-non-US testing/non-US main contrib non-free deb ftp://ftp.debian.org/debian testing contrib main non-free deb http://people.debian.org/~kitame/gnome/release ./ Now the really odd thing is that an awful lot of 'stuff' is missing: I did an apt-get update, and then an apt-get dist-upgrade and let it get on with stuff. But flicking through dselect reveals that the whole of kde is not there, reiserfsprogs, iptables, the perl cdk kit etc and I'm sure many more are not being listed. Is there something stupidly obvious here as to why they are not being listed? What do other peoples' sources.list files look like, and has anyone else come across this? I'm using mainly woody (?) with a 2.4.1 kernel on a PIII with 128Mb Ram. Many thanks for any help you could give me. Matthew
Re: Debian or Linux 7???
Hi, I think that the major security holes are actually in the programs that you run, and have very little to do with the kernel itself. Certainly I use a 2.4.1 kernel which gives me access to the wonderful new firewalling system, thus I simply block all incomming connections on the ppp0 interface. Thus no one else can even see that I'm online. Also, using programs like postfix instead of sendmail, or keeping a really tight dns setup (chroot it for example) will give you a bit more security. But the distributions only provide the tools: you have to work out how to use the tools to make a secure system. Matthew On Mon, Feb 19, 2001 at 06:11:55PM -0500, Steve Rudd wrote: > Hi! > > I am frustrated with the linux 2.2 kernel. I have had two hacks in 3 months > and I am going broke rebuilding my server. > > I went out and bought Redhat 7, and got hacked 6 weeks later. > > I have been placed in contact with a guy who wants me to use Debian. But if > it based upon the same kernel as redhat, how is it going to be more secure? > I checked and found that > > from (http://www.securityfocus.com/) > Security risks for years: 1997-2000 respectively: > Debian 3, 2, 32, 45, 12 > RedHat 6, 10, 49, 85, 20 > > So Debian is about twice as good as redhat, but that is not real reassuring. > > I am considering joining the debian family, but am a bit concerned about > security. > > Just how much more secure is Debian than redhat? > > Thanks! > > Steve Rudd > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > > >
Re: Which IMAP server to use?
Hi there, I use courier-imap which I find is fine: my big debian-user mail box often swells to >2000 messages and it takes about 12 seconds to fetch them all. I use postfix, delivering locally, the .forward file points to maildrop, which filters the emails into the correct mailboxes which I then access through mutt. It works pretty well! ;-) hope this helps Matthew On Fri, Feb 23, 2001 at 10:16:25PM -0500, Jimmy Kaplowitz wrote: > Could someone recommend which IMAP server I should use? The only one I've > tried, the one associated with Pine - or maybe it was the UW one (or are they > the same? maybe) was too slow to use, and every other one I've seen does not > support standard UNIX mbox format. That, however, is OK with me if I can keep > the notification that bash gives me when I have new mail. I am under the > impression that this feature of bash only works for mbox format. How can I > resolve this and set up IMAP while preserving that feature? > > Thanks in advance. > > - Jimmy Kaplowitz > [EMAIL PROTECTED] > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > > >
Re: Which IMAP server to use?
On Sat, Feb 24, 2001 at 11:29:08AM -0500, Jimmy Kaplowitz wrote: > Does bash (or whatever shell you use) notify you when you have new mail? This > is > important to me. If not, how can I enable or replace this functionality? It's > so convenient! No, not currently. I've no idea whether this can be achieved or not, though I suspect that with some reconfiguration and some scripts it probably can be done fairly simply. I'ts not something that I use regularly as I spend 99% of the time in X. Matthew > > - Jimmy Kaplowitz > [EMAIL PROTECTED] > > On Sat, Feb 24, 2001 at 08:22:49AM +, Matthew Sackman wrote: > > Hi there, > > > > I use courier-imap which I find is fine: my big debian-user mail box often > > swells to >2000 messages and it takes about 12 seconds to fetch them all. > > > > I use postfix, delivering locally, the .forward file points to maildrop, > > which > > filters the emails into the correct mailboxes which I then access through > > mutt. > > > > It works pretty well! ;-) > > > > hope this helps > > > > Matthew > > > > On Fri, Feb 23, 2001 at 10:16:25PM -0500, Jimmy Kaplowitz wrote: > > > Could someone recommend which IMAP server I should use? The only one I've > > > tried, the one associated with Pine - or maybe it was the UW one (or are > > > they > > > the same? maybe) was too slow to use, and every other one I've seen does > > > not > > > support standard UNIX mbox format. That, however, is OK with me if I can > > > keep > > > the notification that bash gives me when I have new mail. I am under the > > > impression that this feature of bash only works for mbox format. How can I > > > resolve this and set up IMAP while preserving that feature? > > > > > > Thanks in advance. > > > > > > - Jimmy Kaplowitz > > > [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] > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > > >
Re: MODEM LATENCY
I think you may be suffering from what I've come across lately: the modem gets stuffed up if the serial port baud rate is too high. Try setting the port speed (use pppconfig) down to 56700 instead of 115200, and I think you may find things start to work better for you. (worked here...) As for configuring your sources file, use apt-setup to set things up as you need. Matthew On Mon, Feb 26, 2001 at 12:26:31PM -0800, D. Hoyem wrote: > I hate to ask this cause it is really a DUHHH ... How > do I select the mirror that I want to go to? This is > part of the address in the sources file > htpp://security.debian.org/dists/potato/ binaryi386... > that is not the complete address, but how do I specify > the mirror that I want to use? > Don > --- Forrest English <[EMAIL PROTECTED]> wrote: > > apt-get install hdparm > > > > but, i kinda doubt that it is a latency problem on > > the hard drives. i'm > > guessing your connection to the mirror you have > > selected in your > > sources.list is just poor. try another mirror > > closer to you. > > > > just a suggestion, but the slowest drive i've ever > > seen was .89 megs a > > second. which is still more than enough to not time > > out waiting for data > > for a dsl line, much less a modem. > > > > -- > > Forrest English > > http://truffula.net > > > > "When we have nothing left to give > > There will be no reason for us to live > > But when we have nothing left to lose > > You will have nothing left to use" > > -Fugazi > > > > On Mon, 26 Feb 2001, D. Hoyem wrote: > > > > > I have been trying to use the apt-get update ; > > apt-get > > > install communicator to get Netscape downloaded > > from > > > the Debian non-free site that is in the sources. I > > can > > > connect and download but then I get timed out... I > > was > > > told on #Debian irc.debian.org that it might be a > > > latency problem...So I did a search on the wed and > > > found quite a bit of info and it all is saying to > > use > > > hdparm to check your hard drive. I have 2 hard > > drives > > > and Debian is on hdb2(swap) 156M, hdb3(/) 5.6G. I > > have > > > tried as root and the response is, command not > > found. > > > What am I doing wrong, and will this help resolve > > the > > > latency problem, do I have to be in the /sbin or > > > /usr/sbin to run this comand? After I find the > > best > > > settings when do I need to place them so they will > > be > > > picked up on start up? > > > TIA > > > Don > > > > > > __ > > > Do You Yahoo!? > > > Get email at your own domain with Yahoo! Mail. > > > http://personal.mail.yahoo.com/ > > > > > > > > > -- > > > 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] > > > > > __ > Do You Yahoo!? > Get email at your own domain with Yahoo! Mail. > http://personal.mail.yahoo.com/ > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > > >
Re: Printer recommendations
On Mon, Feb 26, 2001 at 07:56:33PM -0600, Bud Rogers wrote: > On Monday 26 February 2001 19:36, Sean wrote: > > I would second the choice for a laser printer ... I believe they are > > often of better construction than inkjets, and the cost per page is > > significatnly less. Not to mention the benefit of not having all your > > printed type simply disappear into a blurry mess if the paper happens > > to get wet. > > Same here. A couple of years ago we bought an HP 680C inkjet so my > wife to print color graphics. About a year ago I bought an HP 6L > laser for grad school research papers. This year my wife used her > Christmas money to buy another 6L and lugged the inkjet to the upstairs > closet. She decided on her own that crisp monochrome was better than > smudged color. And the cost per page is roughly an order of magnitude > better. Again seconded! I use a LaserJet 6P and have had no problems with it over a number of years which, considering the number of howtos and man pages I print out is very impressive. I use CUPS here, even though I'm not on a LAN. I find that the printing speed is lower than with lpr, but it is far easier to change settings and to monitor the queue etc (full web interface). I'm not sure how CUPS would work with JetDirect, but I would definately recommend investigating this route as I find the software to be rock-solid. Matthew
Re: IMPS/2 mouse protocol and XF86Setup
On Thu, 1 Mar 2001 17:29:10 +0100 Raffaele Sandrini > wrote: > Hi all, > > My problem is, that my mouse is working with gpm but not with X11. > i use the gpm flags: -m /dev/psaux -t imps2 (i have an IntelliMouse) > Thats working fine. Now im entering into XF86Setup to configure various > things... I chose at the mouse tab IMPS/2 and apply the changes. But the > mouse is just going into the right upper corner of the screen and stays > there. It does this eather if i choose /dev/gpmdata or /dev/psaux or if i > choose PS/2 as the protocol. > Has somebody an idea whats wrong? Yeah - this happens to me aswell. Use the keyboard commands to set up your X server mouse as you would expect. Save and then reboot. It manages to sort itself out when you reboot the machine - don't ask me why, but it works! Matthew > > cheers, > Raffaele > -- > Raffaele Sandrini <[EMAIL PROTECTED]> > Check out the most powerfull Linux desktop at www.kde.org !! > Check out the best Linux distribution at www.debian.org (www.linux.org) > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > > > -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing
Re: Reconstructing Var?
On Wed, 28 Feb 2001 19:34:55 -0900 Ethan Benson > wrote: > On Wed, Feb 28, 2001 at 09:05:19PM -0600, John Travis wrote: > > Is there an easy way to reconstruct the basic heirarchy and files for /var? > > I had a reiserfs meltdown after installing 2.4.2. I'm not really sure if it > > was that or the new reiserutils or a combination of both. reiserfsck > > managed to fix everything except for /var which was pretty well hosed. So I > > had to dpkg -X the older version of reiserutils and install it manually. > > Then a reiserfsck --rebuild-tree on var fixed the problem, but as expected > > nuked _everything_ on the partition. So is there anything easier than > > manually creating/touching files as needed for syslogd, dpkg, gdm, etc. > > etc.? > > if /var/lib/dpkg/* is gone and you have no backups your screwed. > reinstall your system from scratch. > Weird - this is exactely what happened to me two weeks ago - I was just doing some simple scans of the hard disc and reiserfs found some bugs. It then completely wiped /var. /var/lib/dpkg was screwed (along with everything else in /var) so I was reduced to reinstalling. Seeing as I track woody, and have potato disks and a 56k modem connection, it has taken some time to get back to normal. Sorry - no choice but to reinstall from scratch. -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing
Re: Optical mice
On Thu, 01 Mar 2001 21:22:52 + Lee Elliott > wrote: > Hello List, > > Would anyone care to recommend any optical wheel mice for use with > Debian - that is, optical mice with a wheel 3rd button? Does anyone know whether any company makes a cordless optical? I've been using my lovely Logitect Cordless Mouseman for some time now, but I really love the idea of the optical series. Does anyone know whether the two have been combined yet? >From my experience, the major difference between M$ mice and logitech mice is that the scroll wheel in the M$ mice is more stable - I find that the logitech's scroll wheel often jumps about which can be annoying when trying to open links in new windows in mozilla just my 2p (sorry - uk here!) Matthew -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing
Re[2]: Reconstructing Var?
On Thu, 1 Mar 2001 14:48:59 -0900 Ethan Benson > wrote: > On Thu, Mar 01, 2001 at 09:58:36PM +, Matthew Sackman wrote: > > > Weird - this is exactely what happened to me two weeks ago - I was just > > doing > > some simple scans of the hard disc and reiserfs found some bugs. It then > > completely wiped /var. /var/lib/dpkg was screwed (along with everything else > > in /var) so I was reduced to reinstalling. Seeing as I track woody, and have > > potato disks and a 56k modem connection, it has taken some time to get back > > to normal. > > > > Sorry - no choice but to reinstall from scratch. > > /usr/bin/yes make backups\!\!\! I'm a student = have no money for backup devices. Plus I'm still really pissed off that I bought a 10gig 7200 rpm seagate about 3 days before a big thread in here on how unreliable seagates are. Agh! Plus I have no IDE slots left so would have to buy a SCSI = very expensive. poo :-( Matthew -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing
Re: Reconstructing Var?
On Thu, Mar 01, 2001 at 10:35:23PM -0900, Ethan Benson wrote: > On Fri, Mar 02, 2001 at 07:29:21AM +0000, Matthew Sackman wrote: > > > > > > /usr/bin/yes make backups\!\!\! > > > > I'm a student = have no money for backup devices. Plus I'm still really > > pissed off that I bought a 10gig 7200 rpm seagate about 3 days before a big > > thread in here on how unreliable seagates are. Agh! > > > > Plus I have no IDE slots left so would have to buy a SCSI = very expensive. > > poo :-( > > well in this case you should split off your partitions, create > seperate / /tmp /usr /var and /home partitions. then backup /var (or > just /var/lib/dpkg) to /home/backup/ *usually* filesystem corruption > does not end up wrecking all your filesystems at the same time, so if > only /var gets trashed (in this case) you still have a backup in > /home/backup. > > if the disk dies your screwed, but this method protects you from > filesystem corruption fairly well. Well yes: currently I use eight partitions, including one /misc which I can use for backups of non-really-big stuff: [EMAIL PROTECTED]:~$ df Filesystem 1k-blocks Used Available Use% Mounted on /dev/hdb8 1542156326000 1216156 22% / /dev/hdb515522 2674 12047 19% /boot /dev/hdb7 1028092110340917752 11% /misc /dev/hdb9 4883556 1128692 3754864 24% /usr /dev/hdd5 10148529 96216 1% /tmp /dev/hdd6 512012116200395812 23% /root /dev/hdd7 717636486296231340 68% /var /dev/hdd8 1552236993120559116 64% /home When /var went down, I was able to tar and bzip2 up /home /etc /root and drop them into a windows partition which meant that I lost no real work - only work done on setting up debian was lost - no school work or such like. Unfortunately, I did have to let go to my mp3 collection: bzip2 couldn't handle a tar file that big! But yes, I suppose the only precaution I can take is to back up the very important stuff to /misc. I guess I really should write a script to do that and then fire it off from cron every couple of days Any ideas about an effective scipt to do this? Any recommended programs? Matthew PS: One other advantage of partitioning this heavily is that when I was writing some web perl scripts a couple of months ago and made a recursive error, /var rapidly filled as apache got screwed with firing out error messages. But the machine didn't crash: I suspect things would have got a lot nastier if I hadn't got /var on a seperate partition. -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing
Re: Reconstructing Var?
On Fri, Mar 02, 2001 at 09:52:27PM -0600, will trillich wrote: > On Fri, Mar 02, 2001 at 03:07:58PM -0900, Ethan Benson wrote: > > On Fri, Mar 02, 2001 at 06:52:17PM +, Matthew Sackman wrote: > > > > > > When /var went down, I was able to tar and bzip2 up /home /etc /root > > > and drop them into a windows partition which meant that I lost no real > > > > why did you do that? if you have seperate partitions there is no need > > to back them up, just don't initialize them in the debian installer. > > mount them instead. > > what ethan means, is, altho there's probably a good reason to > back up just about anything, you don't have to reset your > partitions when you reinstall debian, even from a cd-boot. like > he said, just MOUNT already-initialized (and filled with booty) > partitions and you're off to the races... Yes - I was aware of that, but I wanted to resize some of the partitions already, and as this computer has to be used by some on linux people and previously partitioning in cfdisk had caused some problems, I decided to start things off afresh and set things up under FDISK in DOS to keep that side of things happy. I did spend some time thinking out the best way of getting this done... > > backup whatever you want to back up. early. and often. take it > from the department of redundancy department. backups are a Good > Thing. > Yes. I do hear you. ;-) Matthew -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing
Re: [OT] OpenGL Experiance
On Sat, Mar 03, 2001 at 08:12:34PM -0500, Christopher W. Aiken wrote: > Anyone have OpenGL experience? I'm porting my companies > engineering software from HP/IBM/SGI UNIX systems to Linux. > Our graphics take advantage of OpenGL. From what I can see > OpenGL is a set of graphics libraries that need a special > graphics card to run. I found the OpenGL libraries (actually > Mesa libraries) but I need some recommendations for a graphics > card that is OpenGL capable and also Linux supported. > > I need to be able to tell our customers that we tested and > "certified" card "A", card "B", etc. that our software works > and works correctly with OpenGL. > > Recommendations please? Any special drivers for these cards > or is XFree86 all that is needed? It depends which cards you want to use. Matrox Cards have good easy driver support. 3DFX cards are also pretty well supported. NVidia chipsets are harder to set up - you have to delete afew soft-links and install some of their proprietry drivers and edit a few config files. But it can be done. If you want any hardware acceleration then you have to upgrade to XFree86 4.0 or higher. This is fairly painless, and if you use a 2.4 kernel aswell then you can turn on the direct rendering infrastructure and get a big performance boost. NVidia cards generally have the best OpenGL support in terms of speed and in terms of the implementation of the standard. However, their drivers are not Open Source and are slightly trickier to install than other manufacturers. I use an NVidia TNT2 M64 and am very happy with Quake III performance!! Things are pretty stable with a 2.4.2 kernel, XFree86 4.0.2 and their 0.95 drivers [patched for 2.4.x series kernels]. Their 0.96 drivers don't work for me. Hope this helps, Matthew -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing
Re: non-Debian package
On Sun, Mar 04, 2001 at 09:45:24AM -0500, Holp, John Mr. wrote: > Debian Warriors, > > Is there a general technique or utility that allows one to install a > non-Debina package on a Debian system? > > I am finding very sketchy information about a thing called > "alien" syntax something like; > > alien package_name.rpm > > Will apparently convert/create a Red Hat ".rpm" package to a Debian > ".deb" package?? > > I don't think apt or apt-get need or understand packages with the > .deb suffix. So is it possible to use apt or apt-get on a non-Debian > package to install it on a Debian system? > > man alien, whereis alien, whatis alien, etc., etc produce no > indication of such a thing on my Debian 2.2.17 machine. ...which seems to indicate that you do not have alien installed. Alien is a package in its own right and must be installed firt to work. So: apt-get install alien And then yes, you are correct, doing an alien package_name.rpm will *try* to convert the rpm into a deb. You can then do a dpkg -i package_name.deb and the package will be installed. The package database will be kept current. Hope this helps. Matthew -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing
Re: Need Window Manager recommendation
On Thu, Mar 08, 2001 at 01:21:39PM -0700, Eric Richardson wrote: > Eric Richardson wrote: > > > > Hi, > > I just recently installed Debian 2.2r0 and did a apt-get update apt-get > > dist-upgrade so I'm up to date on stable - I think. > > > > I loaded all the gnome stuff and the default display manager xdm and > > window manager twm. I have since replaced xdm with gdm and I would like > > to replace twm with one that is more fitting for use with gnome - > > besides I don't really like it. I'm interested in stability and I don't > > want it to take too much memory. > > If you want something that is really fast, takes up very little memory and is possible the most functional wm in existance then try ude and uwm. Once I got the hang of the title-bar-less nature of this wm I have never looked back. I know this sounds like a blatent plug, but it really is how I feel: I've looked at kwm, xdm, fvwm, bb, window-maker, sawmill, but none are as good in terms of functionality (IMO) as uwm. Try it. Matthew -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing
Re: Need Window Manager recommendation
On Fri, Mar 09, 2001 at 08:44:51AM +1100, John Griffiths wrote: > At 09:42 PM 3/8/2001 +0000, Matthew Sackman wrote: > >On Thu, Mar 08, 2001 at 01:21:39PM -0700, Eric Richardson wrote: > >> Eric Richardson wrote: > >> > > >> > Hi, > >> > I just recently installed Debian 2.2r0 and did a apt-get update apt-get > >> > dist-upgrade so I'm up to date on stable - I think. > >> > > >> > I loaded all the gnome stuff and the default display manager xdm and > >> > window manager twm. I have since replaced xdm with gdm and I would like > >> > to replace twm with one that is more fitting for use with gnome - > >> > besides I don't really like it. I'm interested in stability and I don't > >> > want it to take too much memory. > > I like ice Well quite: it is always a matter of personal choice: try as many as you can for a week or so and then work out which one works best for you. A month later you might decide you want a change and so try something else. There is never a 'right' answer... Matthew -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing
Re: IMAP and Exim
I run a similar set up here.
Basically, what you do is:
i) Install postfix and maildrop. Postfix will replace exim, but is probably
better anyway (more secure) and is pretty straight forward to configure
(default configuration worked here). Maildrop is a mail forwarding agent that
will understand mail-directories rather than the default unix mail box.
ii) Install courier imap. Getting this to work is not straight-forward: you
will need to install the PAM modules and libpam-pwdb and libdb2 and the utils
package. Get the courier imap server howto from your favourite howtos mirror
and read it! You will have to create some new accounts and use some utils to
create a userdb file. One bit that you may find doesn't work is that the howto
suggests using a full [EMAIL PROTECTED] in the entry in the userdb. I found that
this doesn't work - just use your username.
iii) Once you have a working imap server, with your selection of maildirectories
and sub-mail-directories (remember, when you use maildirmake, use the -f option
to specify multiple depth subfolders within your main mail dir - eg
maildirmake -f mywork.commercial /home/vmail/matthew_maildir to create what will
become imap://localhost/INBOX.mywork.commercial). Now create a ~/.forward file
which contains:
bash# cat /home/matthew/.forward
"| maildrop -d matthew"
(quote marks are needed). This causes postfix to pass all mail for you to
maildrop,
so create a ~/.mailfilter file with your required filtering options. Remember to
add a default catch-all filter at the end. Maildrop shoves the mail into the
mail
directories that courier allows you to access through imap. This sounds
conplicated
but works well and stabily in practice.
You may have to fiddle with the ownerships of the maildirectories - try making
yourself
the owner and vmail the group recursivly:
bash# chown -R username.vmail /home/vmail/username_mail_dir
My ~/.mailfilter looks like this:
bash# cat /home/matthew/.mailfilter
if(/.*boot.*lists\.debian.*/)
{
to /home/vmail/Maildir-Matthew/.linux.debian-boot
}
if(/.*lists\.debian.*/)
{
to /home/vmail/Maildir-Matthew/.linux.debian
}
if(/.*zap.*/)
{
to /home/vmail/Maildir-Matthew/.riscos.zap
}
if(/.*freshmeat.*/)
{
to /home/vmail/Maildir-Matthew/.linux.news
}
if(1)
{
to /home/vmail/Maildir-Matthew
}
Hope this helps.
Matthew
On Tue, Mar 13, 2001 at 10:57:19PM -0300, Carlos Laviola wrote:
>
> On 14-Mar-2001 John M. wrote:
> >
> > I want to set up an IMAP server, but I don't know what package to get. I
> > currently have Exim working great, it sends and receives mail, but I need
> > access to this mail from home and from several machines, so I don't want
> > to use POP, but IMAP seems to be just what I'm looking for. I'm using
> > woody and I looked at the available IMAP packages, but it just got me
> > confused. Could someone please tell me what package I can use with Exim
> > and how I would configure it.
>
> >From what I saw skimming thru apt-cache, you can install task-imap, or
> cyrius-imapd. For information on configuration and exim compatibility, refer
> to
> the programs' own documentation (sorry, I do not run either of them).
>
> >
> > Thanks.
> >
> >
> >
>
> --
> Carlos Laviola - ICQ 55799523
> pub 1024D/3516D372 2000-06-05 Carlos Laviola <[EMAIL PROTECTED]>
> Key fingerprint = 3BE1 6591 C78C 2AA4 31DD AEEF 6406 0227 3516 D372
>
--
Matthew Sackman
Nottingham,
ENGLAND
Using Debian/GNU Linux
Enjoying computing
Re: RDram
On Thu, Mar 15, 2001 at 02:48:44PM -0700, Robert Kerr wrote: > Hi all, > I have the unfortunate task of upgrading one of my users' systems to more > memory. She's using RDRam, and I need to buy (shudder) more of it for > her. I have the numbers off of her current chips, and need to match the > speed. Is there some guide somewhere as to what the numbers mean? > Thanks Not meaning to be rude, but does this have anything to do with debian? -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing
Re: NVdriver problems
I may be completely wrong here, but I was under the impression that the nvidia drivers will only work with XFree 4.01 or greater. And to get any of the real benefits out of XFree 4.01 (hardware 3d renedering etc) you need a 2.4.x kernel. Thus if I were you, I'd get hold of a 2.4.2 kernel, complile with support for AGP and DRI, upgrade XFree86 to the latest, and then try reinstalling the nvidia modules: they work great here. Matthew On Thu, Mar 15, 2001 at 10:39:21AM -0500, Kyle Girard wrote: > Alright I've been trying to get this kernal module to work for a couple > of days off and on without any success I have tried many different > ways with many different kernels and I get natta, zip, zilch. Any help > would be apprieciated... > > Here's my problem: > > I cannot insert the NVdriver module into my kernel I get the following > error: > > > insmod NVdriver > Using /lib/modules/2.2.19pre11/kernel/drivers/video/NVdriver > /lib/modules/2.2.19pre11/kernel/drivers/video/NVdriver: unresolved symbol > create_proc_entry_R0ca385ff > /lib/modules/2.2.19pre11/kernel/drivers/video/NVdriver: unresolved symbol > init_mm_R400c8da0 > /lib/modules/2.2.19pre11/kernel/drivers/video/NVdriver: unresolved symbol > remove_proc_entry_Rc8a0ae1b > /lib/modules/2.2.19pre11/kernel/drivers/video/NVdriver: unresolved symbol > proc_root_R26e7cf1e > /lib/modules/2.2.19pre11/kernel/drivers/video/NVdriver: unresolved symbol > mem_map_R3362b304 > /lib/modules/2.2.19pre11/kernel/drivers/video/NVdriver: unresolved symbol > __wake_up_R0e58d99a > /lib/modules/2.2.19pre11/kernel/drivers/video/NVdriver: unresolved symbol > printk_R1b7d4074 > /lib/modules/2.2.19pre11/kernel/drivers/video/NVdriver: unresolved symbol > register_chrdev_R0429f8ef > /lib/modules/2.2.19pre11/kernel/drivers/video/NVdriver: unresolved symbol > __pollwait_R2cf42003 > > > I've tried building the kernel-module from the tarball straight from > nvidia, using the nvidia-kernel-src_0.9.6-2_i386.deb with kernels 2.2.15 > -> 2.2.19pre11 and I get the same thing every time.. > > I've done everything I can think of, I followed the instructions in > /usr/share/doc/nvidia-kernel-src, I've tried following the instructions > from nvidia's page.. I'm not a guru when is comes to these kinda things > but I can usually plow my way through but this one has me > stumped... > > Thanks in advance, > > Kyle > > ps This might be all my own dum fault since I've never been able to > install a kernel module that I have built.... > > > System info > > Debian unstable > TNT (Diamond Viper V550) > kernel 2.2.19pre11 (was 2.2.15 until a couple of days ago) > Pentium 2 350 > > > > > > > -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing
Re: RDram
On Thu, Mar 15, 2001 at 03:33:16PM -0700, Robert Kerr wrote: > On Thu, 15 Mar 2001, Matthew Sackman wrote: > > > On Thu, Mar 15, 2001 at 02:48:44PM -0700, Robert Kerr wrote: > > > Hi all, > > > I have the unfortunate task of upgrading one of my users' systems to more > > > memory. She's using RDRam, and I need to buy (shudder) more of it for > > > her. I have the numbers off of her current chips, and need to match the > > > speed. Is there some guide somewhere as to what the numbers mean? > > > Thanks > > > > Not meaning to be rude, but does this have anything to do with debian? > > > > Not necessarily, so I suppose I should have put [OT] on the subject > line. The reason I ask here is because I feel a part of the community > fostered on this list, and it seems to be composed of a lot of very smart > individuals who have a wide and deep knowledge of lots of things. Just > like I'd ask the people I work with about things that don't necessarily > have to do with work because I value their opinion, I'd like to think that > I could ask my fellow debianites their opinion on things not strictly > related to Debian. If I have offended, I most humbly and abjectly beg > your pardon. Not at all have you offended: I just thought that it was obscure enough a question (RDRam is still pretty rare in my part of the world) for it to have been unlikely to be answered. Of course I'd be delighted to be proved wrong! :-) Matthew -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing
Re: changing from twm to ??????
On Fri, Mar 16, 2001 at 01:16:03AM -0800, Simon Harvey wrote: > hello : > does anybody know what file tells X11 what window manager to use, ive looked > around and i cant find it. > thanks > simon > ~/.xinitrc controls which window manager is loaded by X11 on startup -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing
Re: security line for source.lists
On Tue, Mar 20, 2001 at 10:18:24AM -0500, Hall Stevenson wrote: > > Do you know about a good site to security for my > > source.lists file? security.debian.org isn't working for me. > > This came up last week when there were problems accessing the site. They > are supposed to be fixed now... are you still unable to access it ?? > > Regardless, it was also mentioned that there is only *ONE* security > site. There are *NO* mirrors of it. At least that's what I think someone > said. Yes, that's right - there is only one security site as i) they are normally fairly small updates so do not consume vast amounts of bandwidth, ii) mirroring the security site is a bad idea because a) it slows the propogation of the updates and b) there is more risk of hacks which would rather negate the purpose of a security site! Matthew -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing
Re: Access & Permissions
On Tue, Mar 20, 2001 at 01:04:27PM -0500, C Mead wrote: > Hi, > > I need to give access to my /www/ dirs to regular users for apache. Can > someone > please take me through the steps of changing the permissions and groups > recursively. > > I would ideally like to make all regular users be able to access all > sub-dirs of dir /var/www and create their own files. As well as that I would > like to create a new group called 'dev' and make all regular users a part of > that as well so we can all at least -rx all the files we create. > > Can someone please walk me through these step it would be greatly > appreciated. Right, what I would recommend is that you create a new user called wwwroot or similar and then have your root website directory as /home/wwwroot. Then, all you need to do is a chmod g+wrx on /home/wwwroot and then for each user that you wish to be able to write to this directory simply do an adduser username wwwroot which will add that username to the wwwroot group and thus will allow them to write to the directory. You could continue using /www/ but it is not quite as neat as you will end up with an empty /home/www directory (yes, I know you can use flags to prevent the creation of a home dir, but it still isn't as straight forward). I believe that the chmod g+rwx with the 'x' is needed because I've found that otherwise some filemanagers will not display the contents of the directory (which is probably not desireable... ;-) Hope this helps, Matthew -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing
AMI RAID Express controllers
Hay guys, Does anyone have any experience with installing the Red Hat drivers for any of the AMI Express RAID controllers? I'm not necessarily bothered about being able to do a complete install from bare-metal straight onto the RAID array, more likely installing onto a spare HD, then installing the drivers and creating the array and then transfering everything across. Just wondered whether anyone had any comments regarding these cards/ drivers? With thanks, Matthew -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing
kernel 2.4.0 test 9 unresolved symbols
Dear all! I've downloaded the 2.4.0 test 9 sources, and have configured the compile options using make menuconfig, then a make-kpkg clean, followed by make-kpkg --revision=custom_0.1 kernel_image and then dpkg -i kernel-image... However, when dpkg is installing, I get an unresolved symbols on every module, and it don't work! I'm trying to upgrade from 2.4.0 test 5, so I don't need to change my /lib/modules/2.4.0-test5 as the new ones should go in /lib/modules/2.4.0-test9. I've also tried uninstalling the 2.4.0 test 5 kernel but that hasn't changed anything either. Does any one know what I've done wrong? Do I need to change anything in /etc/modules or similar files? I didn't get any of these problems when going from 2.2.17 to 2.4.0 test 5, so what extra do I need to do to go from 2.4.0 test 5 to 2.4.0 test 9? Any help would be VERY much appreciated! Yours, Matthew
NVidia kernel module won't compile
Dear all (again)! I'm using a 2.4.0 test 5 kernel, which is working fine, I've build and installed kernel headers, and set up /usr/src/linux to point to the headers. I'm trying to compile the NVidia kernel module so that I can make use of /dev/agpart and dri infrastructure, and the advanced NVidia 'stuff', but the compilation of the kernel module goes as follows: system:~/NVIDIA_kernel-0.9-5# make rm -f nv.o os-interface.o os-registry.o Module-linux NVdriver cc -c -Wall -Wunknown-pragmas -Wno-multichar -O -D__KERNEL__ -DMODULE -D_LOOSE_KERNEL_NAMES -DUNIX -DLINUX -DNV4_HW -DNTRM -DRM20 -D_X86_=1 -Di386=1 -D_GNU_SOURCE -DRM_HEAPMGR -D_LOOSE_KERNEL_NAMES-I. -I/usr/src/linux/include nv.c In file included from nv.c:52: nv.h:131: warning: #warning This driver is not officially supported on post-2.2 kernels nv.c: In function `nv_lock_pages': nv.c:551: warning: implicit declaration of function `virt_to_page' nv.c:551: warning: initialization makes pointer from integer without a cast nv.c:555: array subscript is not an integer nv.c:556: array subscript is not an integer nv.c: In function `nv_unlock_pages': nv.c:579: warning: initialization makes pointer from integer without a cast nv.c:583: array subscript is not an integer nv.c:589: array subscript is not an integer make: *** [nv.o] Error 1 I am doing this as root. I have the XFree 4.0.1 debs installed, and am trying to make use of the specific kernel modules supplied by NVidia. As always, any help is much appreciated. Yours, Matthew
Re: NVidia kernel module won't compile
Just to add a little more info, the kernel modules are NVidia_kernel 0.95-1 and also (but with fewer problems!), the corresponding GLX module. I am trying to compile both from .tar.gz. My vidio card is a NVidia TNT2 M64. Matthew On Tue, Oct 31, 2000 at 10:05:00PM +, Matthew Sackman wrote: > Dear all (again)! > > I'm using a 2.4.0 test 5 kernel, which is working fine, > I've build and installed kernel headers, and set up > /usr/src/linux to point to the headers. > > I'm trying to compile the NVidia kernel module so that I > can make use of /dev/agpart and dri infrastructure, and the > advanced NVidia 'stuff', but the compilation of the kernel > module goes as follows: > > system:~/NVIDIA_kernel-0.9-5# make > rm -f nv.o os-interface.o os-registry.o Module-linux NVdriver > cc -c -Wall -Wunknown-pragmas -Wno-multichar -O -D__KERNEL__ -DMODULE > -D_LOOSE_KERNEL_NAMES -DUNIX -DLINUX -DNV4_HW -DNTRM -DRM20 -D_X86_=1 > -Di386=1 -D_GNU_SOURCE -DRM_HEAPMGR -D_LOOSE_KERNEL_NAMES-I. > -I/usr/src/linux/include nv.c > In file included from nv.c:52: > nv.h:131: warning: #warning This driver is not officially supported on > post-2.2 kernels > nv.c: In function `nv_lock_pages': > nv.c:551: warning: implicit declaration of function `virt_to_page' > nv.c:551: warning: initialization makes pointer from integer without a cast > nv.c:555: array subscript is not an integer > nv.c:556: array subscript is not an integer > nv.c: In function `nv_unlock_pages': > nv.c:579: warning: initialization makes pointer from integer without a cast > nv.c:583: array subscript is not an integer > nv.c:589: array subscript is not an integer > make: *** [nv.o] Error 1 > > I am doing this as root. > > I have the XFree 4.0.1 debs installed, and am trying to make use > of the specific kernel modules supplied by NVidia. > > As always, any help is much appreciated. > > Yours, > > Matthew > > > -- > Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] < /dev/null > >
Compiling kernels
Dear all! I've possibly asked this one before but have not got any where, and am still puzzeled as to how to dort this one out! The problem is that my kernels refuse to install. I have downloaded the kernels from ftp.kernel.org, unpacked them. Configured them using menuconfig, and then cleaned using make-kpkg clean. I then compile using make-kpkg --revision-custom.1.0 kernel_image as per the instructions. Then I use dpkg -i kernel-imagedeb When installing, I get unresolved symbols on EVERY module that is to be installed. Upon looking inside the modules I can see nothing that references any files that don't exist and so am left wondering what on earth is going wrong. The last kernel that I was able to compile was 2.4.0-test5, which was available as a .deb in the woody dist, and what I am currently using. I'm therefore wondering whether there is some action I need to perform on the downloaded sources to prepare them for comilation using kpkg. I hope someone somewhere could help me out here. Matthew -- Using intelligent power: RISC OS, Be OS, Debian Linux Enjoying computing.
Re: Compiling kernels
On Monday 06 November 2000 20:25, David Z. Maze wrote: > Timo Benk <[EMAIL PROTECTED]> writes: > TB> Hi, > TB> On Mon, 6 Nov 2000, Matthew Sackman wrote: > MS> The problem is that my kernels refuse to install. I have > MS> downloaded the kernels from ftp.kernel.org, unpacked > MS> them. Configured them using menuconfig, > TB> If I understand you right you downloaded a tarball from kernel.org and > no TB> deb file. If so you have to do a: > TB> make dep > TB> mek modules > TB> make modules_install > TB> make bzImage > > It's far easier and cleaner to install the Debian kernel-package > package, untar the kernel source tarball, configure it with your > favorite variant on 'make config', and then run 'make-kpkg > buildpackage' to build Debian source, headers, documentation, and > kernel image packages from the source tree. Installing the image > package will prompt you to run lilo. If you decide you want a > new/different/better kernel, you can just install a different > package. If you decide you don't want the one you've installed, you > can remove it as you would any other Debian package. Thank you so much for this info: /usr/doc/kernel-package/README (and other files) give NO info regarding this 'make-kpkg buildpackage'. The man pages only briefly mention it. I'll let you know whether this works or not! With thanks, Matthew -- Using intelligent power: RISC OS, Be OS, Debian Linux Enjoying computing.
Re: Compiling kernels
On Monday 06 November 2000 22:57, Matthew Sackman wrote: > Thank you so much for this info: /usr/doc/kernel-package/README (and other > files) give NO info regarding this 'make-kpkg buildpackage'. The man pages > only briefly mention it. > > I'll let you know whether this works or not! Yeah - it has worked: using a make-kpkg buildpackage, then installing the new kernel source, kernel headers and finally kernel image still produced unresolved symbols, but the reboot (when I think I read a make depmod is run) rebooted properly and there were no kernel panics. However, there does seem to still be a problem with the module: running an 'update-modules' followed by 'modconf' lists NO modules being present. I assume therefore that the module system has altered slightly (though this worked under 2.4.0-test5). I am using the lastest (2.3.19-1) version of modutils, but this problem won't go away. However, the modules are still being loaded properly, but I simply have to edit the /etc/modules file by hand. Does anyone else have this problem? And if so, has anyone managed to resolve it? Oh, and of course, may I thank everyone who responded to my last problem and helped me fix it - thanks are indeed due. Matthew -- Using intelligent power: RISC OS, Be OS, Debian Linux Enjoying computing.
SSL in mozilla M18 and Netscape 6
-- Using intelligent power: RISC OS, Be OS, Debian Linux Enjoying computing.
SSL in mozilla M18 and Netscape 6
Dear all. Firstly apologies for posting an empty email... agh! - wrong key strokes! Without wishing to spark off another bitching match about netscape 6 vs mozilla M18, I would like some help regarding https. So far I've been using Netscape 4, and mainly without fault. I use squid on my box, and have netscape set up to use squid for all http. Under netscape 4, https works fine, bypassing squid. Whenever I try to use Netscape 6 or Mozilla M18, the https connection gets refused. Netscape 6 crashes, whilst Mozilla says `connection was refused when attempting to contact...'. By consulting the squid faq, I've tried getting the browser to stuff the https through squid, but that doesn't work either. I'm using woody, and a 2.4.0 test 10 kernel. I have squid 2.3.4-1 installed, libssl 0-95a-5 and openssl 0-95a-5. Does anyone know what's going on here? Matthew -- Using intelligent power: RISC OS, Be OS, Debian Linux Enjoying computing.
Xfree 4
dear all, this is very urgent as an upgrade is hanging in the balance here - a fast responce would be very much appreciated. Basically, where is /usr/X11R6/lib/X11/app-defaults supposed to point to in Xfree 4.0? Using woody with 2.4.0 test 10 kernel. Please could someone quickly let me know. Cheers, Matthew
Re: Xfree 4
Well I think there was a problem with the xlibs .deb, as I had the symlink set up correctly - /usr/X11R6/lib/X11/app-defaults was pointing to /etc/X11/app-defaults, but xlibs decided it was wrong and removed it. Then when it tried to update its settings it died due to a broken pipe. So I mkdir /usr/X11R6/lib/X11/app-defaults and cp -a /etc/X11/app-defaults/* back into it and then re ran the upgrade and it behaved, though told me that there should be a symlink pointing to /etc/X11/app-defaults. Deja-vu? Thanks for you help, and I hope this may be able to help others who come across this one! Matthew -- Using intelligent power: RISC OS, Be OS, Debian Linux Enjoying computing.
mySQL-server segfault
Dear all! I have been using Pronto for some time with the mySQL server running without problems, and providing very fast access to my many emails. However, on a recent apt-get dist-upgrade, the newly installed mysql-server refuses to install properly, and when trying to start the mysqld, I get a segmentation fault. The output in /var/log/mysql.err is: /usr/bin/mysql_install_db: line 371: 1737 Segmentation fault $execdir/mysqld $defaults --bootstrap --skip-grant-tables --basedir=/usr --datadir=$ldata "$@" <
Re: Stupid question
It generally doesn't matter: so long as you are able to dial up then the other processes take care of themselves. If you use pppconfig to set up the configuration, and remember to add your own account to the configuration then once you dial up, a mail send process will be run, and others: the details are in /etc/ppp/ip-up.d/ Check this for info. You should have one in here for exim (if you use it), and you can add anything you want: I've added a script to run fetchmail. As it is run as a root process, make sure that the copy of .fetchmailrc is in /home/root/ and is owned and grouped by root with mods of 0644. You might want to try changing the ownership of the files in /etc/ppp/ip-up.d/ - you could have a fetchmail script owned by each user on the system, then only their script would be run when they dial up (though I'm not sure if this will work - try it!) Hope this helps, Matthew Sackman On Mon, Dec 18, 2000 at 09:21:02PM +0100, Gary Jones wrote: > Okay, stupid question time. > > What is the best way of connecting to the 'net? I don't mean the > mechanicals, which connection type to use, that sort of thing, but > rather which account(s) should do so. Preferably I don't want to > connect as root, but some things (e.g. collecting mail or news) might > be better done as root or might /need/ to be done as root, or at > least some specific user with the right permissions which might be > different for the different tasks. What's the best thing to do? I've > never really seen a decent discussion about this, since I started > fiddling about with Linux (on and off, about 2 years). > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > >
Re: mysql-server broken
This is the quick fix: edit /etc/mysql/my.cnf remove the line about network listening and then resave the file. start up mysqld as normal - it should work! Apparently a fixed version will appear very shortly. Matthew On 16 Dec 2000 00:54:37 -0500, Jens Gecius said: > Hi folks! > > This time I really need help on short term. > > After my last dist-upgrade tonight mysql-server is broken. It always > gets an error in the postinst script. I tried to get it manually to > work, I tried to dpkg -i the old version (my current version is -8), > didn't work. > > I don't know what to do, I need the server and I'm about to leave for > a couple days, so, if anybody has an idea, please... > > -- > Tschoe,Get my gpg-public-key here > Jens http://gecius.de/gpg-key.txt > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > > > -- Using intelligent power: RISC OS, Be OS, Debian Linux Enjoying computing.
Re: CD Audio tracks/file system
Yep, that's exactely what I've done - both /dev/cdrom and /dev/hdc* are chown
root.cdrom
Then make sure that you are adduser cdrom, and everything works as
it should!
Matthew
On Tue, 26 Dec 2000 10:31:27 -0500, Scott Patterson said:
>
>
>
> >I had done that, but overlooked one small fact; I was in the cdrom group,
> >but not the disk group, to which /dev/hdc was pointed. I had no rights to
> >physically scan the CD for tracks!
> >
> >How has everyone else set up the permissions for their IDE /dev/cdrom links?
> >I don't like being in the disk group, I just hope it's usually automatically
> >set up the *proper* way and I just wrecked it myself...
>
> I'm guessing that ALL /dev/hd? are setup as group "disk". What I did and I
> believe is the *proper* procedure, is simply change the group of /dev/hdc (or
> whatever your CDROM is) from "disk" to "cdrom" ("chgrp cdrom /dev/hdc"). Of
> course, any users that need access should be in the "cdrom" group. Use
> "adduser
> username cdrom" to do that and have the user log out, then log in again for
> the
> changes to take place.
>
> Scott
>
>
>
>
> --
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
>
>
>
--
Using intelligent power:
RISC OS, Be OS, Debian Linux
Enjoying computing.
Re: can't play music cd
On Tue, 26 Dec 2000 00:03:15 -0900, Ethan Benson said: > On Tue, Dec 26, 2000 at 12:19:00AM -0800, Nate Amsden wrote: > > Ethan Benson wrote: > > > > > NEVER *EVER* add yourself to group disk. *NEVER* > > > > [EMAIL PROTECTED]:~$ id > > uid=1000(aphro) gid=1000(aphro) groups=1000(aphro), > > 0(root),4(adm),5(tty),6(disk), 20(dialout),24(cdrom), > > 25(floppy),29(audio),50(staff),60(games),45(usb) > > > > > > *cough* > > > > shoot me :P > > run cat /dev/urandom > /dev/hda > mke2fs /dev/hda > cat /dev/zero > /dev/hda > etc etc > > you might as well just login as root and not bother with a normal > account as your normal account is just as powerful as root is. Thanks for pointing out how dumb it is being a member of disk group. I was just removing myself ( ;-) ) and noticed that in addition to the group dip (needed to be able to use pon), there was a group dialout. What is this group doing there? and under which circumstances would one ever need to be a member of it? Matthew -- Using intelligent power: RISC OS, Be OS, Debian Linux Enjoying computing.
perlqt module (off topic)
I need to develop a new application and it's going to need a GUI. I was hoping to write it in Perl (cos it's easy!) and was hoping to use the PerlQt module, but there is no debian packaged version. I'm fine about installing it manually, I was just wondering whether anyone has had any experience with this module and would specifically recommend it or not or tell me to use the GTK module instead? Any opinions on this? Matthew -- Using intelligent power: RISC OS, Be OS, Debian Linux Enjoying computing.
Re: Switching between KDE & Helix-Gnome
> On Fri, 29 Dec 2000, Noah L. Meyerhans wrote: > > > You don't need KDM in order to start KDE, nor do you need GDM to start > > GNOME. Either of these display managers will allow you to choose a > > "session" from a menu when you log in. > > If I understand you right, it is probably best to do completely without a > display manager and instead use an ".xsession" file in my home directory > from which I start the corresponding session. What script do I have to > execute for starting Helix Gnome? > The way I have it set up is that my ~/.xsession has an "exec gnome-session" in it. I use kdm. On log in, if I leave the session type at default then I go into gnome, if I choose kde2 then I go into kde2. Easy. Matthew -- Using intelligent power: RISC OS, Be OS, Debian Linux Enjoying computing.
Bar code scanners
Dear all! I have been asked to start looking into providing an EPOS system for the shop that I work at. I am hoping to be able to develop all the software myself (using Perl and the GTK+ bindings). However, I have no idea about how well or otherwise bar code scanners work under Linux. Does anyone have any information as to which scanners are particularly good/bad and how they should be connected etc etc, and the same for cash-drawer triggers? Any info is very gratefully received. Matthew -- Using intelligent power: RISC OS, Be OS, Debian Linux Enjoying computing.
AGH! Hard disc faults!
Please help someone, and quickly!! I have found that on my root drive (which is /dev/hdb4), I have a read fault on block 131077. This is disrupting the boot sequence, and putting a whole load of file system errors on other file systems. I have run through with fsck and it reports a : Error reading block 131077 (Attempt to read block from filesystem resulted in short read) while doing inode scan. The boot up gets fairly screwed as it is having major problems in reading from the root file system. Fsck it seems can not fix it, and none of my linux books talk of any other disk repair utilities. Does anyone know what I should do? Please!? With thanks, Matthew -- Using intelligent power: RISC OS, Be OS, Debian Linux Enjoying computing.
Re: AGH! Hard disc faults!
I found that I had a spare 1Gb partition, and using a rescue disk managed to copy accross what I could of the damaged disk. Unfortuantely, that's meant I've lost some network option files and some of the run-levels directories have vanished (/etc/rc0.d, /etc/rc6.d/ went, as did /etc/network/options and some other files plus some i've no idea about!) so I'm trying to recreate those. Does anyone know of a utility that can reinstall all the now missing files - I'm not in anyway sure of what isn't there that should be. The hard disc is a 10Gb 7200rpm 2Mb cache Seagate that is less than 6 months old. I'm fairly furious that it has developed an error this quickly. With thanks, Matthew On Thu, 4 Jan 2001 18:01:11 -0800, kmself@ix.netcom.com said: > on Thu, Jan 04, 2001 at 09:56:43PM +, Matthew Sackman ([EMAIL PROTECTED]) > wrote: > > Please help someone, and quickly!! > > > > I have found that on my root drive (which is /dev/hdb4), I have a read > > fault on block 131077. This is disrupting the boot sequence, and > > putting a whole load of file system errors on other file systems. > > > > I have run through with fsck and it reports a : > > > > Error reading block 131077 (Attempt to read block from filesystem > > resulted in short read) while doing inode scan. > > > > The boot up gets fairly screwed as it is having major problems in > > reading from the root file system. Fsck it seems can not fix it, and > > none of my linux books talk of any other disk repair utilities. Does > > anyone know what I should do? > > If you don't have comprehensive system backups, this is the time time to > make them. > > I'd try running a (nondestructive) badblocks test on the partition -- > you'll likely have to boot another device, possible a floppy distro of > GNU/Linux. > > I tend to suspect bad hardware and toss it when I start getting disk > errors. You might try repairing the issue, but if it returns, your time > and data are more valuable than the disk. > > -- > Karsten M. Self http://kmself.home.netcom.com/ > Evangelist, Zelerate, Inc. http://www.zelerate.org >What part of "Gestalt" don't you understand? There is no K5 cabal > http://gestalt-system.sourceforge.net/http://www.kuro5hin.org > -- Using intelligent power: RISC OS, Be OS, Debian Linux Enjoying computing.
Re: GNOME Mail Client
On Sat, 6 Jan 2001 13:16:09 -0500, David B. Harris said: > To quote Raghavendra Bhat <[EMAIL PROTECTED]>, > # Brad Burns posts: > # > # > Can anyone recommend a good GNOME Mail Client ? > # > > # > # U mean a mail client based on the GTK libs ? > # Plump for Sylpheed. It is good. > > I use Sylpheed too, and I like it. I find that Pronto is great - it can use MySQL for storing emails which makes it ideal for huge amounts of mail from losts of mailing lists. Searches are VERY fast! Matthew -- Using intelligent power: RISC OS, Be OS, Debian Linux Enjoying computing.
Re: Turning off services SOLVED
On Sat, 6 Jan 2001 13:49:46 -0800, Ross Boylan said:
> At various times I have wanted to turn off certain daemons without
> uninstalling their packages. I couldn't find any good way to do this,
> so I wrote a little script. I'm making it available under the GPL.
>
> I've since discovered that the some packages also create a bunch of
> crontab jobs, and they are still cluttering things up a bit.
>
> If there's a better way to do this, I'd love to hear it.
Try finding the line in the file /etc/inet.conf and comment it out for the
services that you don't want to run. You may have to reboot then, or try
restarting the internet superserver - try a "/etc/init.d/inetd restart". You
might have to kill a few other processes or just stop them using their scripts
in /etc/init.d/.
Hope this helps
>
> #! /usr/bin/python
> # switchDemon.py
> #
> # Usage: switchDemon.py (--on | --off)
> #
> # This script will scan, turn off, or turn on selected
> # demons in /etc/rc?.d/. It does so by renaming S* symbolic
> # links so the services won't start.
> # Scan mode (neither --on nor --off) simply reports the status
> # of the demons.
> # Note that demons are not actually started or stopped by this script;
> # it just controls what will happen at system startup.
> #
> # (c) 2001 Ross Boylan [EMAIL PROTECTED]
> # Made available under the GPL (see www.gnu.org)
>
> import glob, os.path, re, sys
>
> class Demon:
> "Information and actions on a demon"
>
> gDisabled = "Disabled-"
>
> def __init__(self, name):
> "Name of demon"
> if not os.path.exists("/etc/init.d/"+name):
> raise name + " is not a known demon"
> self._name = name
> self._statuses = []
>
> def checkStatus(self):
> "Check status at various run levels"
> self.doOverRunLevels(self._checkStatus)
>
> def turnOff(self):
> "Make it so won't run at next system start"
> self.doOverRunLevels(self._turnOff)
> self.checkStatus()
>
> def turnOn(self):
> "We will start on next run level change"
> self.doOverRunLevels(self._turnOn)
> self.checkStatus()
>
> def doOverRunLevels(self, method):
> "Invoke method over all run levels"
> for rl in glob.glob("/etc/rc[0-9Ss].d"):
> method(rl, rl[-3])
>
> def _checkStatus(self, directory, runLevel):
> "See if we are around and fill in _statuses if we are"
> aRegex = re.compile("("+Demon.gDisabled+r")?S(\d\d)"+self._name);
> for file in os.listdir(directory):
> aMatch = aRegex.search(file)
> if aMatch:
> aStatus = DemonStatus(runLevel, aMatch.group(2), not
> aMatch.group(1))
> self._statuses.append(aStatus)
> return
>
> def _turnOff(self, directory, runLevel):
> "Disable demon for future reboots"
> aRegex = re.compile(r"S(\d\d)"+self._name)
> for file in os.listdir(directory):
> aMatch = aRegex.match(file) # must be at start
> if aMatch:
> os.rename(os.path.join(directory, file),
>os.path.join(directory, Demon.gDisabled+file))
> # no return values documented for os.rename
> return
>
> def _turnOn(self, directory, runLevel):
> "Enable demon for future reboots"
> aRegex = re.compile(Demon.gDisabled+r"S(\d\d)"+self._name)
> for file in os.listdir(directory):
> aMatch = aRegex.match(file) # must be at start
> if aMatch:
> os.rename(os.path.join(directory, file),
>os.path.join(directory,
> file[len(Demon.gDisabled):]))
> # no return values documented for os.rename
> return
>
> def reportTo(self, file):
> "Send human readable report to stream"
> file.write("Demon %s:\n"%self._name)
> for aStatus in self._statuses:
> file.write("\t%s\n"%str(aStatus))
>
>
> class DemonStatus:
> "Status of a given demon"
>
> def __init__(self, runlevel, priority, enabled=1):
> "Priority (nn) of demon and whether it is enabled"
> # runlevel is a single 0, 1, ...
> self._runlevel = runlevel
> self._priority = priority
> self._enabled = enabled
>
> def __str__(self):
> if self._enabled:
> msg = "Enabled"
> else:
> msg = "Turned off"
> return "%s for run level %s (priority %s)"%(msg, self._runlevel,
> self._priority)
>
> def isEnabled(self):
> if self._enabled:
> return "on"
> else:
> return "off"
>
>
> aCommand = sys.argv[1]
> if aCommand[0] == "-":
> if aCommand[-1] == "f":
> anAction = "turnOf
Re: hi all
In a stats colum in linux format this month, it said: 64,000 is the maximum number of users that could connect to a linux box before kernel version 2.4 With a 2.4 kernel, you can have 4.3bn connections. I'd image that the structures to handle this number of user ids must be implace otherwise this stat would be somewhat useless (if you believe stats anyway!) Matthew On Tue, 9 Jan 2001 12:39:34 GMT, Cliff Sarginson said: > Well the user id field in the password file is an unsigned int, > i.e. 32bits on a PC... > > So .. lots ! > > Course, it may be limited by other considerations..lol > > Cliff > > I think I remember reading somewhere that because users have 16bit uid's > > by default, the maximum is 64k or so (65,536). But you can have 32bit > > uid's if you're brave I also believe. > > > > Cheers, > > Corey J. Popelier > > http://members.dingoblue.net.au/~pancreas > > > > > > On Tue, 9 Jan 2001, Sathish C wrote: > > > > > Hi All > > > > > > can anybody tell me tha maximum number of users that can be created on > > > debian linux if there i sany such limit. > > > Please don't say that as many users as in any other linux, because I > > > don't know how many can be users created on other variants of linux > > > also. > > > > > > Please tell me how I can find such details. I searched in the man pages( > > > useradd) but I could not. > > > > > > Thanks in advance. > > > > > > > > > Bye > > > Kilaru > > > > > > > > > -- > > 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] > > > -- Using intelligent power: RISC OS, Be OS, Debian Linux Enjoying computing.
Re: hi all
In a stats colum in linux format this month, it said: 64,000 is the maximum number of users that could connect to a linux box before kernel version 2.4 With a 2.4 kernel, you can have 4.3bn connections. I'd image that the structures to handle this number of user ids must be implace otherwise this stat would be somewhat useless (if you believe stats anyway!) Matthew On Tue, 9 Jan 2001 12:39:34 GMT, Cliff Sarginson said: > Well the user id field in the password file is an unsigned int, > i.e. 32bits on a PC... > > So .. lots ! > > Course, it may be limited by other considerations..lol > > Cliff > > I think I remember reading somewhere that because users have 16bit uid's > > by default, the maximum is 64k or so (65,536). But you can have 32bit > > uid's if you're brave I also believe. > > > > Cheers, > > Corey J. Popelier > > http://members.dingoblue.net.au/~pancreas > > > > > > On Tue, 9 Jan 2001, Sathish C wrote: > > > > > Hi All > > > > > > can anybody tell me tha maximum number of users that can be created on > > > debian linux if there i sany such limit. > > > Please don't say that as many users as in any other linux, because I > > > don't know how many can be users created on other variants of linux > > > also. > > > > > > Please tell me how I can find such details. I searched in the man pages( > > > useradd) but I could not. > > > > > > Thanks in advance. > > > > > > > > > Bye > > > Kilaru > > > > > > > > > -- > > 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] > > > -- Using intelligent power: RISC OS, Be OS, Debian Linux Enjoying computing.
Re: X Window Manager
I've just install (yesterday) uwm and found that it created a softlink from
/usr/doc/uwm to /usr/share/doc/udm, so the documentation is easily findable on
my system.
Anyway, I'm really glad I read this list otherwise I'd have never come across
uwm. It really is excellant. If you're bored by the similarities between
Windows and some of the other window managers, and are finding your system is a
little slow and you are changing themes every other day because you basically
don't like them, then give uwm a go - it takes practice to get used to it, but
it really is wonderful.
Thanks to people for mentioning it! One happy user here!
Matthew
On Wed, 10 Jan 2001 09:22:32 -0600, Lance Simmons said:
> There's good documentation for uwm in /usr/doc/ude/. It explains
> exactly how to use and modify uwm.
>
> The documentation is in the ude directory because uwm ("Unix Window
> Manager") is part of the planned ude ("Unix Desktop Environment")--so
> far as I know, it's the only part of ude that's currently usable.
>
> ude is hosted at sourceforge, and one of the message boards there says
> they're working on gnome compatability, so perhaps the plan has changed
> from having a complete desktop environment to having a _great_ window
> manager that can eventually be used with gnome.
>
> If documentation was the only thing keeping you from trying uwm, I
> hope you'll give it another try.
>
> Lance Simmons
>
> On Wed, Jan 10, 2001 at 07:28:39AM -0500, Carl Fink wrote:
> >
> > If anyone else is looking to try it, keep in mind that there's no man
> > page, no documention gets installed in /usr/doc, and no documentation
> > of the program at its home page, meaning you'll have to read the
> > source or something to figure out how it works.
>
>
> --
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
>
>
>
--
Using intelligent power:
RISC OS, Be OS, Debian Linux
Enjoying computing.
Re: video card suggestions
I've got an Nvidia TNT2 M64 in my machine and am more than happy with it: the 2D is not as good as Matrox cards, but the 3D is very good, so long as you are using a 2.4.0 kernel with the patched drivers and an X 4 server. You also have to do some manual removing / re linking of the various mesa libraries - it takes a bit of work to get it set up, but after that it runs great. Matthew On Sun, Jan 14, 2001 at 04:47:31PM -0800, Nicole Zimmerman wrote: > I am going to be purchasing a new system in the near future and am looking > for suggestions on video cards. This system will be with me for quite some > time, so I am looking for something that has at least decent support at > the moment but not necessarily rock-solid. > > The new system is going to be a duron750 or 800 with 128 or 256 MB RAM. I > will (of course) be running debian. > > I am looking for input on performance, installation, drivers, x servers, > anything, related to 3d cards out there today. Direct experience would > help. > > I am leaning toward a radeon, but am open to suggestions. I don't like the > idea of a binary kernel module, but if that's the best out there, it is a > possibility. > > -nicole > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] >
Re: IMAP MUA and filtering
Yes, Postfix does support .forward files, so this works well (I've just
installed maildrop and things seem to be working).
I had no idea that mutt supported IMAP folders, so this comes as a very
pleasent surprise. However, when changing folders, it is a little
cumbersome to have to type in {hostname}INBOX.folder_name each time -
I'd much prefer be presented with a list of folders. Is this possible to
do under mutt?
Thanks for all your help so far!
Matthew
On Wed, Jan 17, 2001 at 11:56:27AM -0500, D-Man wrote:
> On Wed, Jan 17, 2001 at 07:33:29PM +1100, Damon Muller wrote:
> [snip]
> | If postfix supports .forward files, you'll need something like
> |
> | | maildrop -d $USER
>
> "| maildrop -d $USER"
>
> You need the quotes, otherwise the mail disappears. I'm not sure but
> you may also need 'exec' (probably not, it probably is a bash thing to
> fork a separate shell to run in):
>
> "| exec maildrop -d $USER"
>
> |
> | in your .forward file (off the top of my head, could be wrong).
> |
> [snip]
> | In this case, you can't go past mutt. Nothing with a pretty gui (even
> | for windows) comes near it for functionality and spead. It has somewhat
> | of a steep learning curve, but it's worth it.
> |
>
> I agree here, but I didn't find learning mutt to be any trouble. I
> already knew (the basics of) elm so mutt was a breeze. (I even
> installed mutt on the NT box I use at work)
>
> -D
>
>
> --
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
>
>
Re: wish list -> My personal Debian User Manual Generator
I'd just like to pledge my support for this, and if you need any help then please do get in touch! On Mon, Jan 29, 2001 at 12:45:35AM +1100, hogan wrote: > > the boot floppy disk)). Most documents found do not suite debian to > > 100%, that would not be the case in that generator. > ... > > and as you said is you said, the scripts searching for valid information > > is pretty easy, but the structure of the database will have the most > > importance for the whole thing to work. > > Agreed! The database structure is the most important thing of all. technologies or operating systems etc. are introduced (for example, biometric > devices) one won't have to rewrite the system from scratch or hack it to bits. > I mean if the data is put into such a system correctly, the output document > format can be just about anything! :) Yes, the database structure must be very very extensible and must be able to cope with anything the future will hurl at it! > > If for example someone is trying to find a solution for working with Foobar > Corp's new Model X device under environment A.. and someone has a solution for > working with Foobar Corp's older Model T device under environment B - the > database should be able to try and find the closest similarities and show the > instructions for the Model T - flagging that it may not be the solution. That may well involve an almost electronic-level understanding of the peripheral in question, which is not always possible. I think that the greatest problem is that people coming from windows no almost nothing about the hardware they have - if their motherboard has a sound set built on to it then they will not know the chipset (I didn't for a long time ;-). This may become the hampering factor - if you have to specify everything (down to design of motherboard, IDE controllers etc) then it may become a little too complex. > > Also your point about "most documents found do not suite (sic) debian to 100%" > is true - however to discard this information is not good - having the > facility to describe it as "not 100% suitable for debian but worth reading" in > the event that someone hasn't yet written a debian solution would be great. > Indeed writing the system from the word go to be "Debian" specific could be > damaging to the system's popularity - if you could keep it general and > generate enough interest in it, you could have vendors entering information > directly into the system.. That's very true. But as different dists have things set up differently (eg FSH) it may be a little hard for them to enter info directly into the system. Plus, many of the other dists provide a great deal of documentation. Additionally, some of the maufacturers have very poor installation guides and the guides will need re-writing to make them useful in anyway (eg NVidia). This could be an excellant resource, but it needs a lot of sitting down, and thinking and designing to make it as good as it could be. > Anyway - just my ramblings :) To which I've added! Matthew
Re: reiserfs annoyances
Just to say that I've just switched most of my partitions over to reiserfs and am having exactely the same problems: the boot procedure halts on each boot-up because of the request for a file-system check on the root partition. What is going on here? I figured that perhaps there was a problem with the kernel mounting the partition itself on bootup, and then having to unmount and remount the partition as rw (instead of ro), and so changed the read-only option in /etc/lilo.conf. But to no change. Surely more than three people must have a reiserfs root partiton? Could people please post their experiences and extracts of /etc/fstab please? Matthew On Sun, Feb 04, 2001 at 01:10:15AM +1100, Damon Muller wrote: > Hi folks, > > I'm running a woody (as of 2-weeks ago) machine with a 2.4.1-ac2 kernel > on it. I've got reiserfs compiled in, and all of my file systems (except > boot) are now running on reiserfs. > > I'm running > > ii reiserfsprogs 3.x.0a-1 *PRE-RELEASE* Tools for ReiserFS > filesystems > > compiled from the sid sources. > > Overally, I'm very happy with reiserfs, but there are a few annoyances. > > The first is that, presumably resiserfsck doesn't understand some option > passed to it by /etc/init.d/checkroot.sh, as when it gets to that stage > it stops and waits until you confirm to do an fsck. Does anyone know a > way around this, short of commenting out the offending line in > checkroot.sh (as is doing this a Bad Thing?)? > > The other thing that may or may not be a problem is that every time the > system boots, it seems to replay the transaction logs of all of the > filesystems. This suggests to me that it's not umounting them correctly. > Or is this the expected behaviour of reiserfs? > > Another thing is it complains when you have the following in your > /etc/fstab: > > /dev/hda2 / reiserfs defaults,errors=remount-ro 0 1 > > If you take out the `errors=remount-ro' it seems to work fine. Again, > I'm not sure if removing this line is going to have disasterous > consequences, or if it's something that reiserfs doesn't need. > > If anyone has any suggestions I'd be very interested. > > cheers, > > damon > -- > Damon Muller | Did a large procession wave their torches > Criminologist/Linux Geek | As my head fell in the basket, > http://killfilter.com | And was everybody dancing on the casket... > PGP (GnuPG): A136E829 | - TBMG, "Dead" > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > > >
Re: reiserfs annoyances
Yes, this works - removing the pass and dump parameters for ReiserFs disks stops the forced checking of the partitions. However, having defaults,errors=remount-ro still causes an error: unrecognised mount options. So I guess the errors=remount-ro bit is being passed to reiserfs as I don't get this message when I remove that bit. Many thanks, Matthew On Sun, Feb 04, 2001 at 01:13:08PM -0800, Jim Nutt wrote: > Matthew Sackman writes: > > Just to say that I've just switched most of my partitions over to reiserfs > > and am having exactely the same problems: the boot procedure halts on each > > boot-up because of the request for a file-system check on the root > partition. > > I converted both my partitions to reiserfs (mainly because my laptop > doesn't always wake up correctly and I hated waiting for fsck after > hard reboots). I did a couple of things to get rid of the checks. > First off, I removed the dump and pass parameters for the reiser > partitions in fstab, ie., the entries for the reiser partitions look > like this: > > /dev/discs/disc0/part3/ reiserfsdefaults > /dev/discs/disc0/part3/mp3reiserfsdefaults > > (gee what do you think is on the second partition?) Anyway, I'm also > using devfs, so that's why the device entries look a bit funky. This > got rid of the checks for the at least the non-root partitions. I > thought I had to do something else for the root partition, but I can't > find the change now and can't remember what it was . > > I don't have any problems with the questions now. > > jim > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > > >
Re: reiserfs annoyances
On Sun, Feb 04, 2001 at 08:52:36PM -0500, Dan Christensen wrote: > Is there a way to convert an existing partition to reiserfs? Or does > one have to create a new reiserfs partition and copy stuff over? Is > there documentation outlining the procedure? > > Dan You have to have a spare partition and to be able to copy other partitions into it, reformat the original partition and then copy things back. This may mean that certain partitions can not be changed: if it's too big to be moved: you could try using tar and gzip or bzip2. Unless you have a spare partition then you may find yourself a bit stumped ;-) Matthew
Re: Old news : Opera Free 4 Linux
On Mon, Feb 05, 2001 at 06:12:19PM +, Colin Watson wrote: > Anthony Campbell <[EMAIL PROTECTED]> wrote: > >On 05 Feb 2001, Colin Watson wrote: > >> I've been using Konqueror at work for a week or so now after Netscape > >> somehow started crashing my machine hard. I must say that it's a very > >> nice browser, and much faster than Mozilla. You don't need to be running > >> KDE either, which is good for me as I'm a GNOME/sawfish type. > > > >Do you need gnome, or will it work with any window manager, e.g. icewm > >(in my case)? > > AFAIK it doesn't need any special window manager or desktop environment > support. I've sometimes used it at home as well, where I run sawfish > without most of the GNOME stuff. > It just requires the qt toolkit libs (as used by kde), apart from that it really doesn't care which wm you are using (works fine here on ude). Matthew > -- > Colin Watson [EMAIL PROTECTED] > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > > >
Re: Old news : Opera Free 4 Linux
On Mon, Feb 05, 2001 at 07:02:06AM +0100, John Travis wrote: > On Monday 05 February 2001 13:12, Kerstin Hoef-Emden wrote: > > Hi, > > > > On Mon, 5 Feb 2001, Colin Watson wrote: > > > (Where was the link?) > > > > > > IIRC, Opera is now free-as-in-beer, but it's still not free in the > > > DFSG sense, which is a great shame; it would be great if it could > > > go into Debian. > > > > > > Perhaps it could be distributed by Debian in non-free, though. I > > > might have a look at its licence. > > > > Will the sources be available? > > > > Regards, > > > > Kerstin > > > I don't see that happening unless Opera runs out of cash and goes > under. I believe they meant the same sense of "free" as the windows > version when it hit 5.01. It was no longer an evaluation period deal. > But the "free" version displayed adds (although not terribly > obtrusive). But I would *love* it if it did go open source. 5.02 for > windows is probably the fastest browser I have ever used (IE for the > mac is actually pretty nice too). The idea of blocking the adds is > interesting. Although they could (and maybe have) thrown in a check to > make sure they are being displayed properly in order to function. > It may be possible to use a special iptables firewall filter rule on the 2.4 kernels to try and block the browser fetching emails from wherever it fetches them from. One assumes that opera will continually try to fetch the ads from a specific server (which may or may not forward it to another ad server), so blocking that ip may well work. Just a thought. Matthew
Win4Lin 2.0
Hay! Has anyone had any experience installing win4lin 2.0 on a debian box? Does it work? Is it any good etc? Anything particular to look out for? How well do applications that are not those listed on the win4lin site work under win4lin? Any info most gratefully received. Matthew
Re: Win4Lin 2.0
Sorry - just found the last thread on this - ignore this thread please. Sorry, Matthew On Sat, Feb 10, 2001 at 06:10:01PM +, Matthew Sackman wrote: > Hay! > > Has anyone had any experience installing win4lin 2.0 on a debian box? > > Does it work? Is it any good etc? Anything particular to look out for? > How well do applications that are not those listed on the win4lin site > work under win4lin? > > Any info most gratefully received. > > Matthew > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > > >
RAID port on SCSI-chiped motherboards
Hay all! Basically, I'm in the process of setting up an entry-level server and need some info on the various raid options supported in Debian. I'm thinking about using a supermicro dual PIII 440GX board which has a built in SCSI chipset with support for a RAID port. The adaptec pages for the RAIDport ARO-1130U do not talk about linux support, though some of their other raid controller cards do. So, does anyone have any experience of setting these cards up, what performance is achieveable, kernel compilation options, configuration etc.? Any info gratefully received. With thanks, Matthew
client-server backup using NFS and tar
Hi all, I wish to set up a client to backup to a standard tape drive (/dev/st0) on the server. Using tar. I really don't want to have to resort to using a commercial product or something like amanda for simplicity reasons. Now, as I understand it, tar when used in this situation writes directly to the tape device - the tape is not 'mounted' as such as it has no filesystem. So does that mean I need to export the entire server /dev/ via NFS to allow the tar process on the client to write to the device? I would rather not as that would seem a little in secure. Can I copy the device node from dev to another directory and simply export that? Will the node still function? I notice that from the tar man page the -f option will write to a device on a remote system: -f, --file [HOSTNAME:]F use archive file or device F (default "-", meaning stdin/stdout) but how do you set the remote system up (in this case the server) to allow the tar process access to the device? Many thanks for your help, Matthew -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing It said 'Required Windows XP or better.' So I installed Linux. pgp6Up6aqgdqf.pgp Description: PGP signature
Re: client-server backup using NFS and tar
Hi all (again), I've discovered that I can use the option --rsh-command and specify /usr/bin/ssh and then, if you generate a key pair and set it up properly then you're not prompted for a password, the authentication happens and then the tarring happens over a secure channel. So this basically solves my problem, but does anyone have any comments on the security of this for backing up a server to another server: two servers, one is internet facing, the other serves as a database server and has a tape drive. A 100TX connection connects the two. Is the above considered secure enough to backup the internet facing machine via a 2nd NIC to the backup device on the database server? Your comments are much appreciated. Matthew On Thu, Jun 21, 2001 at 11:22:02AM +0100, wrote: > Hi all, > > I wish to set up a client to backup to a standard tape > drive (/dev/st0) on the server. Using tar. I really don't > want to have to resort to using a commercial product or > something like amanda for simplicity reasons. > > Now, as I understand it, tar when used in this situation > writes directly to the tape device - the tape is not > 'mounted' as such as it has no filesystem. So does that > mean I need to export the entire server /dev/ via NFS to > allow the tar process on the client to write to the device? > I would rather not as that would seem a little in secure. > Can I copy the device node from dev to another directory > and simply export that? Will the node still function? > > I notice that from the tar man page the -f option will > write to a device on a remote system: > > -f, --file [HOSTNAME:]F > use archive file or device F (default "-", meaning > stdin/stdout) > > but how do you set the remote system up (in this case the > server) to allow the tar process access to the device? > > Many thanks for your help, > > Matthew > > -- > > Matthew Sackman > Nottingham, > ENGLAND > > Using Debian/GNU Linux > Enjoying computing > > It said 'Required Windows XP or better.' > So I installed Linux. -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing It said 'Required Windows XP or better.' So I installed Linux. pgp9Rz9IKiuwQ.pgp Description: PGP signature
Re: mod_ssl vs. apache-ssl
On Tue, Jul 10, 2001 at 11:58:50AM -0700, der.hans wrote: > moin, moin, > > which should one use and why? > > At a consulting gig with another dist the web developers suggested mod_ssl > was what got approval from the apache group. > > Does anyone have any input as to which is better? Does debian prefer one or > the other? > > I need both ports 80 and 443. Initially the ssl is just to lock down a new > tree, but eventually the main web site should be available encrypted as well > as not encrypted. With something as important as apache, I'd *alway* suggest compiling it from source. Along with OpenSSL, and mod_ssl, then follow the instructions, and it should work fine. You might want to set up a Debian-style layout for installation to 'debian' places: simply add the following to the $apache_src/config.layout: prefix:/ exec_prefix: $prefix bindir:$prefix/usr/bin sbindir: $prefix/usr/sbin libexecdir:$prefix/usr/lib/apache/1.3.19 mandir:$prefix/usr/man sysconfdir:/etc/apache datadir: /usr/lib/apache/data iconsdir: $datadir/icons htdocsdir: $datadir/htdocs cgidir:$datadir/cgi-bin includedir:$prefix/include/apache localstatedir: /var runtimedir:$localstatedir/run logfiledir:$localstatedir/log/apache proxycachedir: $localstatedir/cache/apache Then when you run ./configure, use a --with-layout=Debian That's how I do it... Matthew -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing It said 'Required Windows XP or better.' So I installed Linux. pgpj5Qh3OBDw7.pgp Description: PGP signature
Re: Squid - a bit off-topic
On Tue, Jul 10, 2001 at 05:58:33PM +, Victor wrote: > Sorry for this question somewhat astray. > > I've convinced my Information Tech Dept. to adopt a linux proxy server > (instead of Win NT stuff!!) and they're thinking of Squid. > > Now the problem they are dealing with as newbies with linux is about > user authentication. > > In a nutshell, because only few of us (100 people out of 650) are > allowed to connect to the web, the present Win NT proxy server connect > to the PDC server where a group of users authorised to connect to the > Internet is defined. Of course the users don't need to provide their > account and passwords. > > They now wonder if this same group on the PDC server can be used with > Squid for the same purpose without dismantling that magnificent :-( NT > server or - which is worst - duplicate list of users? > > Any suggestion? Hmm, not sure. Squid supports PAM, so if you can set up PAM to look at the list of users using a database module (investigate the berkley style modules) then that should allow you to just get away with copying the list. If you use DNS for all your internal machines and DHCP then you should be able to set up a subdomain for the machines that can connect to the net and then authenticate on this. Or, if you use fixed IPs then you can just authenticate on that. HTH Matthew -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing It said 'Required Windows XP or better.' So I installed Linux. pgpkcgIXLRCjU.pgp Description: PGP signature
Re: Reiserfs and disk spindown: separate /var partition?
On Thu, Jul 05, 2001 at 02:24:56PM -0700, Craig Dickson wrote: > Christian Jaeger wrote: > > > - Are there no problems making /var (and /var/cache) and /tmp > > symlinks to another partition? > > Don't make them symlinks; just make them mount points. I always set up > separate filesystems for /var and /tmp. > Yep - me too! > > - Getting it right for the boot process (should I make / reiser or > > ext2?): I assume I need to have /etc, /bin and /sbin on the root > > partition. What about /usr? > > It's never occurred to me to put /etc, /bin, or /sbin on any filesystem > other than the root. I would guess that that would not be a good idea, > since the boot process might (?) need access to them before it gets > around to mounting all the default filesystems (for one thing, the mount > command itself lives in /bin), but I'm not really sure. In the interests of maximising hard-disc usage, I once moved the contents of /lib to another partition (non-root) and created a sym-link. An hour later I had repaired the damage - it really don't like it!!! :-( Matthew -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing It said 'Required Windows XP or better.' So I installed Linux. pgpbTyYgFDQ6W.pgp Description: PGP signature
woody cd pseudo-images
Hi all, I'm trying to set up a pseudo image for woody on my localhost, and have downloaded the list for the unofficial images from ftp://ftp.fsn.hu/pub/CDROM-Images/debian-unofficial and I've got the pseudo package set up and read the docs. However, when trying to download the packages from ftp.uk.debian.org I've been hitting lots of errors - the packages arn't actually where they're supposed to be according to the image list. However, if I find the corresponding place in potato then they are in the right place. Example is pciutils which should be in /woody/binary-i386/main/admin From checking the packages.debian.org I've found the packages all under /debain/pool/... Thus what should I do? Does anyone know of a package mirror that has all the woody files in the correct place according to the unoffical lists? I've tried the sites that are listed in the email announcing the unofficial images, but even there they are not in the 'correct' place. Any help is much appreciated? Matthew -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing It said 'Required Windows XP or better.' So I installed Linux. pgpHabPnJrrT3.pgp Description: PGP signature
Re: woody cd pseudo-images
Hmm, I seem to have realised that the images are only available as images and not as a package tree. Is this correct? Matthew On Sat, Jul 14, 2001 at 03:37:00PM +0100, Matthew Sackman wrote: > Hi all, > > I'm trying to set up a pseudo image for woody on > my localhost, and have downloaded the list for the > unofficial images from > ftp://ftp.fsn.hu/pub/CDROM-Images/debian-unofficial > > and I've got the pseudo package set up and read the > docs. However, when trying to download the packages > from ftp.uk.debian.org I've been hitting lots of > errors - the packages arn't actually where they're > supposed to be according to the image list. However, > if I find the corresponding place in potato then > they are in the right place. Example is pciutils > which should be in /woody/binary-i386/main/admin > > From checking the packages.debian.org I've found > the packages all under /debain/pool/... > > Thus what should I do? Does anyone know of a > package mirror that has all the woody files in > the correct place according to the unoffical lists? > > I've tried the sites that are listed in the email > announcing the unofficial images, but even there > they are not in the 'correct' place. > > Any help is much appreciated? > > Matthew > > > -- > > Matthew Sackman > Nottingham, > ENGLAND > > Using Debian/GNU Linux > Enjoying computing > > It said 'Required Windows XP or better.' > So I installed Linux. -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing It said 'Required Windows XP or better.' So I installed Linux.
Re: FW: Careful. This is for information only.
Yes, but I think the point that Cringely was making is that MS could deliberately cause the internet to become crippled with DDoS attacks with the Windows XP. So if they've already built in a new version of TCP/IP without telling anyone then when the internet 'crashes', everyone using winXP could discover that they have access to a new internet. With 100 million users Cisco would have to be pretty dumb not to move quickly to make sure a non-crippled internet was available otherwise who'd get the blame? Surely not Microsoft? I hope it won't happen. Matthew Sackman On Thu, Aug 09, 2001 at 10:57:22AM +0100, P Kirk wrote: > Lets suppose that someone launches a proprietary version of the > Interent, called MSN or AOL for the sake of example. It will have to be paid > for. The free Internet will remain free. It cannot be taken away from > us. And even if the new proprietary version were subsidised for a few > years and thus got the majority of people used to it, when the price > gets passed on, people will return to the free Internet. > > That's why I think Cringely is a little wrong on this score. I remember > in the mid-90s the English football clubs tried to charge for access to > the "official team web sites." It failed because the fans own sites > were free, more passionate and better informed because they weren't > filtered by Pr/marketing types. Anyone chucking a few billion into > replacing the free Internet will have to have a very convincing case as > to how they can get their money back. I can't see what that case would > be. > > -- > > Patrick "sig free and joyful"" Kirk > > GSM: +44 7876 560 646 > ICQ: 42219699 > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > > -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing It said 'Required Windows XP or better.' So I installed Linux. - End forwarded message - -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing It said 'Required Windows XP or better.' So I installed Linux.
Re: File Manager for wmaker
On Fri, Aug 24, 2001 at 11:01:27AM -0700, Karsten M. Self wrote: > on Fri, Aug 24, 2001 at 08:28:33AM -0500, usucapiao ([EMAIL PROTECTED]) wrote: > > > Hi i would like to know any url of a good file manager for wmaker...i > > use gmc, but dont like it...if somebody can help me thanks!!! > > bash? ;-) > > There's a WINGS-based filemanager called FSViewer, if you want the > WinodwMaker look'n'feel. I usually just use the shell, sometimes mc. Yes, bash is a very good filemanager indeed - infact, rumour has it that it can do all kinds of other things too When I use a graphical filemanager, I use ROX - very fast and very straight- forward. http://rox.sourceforge.net It's based on the RISC OS filemanager. HTH Matthew -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing It said 'Required Windows XP or better.' So I installed Linux.
Re: ARRGH!!! Tulip card again
On Wed, Nov 14, 2001 at 02:01:41PM -0800, Jeffrey W. Baker wrote: > On Wed, 2001-11-14 at 13:39, [EMAIL PROTECTED] wrote: > > HELP! > > > > This is driving me absolutely positively mad. > > > > I have a Netgear F? 310TX network card, which is a PNIC 82c168 (tulip > > clone). > > > > I *cannot* get this card to run reliably on the latest Debian stable > > (2.2.Rev4). > > As you note, these are common cards, and the tulip is by far the best > widespread ethernet controller. So, why are you having problems with > yours? :) > > I'm using these cards in machines running 2.4 without problems. The > only thing that strikes me about your problem is that it seems related > to autonegotiation of speed and duplex settings. I presume that your > card is attached to an autonegotiating switch. If so, try disabling the > autonegotiation and lock the port at 100, full duplex. If the switch > cannot do that, lock your NIC to 100, full duplex. I think you can use > module parameters for this. > > I don't think there is any reason to go with old_tulip; the new driver > is fine. If it is possible, go with the latest kernel revision > (2.4.14). > > There is a linux-tulip mailing list if you get really frustrated. Kinda related... I had big problems with the natsemi driver until I compiled it as a module - it seems it's a lot more stable as a module rather than compiled into the kernel. I would also very much suggest using a new kernel. HTH, Matthew -- Matthew Sackman Nottingham, ENGLAND - The contents of this email are intended for the indicated recipient(s) only. This may or may not be indicated in the above email as it is enormously easy to fake email addresses (see the relevant RFCs). For security reasons this email is likely to be gnupg signed. On the other hand it may not be if I forgot to do so. In any case, if you are reading this on a Windows based computer then there was no point in me doing so (provided that I remembered) as your computer is most likely being used by yourself and 2.8 other people at the same time (normally without your consent). No responsibility will be accepted by anyone for any of the contents of this email. So tough. If in doubt, go compile Mozilla. pgpdD5TbvfcTi.pgp Description: PGP signature
[OT] Dual head console?
Hi People! I've got dual head working in XFree, but am now curious - if I don't start up XFree, is there any way I can run consoles on both monitors? inittab doesn't seem to help and I can't find any kernel options and I can't find anything on the web. Does anyone have any ideas or is this just not possible? Using 2.4.13 and woody. As always, any help is much appreciated. Matthew -- Matthew Sackman Nottingham, ENGLAND - The contents of this email are intended for the indicated recipient(s) only. This may or may not be indicated in the above email as it is enormously easy to fake email addresses (see the relevant RFCs). For security reasons this email is likely to be gnupg signed. On the other hand it may not be if I forgot to do so. In any case, if you are reading this on a Windows based computer then there was no point in me doing so (provided that I remembered) as your computer is most likely being used by yourself and 2.8 other people at the same time (normally without your consent). No responsibility will be accepted by anyone for any of the contents of this email. So tough. If in doubt, go compile Mozilla. pgpzCg1rfaAxO.pgp Description: PGP signature
Re: NIC Netgear FA311 : how do i make it work ?
On Sun, Dec 02, 2001 at 10:51:15AM +0100, spear wrote: > Hi there ! > > It seems that the drivers delivered with the mainboard are " > Slackware/Redhat " designed ... I found another one, a natsemi.c file, but i > never had to manage anything like this : how should i proceed ? What do i > need installed to use it ? > > Thanks in advance ! > > Mathias Ignore all the drivers on the disk. Recompile your kernel with the natsemi card driver set as a module. Then simply reboot, set up /etc/network/interfaces with something along the lines of: auto eth0 iface eth0 inet static address 192.168.1.2 network 192.168.1.0 netmask 255.255.255.0 broadcast 192.168.1.255 gateway 127.0.0.1 (obviously adjusting as necessary), and then run modconf and load the module. Note that gateway is critical as it determines the routing of packets going out of your machine. You may well need to initially use the route command to display the routing tables and/or modify them. HTH, Matthew -- Matthew Sackman Nottingham England BOFH Excuse Board: pgprD20cKFark.pgp Description: PGP signature
Re: nfs woes
y autodefaults,user,noauto0 0 > > --- Hmm, I think this is the major problem - I don't think that you're allowed to do such things in fstab because it has no reason to supply the -o remount option when trying to mount the fs the second time so the mount will fail. I would suspect so anyway - I may be wrong. I would therefore suggest that you create a script that is run once the system is up (use /etc/init.d/skeleton as a guide if you want) that i) mount -o remount,rsize=8192,wsize=8192 -t nfs 192.168.1.1/nfsroot/casper / ii) mounts the rest. Really this is still an ugly solution. A more elegant system would go: i) client logs in an grabs a filesystem image which is uncompressed into RAM, all over the 10MBit card. ii) client inflates image, mounts it and boots it (initrd style). iii) client drops 10MBit connection, starts up 100MBit connection, mounts nfs, moves ramdisk off. I would imagine that you have a lot of this already going. The problem is that your setup to too complicated IMHO, simply use the 10MBit card to boot, then drop 10mbps and start everything on the 100mbps. I'm assuming that you're using a boot chip on your 10mbps card to tell it where to get the kernel etc, in which case it should just be a matter of redesigning the initrd image. HTH, Matthew -- Matthew Sackman Nottingham England BOFH Excuse Board: Bogon emissions pgpJJoJydmE6n.pgp Description: PGP signature
Re: ISP asking about switching to Debian from OpenBSD
On Wed, Nov 21, 2001 at 09:08:25AM +1000, john wrote: > Fred Bloom wrote: > > > Cobalt is a linux ISP out of the box. > > > > J.H.M. Dassen (Ray) wrote: > > > > Many people regard Cobalt as a piece of junk aimed at wannabe ISPs. My > impression > is that the original poster is > working for an already established ISP. > > They are not secure out of the box, and can be a pain to modify to suit a > serious > environment. They also tend to crash a > lot unless you put a lot of time into them, based on my experience. > > My 2cents, which is of course really 1 cents at current exchange rates. I work for an ISP and I hate the damn things. They are really badly conceived things - no firewall tools enabled in the kernel at all, really bad implementation of certain cervices, and generally underpowered machines. If I was mad enough to actually own one of those things then I'd strip out cobalt's version of linux and hack debian until it worked. I would very seriously not recommend you touch these things for ISPs - they look easy, but so does MS Windows, and most of us know how difficult it is to fix that when something screws up... The fact that they come with no firewall tools and have telnetd enabled by default should be enough to scare off anyone with any sense! ;-) -- Matthew Sackman Nottingham England BOFH Excuse Board: Your/our computer(s) had suffered a memory leak, and we are waiting for them to be topped up.
OT: BOFH sig gen
I'm assuming that most people have come across/read the BOFH
stories...
...well I was bored and so put together a little script that
pulls an excuse off the board and puts it in the back of your
.signature file.
I realise this is off topic, just figured some people might
enjoy it! (obviously edit it as needs be).
#! /usr/bin/perl
my $ppp = `/bin/ping -qc1 google.com`;
if ($ppp =~ /1 packets received/)
{
my @text = `/usr/bin/telnet bofh.jive.org 666`;
my $bofh;
foreach (@text)
{
if ($_ =~ /Your excuse is\: (.*)/) {
$bofh = $1; }
}
my $sig = q/
Matthew Sackman
Nottingham
England
BOFH Excuse Board:
#bofh#
/;
$sig =~ s/#bofh#/$bofh/msg;
#print $sig, "\n";
open(FILE, ">/home/matthew/.signature") || die "Can not open file";
print FILE $sig;
close FILE;
}
========
enjoy,
Matthew
--
Matthew Sackman
Nottingham
England
BOFH Excuse Board:
Electrons on a bender
pgpZxJEJclsJ6.pgp
Description: PGP signature
Re: 100mbit nic: intel or 3com?
On Fri, Feb 01, 2002 at 05:40:52PM +0100, Victor Julien wrote: > Hi, > > I want to build a debian based router/gateway/fileserver/mailserver for a > home network with 12 clients. It will be quite low budget as the server is a > Pentium 166Mhz. I want the network to be 100mbit fullduplex, so I want to buy > a Nic for the server. Which one is best for maximum performance and > stability? Intel, 3com, SMC or just a cheap Realtek? I think the nic should > be using the cpu as little as possible... > > Thanks for your advice, > > Victor Julien Well I'm a poor student... so I started with Netgear FA311's which are terrible and gave me a *LOT* of agro -> steer well clear (hint: they drop packets). I've now got a pair of realtek's which are a damn site better, but I still have problems with enbd under high load, which I suspect are attributed to them. Plus when one machine is under high load, an ssh can take up to 15 seconds to connect. I don't know any of the more expensive ones - can't afford them myself. At work we use Intel and 3com nics in the big machines and realteks in the small machines. Switches are 3com, SMC and Intel. HTH, Matthew -- Matthew Sackman Nottingham England BOFH Excuse Board: positron router malfunction
Re: nVidia GeForce Ti
On Wed, Jan 30, 2002 at 12:00:55PM -0800, Didier Malenfant wrote:
> Hi,
>
> I'm also trying to get a Leadtek card with the GeForce3 Ti 500 to work
> under 2.2r4.
>
> I couldn't get it to work under XFree3.3 (would only display in 640x480
> 8bit) so I upgraded to XFree4.1.0 with Nvidia's own drivers.
>
> I can now get a decent resolution but X is acting very weird, the system
> locks up every now and then for 5-10 seconds (the mouse cursor
> dissapears when that happens) and then returns to normal.
>
> It's obviously not happy. :o)
>
> The machine is a dual proc running 2.4.17 but I removed smp support from
> the kernel to check if that was an issue with the video drivers.
>
> I know that won't help you much but maybe we can share experiences.
I've got a TNT2 M64 here. Using NVidia drivers and GL system and X4. No
problems. Have you tried compiling the DRI extensions in the kernel?
Also, check what your XF86Config-4 file is loading, should be something
like:
Section "Module"
Load"ddc"
Load"dbe"
Load"extmod"
Load"glx"
Load"pex5"
Load"record"
Load"xie"
Load"bitmap"
Load"freetype"
Load"speedo"
Load"type1"
Load"vbe"
Load"int10"
EndSection
Also check that your config file is loading the "nvidia" driver rather
than the "nv" driver.
HTH,
Matthew
--
Matthew Sackman
Nottingham
England
BOFH Excuse Board:
Too much radiation coming from the soil.
Re: 100mbit nic: intel or 3com?
On Fri, Feb 01, 2002 at 05:15:20PM -0600, Brian McGroarty wrote: > On Fri, Feb 01, 2002 at 10:48:21PM +0000, Matthew Sackman wrote: > > > > I've now got a pair of realtek's which are a damn site better, but I still > > have problems with enbd under high load, which I suspect are attributed to > > them. Plus when one machine is under high load, an ssh can take up to 15 > > seconds to connect. > > The realtek supposedly has a really wonky and expensive DMA scheme > that makes performance similar to a PIO card. Tulip-based cards are > very cheap as well, and may serve about as well as the 3Com and Intel > cards. Thanks for that, I didn't know that you could get cheap Tulips. Since some manufacturers arn't too upfront about chipsets, do you know any particular makes/models that are tulip chipsets? Matthew -- Matthew Sackman Nottingham England BOFH Excuse Board: Too much radiation coming from the soil.
Re: C Integrated Development Environment
On Thu, Jan 31, 2002 at 07:01:37AM -0500, Michael P. Soulier wrote: > On 31/01/02 Scott Henson did speaketh: > > > Sorry all. I guess i should mention he is using gnome. Its going to be > > a woody system. And before anyone suggests I switch him to KDE. I > > showed him both and he perfered gnome. It is ximian-gnome and he liked > > it alot. And he isnt interested in emacs or vim. I introduced him to > > both and he got lost.(GO EMACS) Thanks to all. > > It amazes me how so many people are so unwilling to invest time to learn > something worthwhile. I've taken the time to learn both Emacs and Vi/Vim, and > find both suitable in different situations. With some tools, you just don't > know you need them until you try them. I agree, and it's something I keep coming across. Personally, I believe that it is a deficiency of the society in which we live: there is never any time left to learn new things: people have to be able to pick up a tool and use it immediately. That is certainly the attitude of people that come across from windoze and try to use linux and can't even be bothered to read any books/man pages on apt-get. Seriously, almost every product that is on sale is designed to be used with next to no knowledge required: almost every car radio can be operated in the same manner, every TV or VCR conforms now to a standard usage pattern, even political parties... (whoops!). This is actually one of the reasons I love linux so much: there is so much wealth of documentation available, barely a day passes when you don't learn something new that can help you. The effort of actually learning something new is by far outweighted by the joy and satisfaction of discovering something new and exciting that makes your usage of your computer more productive. Just a few pennies. Matthew -- Matthew Sackman Nottingham England BOFH Excuse Board: Too much radiation coming from the soil.
Re: user homedir chroot jail..
On Sat, Feb 02, 2002 at 10:23:29AM -0500, Ben Collins wrote: > On Sat, Feb 02, 2002 at 04:37:01PM -0800, Petre Daniel wrote: > > how can i deny to a user with shell access the browsing of /home ? > > thanx > > chmod 711 /home > > So they can go into directories they know about, but they wont be able > to do "ls -l /home" Of course, if they know the usernames (which they will from cat /etc/passwd) then they can do a straight cd /home/username/ even if /home is chmod 711. Only by setting the user ~/ to 750 can you prevent them entering the directory. So yes, the best solution is chmod 711 /home; chmod 750 /home/* and then set the umask for all users to 027. HTH, Matthew -- Matthew Sackman Nottingham England BOFH Excuse Board: Interference between the keyboard and the chair.
Re: Ask Slashdot: How Well Does Windows Cluster? Turned around
Hi, linux-ha.org does very good stuff for high availability, though currently only monitors the status of nodes rather than the resources on the nodes. Red Hat is apparently working on their own version of it. IPVS is load balancing stuff. Basically it allows you to turn a machine into a load balancer via a range of options. There's a good HOWTO on this at http://www.linuxvirtualserver.org/Joseph.Mack/HOWTO/LVS-HOWTO.html whilst the main site is: www.linuxvirtualserver.org The type of load balancing you implement depends massively on the hardware and the network that you have. All the details are in the HOWTOs. If you want to do resource monitoring then you need to use something like mons and write your own scripts. Useful resources: http://www.linuxvirtualserver.org/Documents.html http://www.linux-ha.org/#Links Be wary about the HA howto which is from '97: much has changed since then. If in doubt, join either the HA mailing list or the LVSP mailing lists (or both). Neither are high volume. Have fun! Matthew On Thu, Feb 21, 2002 at 02:21:50PM -0600, Michael Jinks wrote: > well there's this... > > http://linux-ha.org/ > > ...though to be honest i haven't tried to do HA in quite a while, so > have no idea how up to date or workable any of it is. > > TurboLinux used to (does?) market a distro with some proprietary kernel > and userland enhancements for doing load balancing and failover. Two > jobs ago I got as far as having it set up and not-quite-working, then I > switched jobs. At that time (mid to late 2000) the GPL'ed HA stuff had > some fairly serious limitations not present in the TL package which is > why we coughed up the money, but with luck GPL-land has caught up some > since then. > > -mrj > > > > On Thu, Feb 21, 2002 at 08:09:23PM +, Andrew Pritchard wrote: > > For those of you who read Slashdot (http://slashdot.org), there was a > > recent > > question about Linux/MS clustering. The query was about a computational > > clustering and everyone was suggesting that Beowolf was good for that. The > > consensus seemed to be that MS Clustering was good for High Availability, > > not > > good for Computational Clustering (this was disputed, but I'm not going to > > go > > there :) > > > > However, no one turned the question round, and asked or suggested what you > > can > > do for High Availability/Load Balancing Linux. Is there a package (or set > > of > > packages) in debian for this? What are people's experiances with this kind > > of > > setup. Obviously Linux is good for High Availability, but the Load > > Balancing is > > another matter. > > > > Cheers, > > > > Andrew > > > > "I do not agree with what you say, > > but I will defend to the death your right to say it." > > Francois Marie Arouet Voltaire (1694-1778) > > > > > > -- > > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > > > > -- > ## Michael Jinks, IB ## JFI/MRSEC Computing ## University of Chicago ## > Reader! Think not that > technical information > ought not be called speech; -- Anonymous, "How to decrypt a DVD" > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > > > -- Matthew Sackman Nottingham England BOFH Excuse Board: Computers under water due to SYN flooding. pgpWwVpvHpaSX.pgp Description: PGP signature
Re: Diagnosing poor performance
On Thu, Feb 21, 2002 at 10:30:53PM +, Patrick Kirk wrote: > Thanks to all who pointed out ways forward. > > I have switched to twm which offers all the window management I need. > > It seems that I was simply loading the equivalent of three operating > systems by using kde and gnome apps interchangably on Enlightenment. > That's sad. > > Tried mc. I don't like it. Konqueror is a great file manager. GMC is > OK but it keeps asking how to open text files which seems an odd thing > to need to configure. Are there other light file managers? rox (rox.sourceforge.net): *The* fastest file manager. Also, for ages I've been using uwm/ude as the wm, but eventually got bored with it. I tried windowmaker for about a day and then came across xfce and xfwm (www.xfce.org). Try it: I'm sure you'll like it. Have fun. Matthew -- Matthew Sackman Nottingham England BOFH Excuse Board: Computers under water due to SYN flooding. pgpytSPyUD6jH.pgp Description: PGP signature
Re: Diagnosing poor performance
On Sat, Feb 23, 2002 at 09:10:05PM +, Patrick Kirk wrote: > On Sat, 2002-02-23 at 12:14, Matthew Sackman wrote: > > On Thu, Feb 21, 2002 at 10:30:53PM +, Patrick Kirk wrote: > [snip] > > with it. I tried windowmaker for about a day and then came across xfce and > > xfwm (www.xfce.org). Try it: I'm sure you'll like it. > > > > Have fun. > I'm trying it now. The overall impression is nice but there are some > real oddities. > > 1. I can't change the default buttons. Really? .xfce/xfce3rc allows you to set all the buttons you want. (remove the bak file when you're editing this, or edit it from the console). You can do it non-manually too: just right click on the icon. > 2. Menus are really difficult to edit and don't accept png images. If I > edit xfce3rc by hand, the sodding thing reverts to the default settings > next time I login. Just use the menu editor (from the menu). Also try enabling the menu modules as in the setup panel (Debian, Gnome, Kde). So what if it doesn't support pngs: the png library takes longer to load and is bigger than xpm. Just use GIMP to convert any icons you really want (not too big a job). > 3. There is no place to configure useful things like what Alt-Tab does. > Currently it shows ative windows on this desktop which really isn't much > use if the application I want is on another. Possibly not, but there is a window menu available (3rd button). Also, the debian docs for this arn't too hot. Download the src tgz from the xfce.org site and go through the docs there: there really is a lot there. > 4. It says on man xfce you can drag programs onto the Add to Menu > button. How? All I can achieve is moving the program about the > screen. ITs like the button is waitng for some event that I don;t know > about. Um, yeah, it supports DnD protocol. Just drag the icon from a DnD compliant file manager (such as Rox!). [DnD = Drag and Drop] > I like xfce but would like to see a way to sort out these bugbears. Ha! You think these ones are bad. Try using a 'minority' window manager! Matthew -- Matthew Sackman Nottingham England BOFH Excuse Board: Computers under water due to SYN flooding.
imaps - using courier and stunnel
hi all, I've been trying to get courier-imap to work under ssl and have all the relevant openssl and stunnel programs installed, but can't for the life of me work out what's wrong. I've edited the /etc/init.d/courier-imap so that it launches the ssl version of courier, and a ps ax shows it to be running. But whenever I try to get mutt to connect to it, it just freezes. The logs show it trying to startup, but then nothing, and when I kill off mutt there is an error message in the logs. I suspect that the certificates are not set up properly, but I've yet to come across any document that shows the openssl commands you use to create a x509 cert, and then the process you use to sign it yourself. Is it just a matter of using your gpg key to sign it? I hope someone knows what I need to do cos I've been scratching myself for several days on this one. Just pointers to good documentation would be a start. With thanks, Matthew using a 2.4.2 kernel and versions as in woody. -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing pgpGPuvnECT3G.pgp Description: PGP signature
[OT] Cordless MouseMan Optical
well, this is new and seems wonderful: a cordless optical. But, does anyone already use it / have any comments regarding it? Also, it has 4 buttons. How useful is this? Is there software under linux to actually be able to make use of it? Any info gratefully received. Matthew -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing pgp9yjGCLlw2Z.pgp Description: PGP signature
Re: [OT] Cordless MouseMan Optical
On Sat, Mar 31, 2001 at 08:05:49AM -0500, Thomas J. Hamman wrote: > Well, I recently bought a non-cordless (cordful? corded? > cord-bearing?) optical MouseMan, and it's very nice. You'll > appreciate having an optical mouse, especially if you're at the point > where it seems like you have to clean your current mouse every couple > days. > > I'm not sure if mine is supposed to have four buttons (yeah, I know > how stupid that sounds) -- I vaguely remember it being advertised as > 4-button, but I didn't care as long as it had at least 3 and it didn't > come in a retail box (I got a $19 price on pricewatch). Basically, it > has a wheel with a typical button on each side of the wheel, and a > third button where your thumb rests (it's comfortably ergonomic, BTW). > Also, pressing down on the wheel counts as a button--basically, it > seems to work as another third button. > > Being optical doesn't present any compatibility issues. Mine is USB > but comes with a PS/2 attachment so I can use it as PS/2, and I just > used the PS/2 MouseMan driver in X. > > I guess the real question for you is whether or not cordless mice have > compatibility problems... that I can't answer, since I know very > little about cordless mice. Well my current mouse is a logitech cordless mouseman, and there are no problems there, but the scroll wheel has just gone bust - I figured that I'd got it set up wrongly in Linux, but it turns out it was actually broken. I guess I'm really just curious as to whether it's possible to actually put the fourth mouse button (under the thumb) to any use. Matthew -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing pgpyeG7yBZUt6.pgp Description: PGP signature
Re: [OT] running a PIII with no fan?
On Mon, Mar 26, 2001 at 12:51:11PM -0800, [EMAIL PROTECTED] wrote: > On Sun, Mar 25, 2001 at 11:19:19PM -0800, Krzys Majewski wrote: > > Noise. > > -chris > > > > On Mon, 26 Mar 2001, Roberto Diaz wrote: > > > > > > > > Why do you want to burn you CPU? a fan is very cheap.. less than $15 some > > > models. You can buy one in all computer stores. > > > > > > Just curious.. why do you want to make this? (maybe you have other > > > solutions) > [snip: wanting to run without a fan] > > I hear there exist fans that can turn on and off according to > temperature. The local computer shop down the road is trying to find one > for me. Aparantly the newer P3 box sets have temperature-driven > variable-speed fans, too. And I believe some motherboards can even > control the fans through i2c, though I've not personally run across one. > > I was running my K6/300 for a while with the fan power disconnected, and > attaching it when doing CPU-intensive things like compiles and music. I > also have noflushd installed and have the hard drive spin down. You might want to take a look at www.supermicro.com. They make some excellant motherboards, all of the newer ones use variable speed temperature control feedback - the temperature of the CPU is available off two pins on the CPU which is then used to control the speed of the fan. You may want to contact them to ask them whose fans they use. Matthew -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing pgphlNT4qFYGd.pgp Description: PGP signature
WARNING: ReiserFS (and request for help!)
Hay all. A few nights ago, I left my machine on over night whilst ripping some cds. The file names were rather long (>20). In the morning all was fine until I rebooted at which point the / partition refused to mount and the entire system became unusable. The fault was ReiserFS: I seems that BugTraq amoungst others have found faults with reiserfs in its handling of files under certain conditions. This is the second time in 2 months that reiserfs has caused me to do a reinstallation, and after this one, I'm now starting to see kernel oops' elsewhere. I seriously recommend that if at all possible you get off reiserfs and back onto ext2 for at least a couple of months. I rapidly compiling 2.4.3 here to try and remove the bugs in reiserfs. But if it takes out my /usr partition then I really am *ucked as there is no backup and I don't have enough spare space to copy and convert back to ext2. There are already areas which I can not access - I get reiserfs errors - when trying to ls certain directories which unfortunately is part of the perl libs structure, thus rendering debconf useless and thus I am having huge problems installing anything that depends on debconf. Does anyone know whether these faults can be fixed? I've tried reiserfsck to no avail: it doesn't even report a fault. But I'm sitting here wondering just how much time I'll have before the whole thing goes Despondantly, Matthew -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing pgpLvo3pINoiU.pgp Description: PGP signature
Re: WARNING: ReiserFS (and request for help!)
On Wed, Apr 04, 2001 at 08:04:46PM +0200, Mario Vukelic wrote: > On 03 Apr 2001 22:48:57 +0100, Matthew Sackman wrote: > > > Does anyone know whether these faults can be fixed? I've tried > > reiserfsck to no avail: it doesn't even report a fault. But > > I'm sitting here wondering just how much time I'll have before > > the whole thing goes > > I can't run fortune anymore due to what seems to me to be reiserfs > errors. reiserfsck on an unmounted /usr gives no errors, but with -x it > segfaults due to unexpected values. Also I get kernel logs (BTW I'm > running 2.4.2) like these regularly: > > Mar 5 10:35:50 chello213047079152 kernel: vs-13070: > reiserfs_read_inode2: i/o failure occurred trying to find stat data of > [60448 60491 0x0 SD] > Mar 5 10:35:50 chello213047079152 kernel: vs-13048: reiserfs_iget: > bad_inode. Stat data of (60448 60491) not found > Mar 5 10:35:50 chello213047079152 kernel: vs-13048: reiserfs_iget: > bad_inode. Stat data of (60448 60491) not found > > And the ids are always the same. I had my fair share of manual > power-offs, so this might have caused it. Fortunately I have enough > space to backup /usr, so I'll try reiserfsck --rebuild-tree some day > > I'm very afraid ;o) These are exactely the kind of faults that I've been getting. This is with a 2.4.2 kernel. Just to let you know, 2.4.3 refused to compile (I suspect there was some corruption of the source code), so because things were getting worse, I tarred up /home and plonked it on a spare partition and have now started a fresh, going back to potato (with long overnight upgrades) and ext2. I'm not gonna try reiserfs until at least 2.4.8 if not later. I'll also be very interested to see how work with ext3 develops as there is a certain amount of safety built in to ext3 in that if it fails, the filing system reverts back to ext2 and so severly limits any damage. ho hum. Live and learn... Matthew -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing
