Re: [gentoo-user] Re: Checking sanity of system...
Nikos Chantziaras [10-04-04 08:28]: > On 04/04/2010 08:18 AM, meino.cra...@gmx.de wrote: > > > >Hi, > > > >this is no security issue in sense of attacks...it is related > >to the consistency of the system. > > > >Simple question (and may be complicate to answer... ;) ) > > > >How can I check, that my Gentoo system is uptodate > > emerge --sync && emerge -uDN world > > > >consistent and sane? > > Define "consistent" and "sane". Those words don't say anything, > really. > Ok... Consistency: Following each logical branch of each logical tree of control relationship, which is for example but not only the tree of dependancies, will end in a leave. These are the control paths. Sane: Following each logical branch of each logical tree of data relationships. which is for example but not only the interaction of scripts under /etc, will end in a leave. These are the data paths. HTH Best regards, mcc -- Please don't send me any Word- or Powerpoint-Attachments unless it's absolutely neccessary. - Send simply Text. See http://www.gnu.org/philosophy/no-word-attachments.html In a world without fences and walls nobody needs gates and windows.
Re: [gentoo-user] Checking sanity of system...
Dale [10-04-04 08:20]: > meino.cra...@gmx.de wrote: > >Hi, > > > >this is no security issue in sense of attacks...it is related > >to the consistency of the system. > > > >Simple question (and may be complicate to answer... ;) ) > > > >How can I check, that my Gentoo system is uptodate, consistent > >and sane? > > > >Best regards, > >mcc > > > > > > > > I think this is what you want. man glsa-check I don't use it so you > will have to read or wait until someone who uses it comes along. No > real clue how it works. > > I hope that helps and is what you are looking for. > > Dale > > :-) :-) > Hi Dale, As far as I can understand the man-page, glsa-check is a tool to check security settings in the system: _G_entoo _L_inux _S_ecurity _A_visory. I didnt know of this tool before so it is a goog hint anyway, but for the moment I want only check, whether the system is not cleanly setup/installed/updated in the sense of "it is working well" ;) Best regards, mcc -- Please don't send me any Word- or Powerpoint-Attachments unless it's absolutely neccessary. - Send simply Text. See http://www.gnu.org/philosophy/no-word-attachments.html In a world without fences and walls nobody needs gates and windows.
Re: [gentoo-user] Re: Checking sanity of system...
Dale [10-04-04 08:36]: > Nikos Chantziaras wrote: > >On 04/04/2010 08:18 AM, meino.cra...@gmx.de wrote: > >> > >>Hi, > >> > >>this is no security issue in sense of attacks...it is related > >>to the consistency of the system. > >> > >>Simple question (and may be complicate to answer... ;) ) > >> > >>How can I check, that my Gentoo system is uptodate > > > >emerge --sync && emerge -uDN world > > > > > >>consistent and sane? > > > >Define "consistent" and "sane". Those words don't say anything, > >really. > > > > > > > > You may also want to run revdep-rebuild as well. If you are talking > about your packages being "sane". That should catch anything that has > broken links or something else that leads to a package needing to be > recompiled. > > Dale > > :-) :-) > Hi Dale, revdep-rebuild is currently running! :) This was the only tool I knew before posting my question here :) :) Best regards, mcc -- Please don't send me any Word- or Powerpoint-Attachments unless it's absolutely neccessary. - Send simply Text. See http://www.gnu.org/philosophy/no-word-attachments.html In a world without fences and walls nobody needs gates and windows.
[gentoo-user] Re: Checking sanity of system...
On 04/04/2010 10:07 AM, meino.cra...@gmx.de wrote: Nikos Chantziaras [10-04-04 08:28]: On 04/04/2010 08:18 AM, meino.cra...@gmx.de wrote: Hi, this is no security issue in sense of attacks...it is related to the consistency of the system. Simple question (and may be complicate to answer... ;) ) How can I check, that my Gentoo system is uptodate emerge --sync&& emerge -uDN world consistent and sane? Define "consistent" and "sane". Those words don't say anything, really. Ok... Consistency: Following each logical branch of each logical tree of control relationship, which is for example but not only the tree of dependancies, will end in a leave. These are the control paths. Sane: Following each logical branch of each logical tree of data relationships. which is for example but not only the interaction of scripts under /etc, will end in a leave. These are the data paths. These are very abstract things you speak of. But I guess it boils down to "are there bugs in my system." Yes, there are. All software has bugs. There is a tool to check whether there are bugs: you. You use the software and check whether it works correctly or not. For anything more specific, you also need to be more specific with your questions. :)
Re: [gentoo-user] Re: Checking sanity of system...
meino.cra...@gmx.de wrote: Dale [10-04-04 08:36]: Nikos Chantziaras wrote: On 04/04/2010 08:18 AM, meino.cra...@gmx.de wrote: Hi, this is no security issue in sense of attacks...it is related to the consistency of the system. Simple question (and may be complicate to answer... ;) ) How can I check, that my Gentoo system is uptodate emerge --sync&& emerge -uDN world consistent and sane? Define "consistent" and "sane". Those words don't say anything, really. You may also want to run revdep-rebuild as well. If you are talking about your packages being "sane". That should catch anything that has broken links or something else that leads to a package needing to be recompiled. Dale :-) :-) Hi Dale, revdep-rebuild is currently running! :) This was the only tool I knew before posting my question here :) :) Best regards, mcc When I do my updates, I always do the following: eix-sync # This does my emerge --sync for me and updates eix since I use it sometimes for the FAST searches emerge -uvDNa world # My system is set up so that world includes system so this catches everything including deep dependencies (D) and changes of USE flags (N). I then check USE flags and anything else that may be odd. If I need to change something, I answer NO and repeat until I get it like it should be. After emerge finishes: emerge -p --depclean # I run that about once a month or if I know something is unneeded and should be removed. If it looks sane, I rerun without the -p. You could use -a I guess. revdep-rebuild -i # This makes sure nothing is broken and I run it each time after the emerge world whether --depclean is ran or not. It sometimes finds something broke and fixes it so I figure it is safer to run it and it do nothing than to not run it and something be broken. One more optional thing to run, python-updater. If I see python being updated, I run that too. Note: Python 3 should be popping its head up if it hasn't already. If it does, do NOT switch the system to it. A lot of packages do not work with it yet. If you switch to it, you can keep the pieces if it breaks. Sane thoughts did not prevail on -dev so you either have to mask it locally or it will be there basically doing nothing. Don't ask me why they did it. I was against the idea myself. < shrugs > That's how I do it and my little rig runs pretty good. I do have hiccups on occasion but everyone does eventually. Dale :-) :-)
Re: [gentoo-user] Disk or filesystem issue?
On Sunday 04 April 2010 01:52:17 Adam wrote: > > Probably neither. Can you Ctrl+F12 to see what the logs are > > saying? I've been getting kernel Oops! on shutdown on one > > machine of mine with the 2.6.31-gentoo-r10 kernel. Might be > > similar > > Ok, i found this, which is mentioning fglrx, and the problem may have > started when i changed driver versions. > > So is it possible for an X driver to cause filesystem issues? I would > have assumed not. > > > > Apr 4 10:38:08 sphinx shutdown[5768]: shutting down for system reboot > Apr 4 10:38:09 sphinx init: Switching to runlevel: 6 > Apr 4 10:38:09 sphinx gnome-session[5585]: WARNING: Detected that > screensaver has left the bus > Apr 4 10:38:09 sphinx gnome-keyring-daemon[5614]: dbus failure > unregistering from session: Connection is closed > Apr 4 10:38:09 sphinx su[5753]: pam_unix(su:session): session closed > for user root > Apr 4 10:38:09 sphinx kernel: uhci_hcd :00:1d.1: release dev 2 > ep81-INT, period 8, phase 4, 93 us > Apr 4 10:38:09 sphinx kernel: uhci_hcd :00:1a.0: release dev 4 > ep81-INT, period 8, phase 4, 14 us > Apr 4 10:38:09 sphinx kernel: __ratelimit: 20 callbacks suppressed > Apr 4 10:38:09 sphinx kernel: BUG: using smp_processor_id() in > preemptible [] code: X/5275 > Apr 4 10:38:09 sphinx kernel: caller is KAS_GetExecutionLevel+0xd/0x120 > [fglrx] > Apr 4 10:38:09 sphinx kernel: Pid: 5275, comm: X Tainted: P > 2.6.31-gentoo-r6 #4 > Apr 4 10:38:09 sphinx kernel: Call Trace: > Apr 4 10:38:09 sphinx kernel: [] ? > debug_smp_processor_id+0xd9/0xe0 > Apr 4 10:38:09 sphinx kernel: [] ? > KAS_GetExecutionLevel+0xd/0x120 [fglrx] > Apr 4 10:38:09 sphinx kernel: [] ? > MCIL_GetExecutionLevel+0x39/0x80 [fglrx] > Apr 4 10:38:09 sphinx kernel: [] ? > CallbackQueueAccess+0x2b/0x2a0 [fglrx] > Apr 4 10:38:09 sphinx kernel: [] ? > UnRegisterIRQClient_Worker+0x0/0xe0 [fglrx] > Apr 4 10:38:09 sphinx kernel: [] ? > MCIL_bMiniportCapEnabled+0x82/0xa0 [fglrx] > Apr 4 10:38:09 sphinx kernel: [] ? > UnRegisterIRQClient+0x83/0xe0 [fglrx] > Apr 4 10:38:09 sphinx kernel: [] ? > IRQMGR_IRQSourceSupported+0x7e/0xa0 [fglrx] > Apr 4 10:38:09 sphinx kernel: [] ? > IRQMGR_Access+0x131/0x190 [fglrx] > Apr 4 10:38:09 sphinx kernel: [] ? > fireglAsyncioUnregisterIntMsgHandlers+0x2c3/0x3c0 [fglrx] > Apr 4 10:38:09 sphinx kernel: [] ? > asyncIONotifyMsg+0x234/0x3e0 [fglrx] > Apr 4 10:38:09 sphinx kernel: [] ? > firegl_asyncio_write+0x189/0x250 [fglrx] > Apr 4 10:38:09 sphinx kernel: [] ? > ip_firegl_write+0x58/0xa0 [fglrx] > Apr 4 10:38:09 sphinx kernel: [] ? vfs_write+0xcb/0x180 > Apr 4 10:38:09 sphinx kernel: [] ? sys_write+0x53/0xa0 > Apr 4 10:38:09 sphinx kernel: [] ? > system_call_fastpath+0x16/0x1b > Apr 4 10:38:09 sphinx kernel: BUG: using smp_processor_id() in > preemptible [] code: X/5275 > Apr 4 10:38:09 sphinx kernel: caller is > KCL_SPINLOCK_STATIC_Grab+0x25/0x110 [fglrx] > Apr 4 10:38:09 sphinx kernel: Pid: 5275, comm: X Tainted: P > 2.6.31-gentoo-r6 #4 > Apr 4 10:38:09 sphinx kernel: Call Trace: > Apr 4 10:38:09 sphinx kernel: [] ? > debug_smp_processor_id+0xd9/0xe0 It could be that this kernel does not play well with fglrx? I would unmask the next kernel up the tree and try that out. This particular kernel version you are running had a few bugs hanging around, which may affect your problem. PS. My machine that keeps crashing on shutdown doesn't always crash at exactly the same point - but it is well after X has been stopped and invariably after the fs has been unmounted. PPS. Clutching at straws, but just for testing purposes you may want to boot with no X, just on a console and try shuting down from there? -- Regards, Mick signature.asc Description: This is a digitally signed message part.
Re: [gentoo-user] Checking sanity of system...
Am 04.04.2010 07:18, schrieb meino.cra...@gmx.de: > Hi, > > this is no security issue in sense of attacks...it is related > to the consistency of the system. > > Simple question (and may be complicate to answer... ;) ) > > How can I check, that my Gentoo system is uptodate, consistent > and sane? > > Best regards, > mcc > > > Hi, Every Update: emerge --sync && emerge -uDN world --> Read the Packagemessages for Instructions. etc-update# Merge new Configfiles revdep-rebuild# Identify broken libraries >From time to time: emerge --deplcean (-p) && revdep-rebuild# Delete old packages and sort out the resulting broken packages eclean distfiles# Delete the old source-packages in your distfile repo. Regards, Norman
Re: [gentoo-user] How does grub assemble a RAID1 for / ??
Try appending md=3,/dev/sdb3,/dev/sdc3 to the kernel command line parameters. On 04/04/2010 01:45 AM, Albert Hopkins wrote: > On Sat, 2010-04-03 at 16:07 -0700, Mark Knecht wrote: >> The install is complete but it won't boot. grub finds the kernel >> and starts booting but then I get the typical VFS file sync error as >> the kernel starts looking for the install on /dev/md3. What I'm not >> understanding is how does the boot process get the information >> required to assemble the RAID device. > > GRUB does not assemble raid. That's why it only works with RAID1. > > By your own account, GRUB has succeeded, therefore GRUB is not the > problem. > > The problem is the kernel > > The kernel assembles RAID by looking for partitions of with the Linux > RAID partition type, finding out what kind of RAID they are, and > assembling them (according to their RAID volume UUID). > > You apparently only have one RAID volume. It's probably being assigned > to /dev/md0, yet you are passing root=/dev/md3.. not sure why you are > doing that. -- Xavier Parizet YaGB : http://gentooist.com GPG :C7DC B10E FC21 63BE B453 D239 F6E6 DF65 1569 91BF signature.asc Description: OpenPGP digital signature
Re: [gentoo-user] Re: How does grub assemble a RAID1 for / ??
On Sat, Apr 3, 2010 at 7:38 PM, Kerin Millar wrote: > On 04/04/2010 02:01, Mark Knecht wrote: >> >> Tried changing root=/dev/md0. No change. >> >> The actual failure message is the fairly standard >> >> VFS - Unable to mount root fs on unknown-block(9,0) > > [snip] > >> CONFIG_MD_RAID1=y > > That's all that needs to be enabled within the RAID section of the kernel. > However, all the other options that would normally be required to boot must > also be compiled in statically for things to work as expected (ATA/SCSI > controller driver, filesystem of choice, CONFIG_BLK_DEV_SD and so forth). It > seems that you may have overlooked something. However, it's impossible to > determine whether that's the case based on the information presented thus > far. > > I would suggest that you double-check your .config in full, or present it > here for review, along with the output of lspci -nn. > > Cheers, > > --Kerin Hi Kerin, Happy for any help I can get. Instead of the whole .config file here's a diff. Remember that the machine already boots non-RAID from /dev/sda and I'm trying to build my first RAID boot on /dev/sdb & sdc. First, here's the RAID I would like to boot from: keeper ~ # mdadm -A /dev/md3 /dev/sdb3 /dev/sdc3 mdadm: /dev/md3 has been started with 2 drives. keeper ~ # mdadm --detail /dev/md3 /dev/md3: Version : 1.01 Creation Time : Sat Apr 3 11:43:39 2010 Raid Level : raid1 Array Size : 52436092 (50.01 GiB 53.69 GB) Used Dev Size : 52436092 (50.01 GiB 53.69 GB) Raid Devices : 2 Total Devices : 2 Persistence : Superblock is persistent Update Time : Sun Apr 4 06:40:54 2010 State : clean Active Devices : 2 Working Devices : 2 Failed Devices : 0 Spare Devices : 0 Name : keeper:3 (local to host keeper) UUID : 6dcf5ddb:c4a2d5ea:ba59df10:f5473502 Events : 3703 Number Major Minor RaidDevice State 0 8 190 active sync /dev/sdb3 1 8 351 active sync /dev/sdc3 keeper ~ # cat /proc/mdstat Personalities : [raid1] md3 : active raid1 sdb3[0] sdc3[1] 52436092 blocks super 1.1 [2/2] [UU] unused devices: keeper ~ # Here's the diff of the running kernel without RAID and the kernel I created while in the install chroot on the RAID device: keeper ~ # diff /usr/src/linux/.config /mnt/gentoo/usr/src/linux/.config 4c4 < # Mon Mar 29 01:02:31 2010 --- > # Sun Apr 4 06:28:53 2010 893,912c893,906 < CONFIG_MD_LINEAR=m < CONFIG_MD_RAID0=m < CONFIG_MD_RAID1=m < CONFIG_MD_RAID10=m < CONFIG_MD_RAID456=m < # CONFIG_MULTICORE_RAID456 is not set < CONFIG_MD_RAID6_PQ=m < # CONFIG_ASYNC_RAID6_TEST is not set < CONFIG_MD_MULTIPATH=m < CONFIG_MD_FAULTY=m < CONFIG_BLK_DEV_DM=m < CONFIG_DM_DEBUG=y < CONFIG_DM_CRYPT=m < CONFIG_DM_SNAPSHOT=m < CONFIG_DM_MIRROR=m < # CONFIG_DM_LOG_USERSPACE is not set < CONFIG_DM_ZERO=m < CONFIG_DM_MULTIPATH=m < # CONFIG_DM_MULTIPATH_QL is not set < # CONFIG_DM_MULTIPATH_ST is not set --- > # CONFIG_MD_LINEAR is not set > CONFIG_MD_RAID0=y > CONFIG_MD_RAID1=y > # CONFIG_MD_RAID10 is not set > # CONFIG_MD_RAID456 is not set > # CONFIG_MD_MULTIPATH is not set > # CONFIG_MD_FAULTY is not set > CONFIG_BLK_DEV_DM=y > # CONFIG_DM_DEBUG is not set > # CONFIG_DM_CRYPT is not set > # CONFIG_DM_SNAPSHOT is not set > # CONFIG_DM_MIRROR is not set > # CONFIG_DM_ZERO is not set > # CONFIG_DM_MULTIPATH is not set 914,915c908,909 < CONFIG_DM_UEVENT=y < CONFIG_BLK_DEV_DM_BBR=m --- > # CONFIG_DM_UEVENT is not set > # CONFIG_BLK_DEV_DM_BBR is not set 2293,2298d2286 < CONFIG_XOR_BLOCKS=m < CONFIG_ASYNC_CORE=m < CONFIG_ASYNC_MEMCPY=m < CONFIG_ASYNC_XOR=m < CONFIG_ASYNC_PQ=m < CONFIG_ASYNC_RAID6_RECOV=m keeper ~ # One additional thing I thought of last night was some message that came up when I first built the RAID about the partitions having metadata and to be sure that the bootloader understands metadata. In the cool light of morning that seems fairly important. I am using grub-static on this machine. I assumed this would be OK but possibly it isn't? If rebuilding the RAID from scratch is important, or just makes things more straight forward, then don't hesitate to suggest it and I'll document the build step by step. This install isn't important. I'm just doing it to learn how to do RAID and most importantly to test the disk drives. I purchased other disk drives that aren't working with RAID at all so I wanted to test these a bit before I did anything real. The final install with be a 3 disk RAID1 but the 3rd drive hasn't arrived yet so none of this is critically important. Thanks! Cheers, Mark
Re: [gentoo-user] How does grub assemble a RAID1 for / ??
On Sun, Apr 4, 2010 at 2:57 AM, Xavier Parizet wrote: > Try appending md=3,/dev/sdb3,/dev/sdc3 to the kernel command line > parameters. > Thanks. Tried that one last night but no luck, although it does change the message to Unknown-block(9,3) from Unknown-block(9,0). Cheers, Mark
[gentoo-user] Problems with booting from SD card on EEE 1201H
Hello Everyone! I am trying to install gentoo on my eee1201h and used this howto: http://www.gentoo.org/doc/en/liveusb.xml to setup a bootable SD card. The system boots but depscan fail with "line 128: /bin/chmod: Input/Output Error" "line 139: Bus Error" and a lot more. Almost every command i issue in the shell fails with Input/Output error (even ls). Any ideas how I can get this to work? Thanks, jdb
[gentoo-user] Re: How does grub assemble a RAID1 for / ??
On 04/04/2010 15:20, Mark Knecht wrote: On Sat, Apr 3, 2010 at 7:38 PM, Kerin Millar wrote: On 04/04/2010 02:01, Mark Knecht wrote: Tried changing root=/dev/md0. No change. The actual failure message is the fairly standard VFS - Unable to mount root fs on unknown-block(9,0) [snip] CONFIG_MD_RAID1=y That's all that needs to be enabled within the RAID section of the kernel. However, all the other options that would normally be required to boot must also be compiled in statically for things to work as expected (ATA/SCSI controller driver, filesystem of choice, CONFIG_BLK_DEV_SD and so forth). It seems that you may have overlooked something. However, it's impossible to determine whether that's the case based on the information presented thus far. I would suggest that you double-check your .config in full, or present it here for review, along with the output of lspci -nn. Cheers, --Kerin Hi Kerin, Happy for any help I can get. Instead of the whole .config file here's a diff. Remember that the machine already boots non-RAID from /dev/sda and I'm trying to build my first RAID boot on /dev/sdb& sdc. No, really, the whole thing needs to be seen, along with the lspci data. It's very likely that this thread can be drawn to a close if you provide exactly what's being asked for :) One additional thing I thought of last night was some message that came up when I first built the RAID about the partitions having metadata and to be sure that the bootloader understands metadata. In The bootloader does not enter into this. If the kernel is being loaded - which, by your own admission it is - the bootloader has done its job. What happens thereafter is entirely the responsibility of the kernel. Essentially, the subject of this thread is a misnomer and the issue lies with your kernel. As for the warning regarding metadata, it's applicable to legacy bootloaders which may not be able to fathom the presence of the md superblock data at the beginning of a block device that happens to be a member of a raid1 volume. As far as grub is concerned, this is a non-issue. Even if it were an issue, you wouldn't even get as far as being able to load the kernel in the first instance. Indeed, the bootloader itself would likely fail to initialise properly. >If rebuilding the RAID from scratch is important, or just makes > things more straight forward, then don't hesitate to suggest it and > I'll document the build step by step. This install isn't important. On the other hand, if you don't get to the point of understanding why the kernel isn't configured so as to be able to assemble the array on this occasion, a re-install isn't going to change that. Moreover, you won't be able to fix any such problem that may occur again unaided. Cheers, --Kerin
Re: [gentoo-user] Re: How does grub assemble a RAID1 for / ??
On Sun, Apr 4, 2010 at 10:32 AM, Kerin Millar wrote: > On 04/04/2010 15:20, Mark Knecht wrote: >> >> On Sat, Apr 3, 2010 at 7:38 PM, Kerin Millar wrote: >>> >>> On 04/04/2010 02:01, Mark Knecht wrote: Tried changing root=/dev/md0. No change. The actual failure message is the fairly standard VFS - Unable to mount root fs on unknown-block(9,0) >>> >>> [snip] >>> CONFIG_MD_RAID1=y >>> >>> That's all that needs to be enabled within the RAID section of the >>> kernel. >>> However, all the other options that would normally be required to boot >>> must >>> also be compiled in statically for things to work as expected (ATA/SCSI >>> controller driver, filesystem of choice, CONFIG_BLK_DEV_SD and so forth). >>> It >>> seems that you may have overlooked something. However, it's impossible to >>> determine whether that's the case based on the information presented thus >>> far. >>> >>> I would suggest that you double-check your .config in full, or present it >>> here for review, along with the output of lspci -nn. >>> >>> Cheers, >>> >>> --Kerin >> >> Hi Kerin, >> Happy for any help I can get. >> >> Instead of the whole .config file here's a diff. Remember that the >> machine already boots non-RAID from /dev/sda and I'm trying to build >> my first RAID boot on /dev/sdb& sdc. >> > > No, really, the whole thing needs to be seen, along with the lspci data. > It's very likely that this thread can be drawn to a close if you provide > exactly what's being asked for :) > >> One additional thing I thought of last night was some message that >> came up when I first built the RAID about the partitions having >> metadata and to be sure that the bootloader understands metadata. In > > The bootloader does not enter into this. If the kernel is being loaded - > which, by your own admission it is - the bootloader has done its job. What > happens thereafter is entirely the responsibility of the kernel. > > Essentially, the subject of this thread is a misnomer and the issue lies > with your kernel. > > As for the warning regarding metadata, it's applicable to legacy bootloaders > which may not be able to fathom the presence of the md superblock data at > the beginning of a block device that happens to be a member of a raid1 > volume. As far as grub is concerned, this is a non-issue. Even if it were an > issue, you wouldn't even get as far as being able to load the kernel in the > first instance. Indeed, the bootloader itself would likely fail to > initialise properly. > >> If rebuilding the RAID from scratch is important, or just makes >> things more straight forward, then don't hesitate to suggest it and >> I'll document the build step by step. This install isn't important. > > On the other hand, if you don't get to the point of understanding why the > kernel isn't configured so as to be able to assemble the array on this > occasion, a re-install isn't going to change that. Moreover, you won't be > able to fix any such problem that may occur again unaided. > > Cheers, > > --Kerin No problem supplying it. I did the rebuild this morning but forced metadata to Type 1.0. No change as you suggested there wouldn't be. OK, here's: 1) lspci to read & lspci -k to see drivers both from the non-RAID kernel 2) The RAID kernel config 3) At the very end a diff between the kernel config in this email and the running one without RAID. (I.e. - the changes I made to attempt to mount / which is on RAID.) Note that the Marvell SATA controller is for two external eSATA ports that have nothing attached at this time. It's the Intel controller that's in play here. Thanks, Mark keeper ~ # lspci 00:00.0 Host bridge: Intel Corporation X58 I/O Hub to ESI Port (rev 13) 00:01.0 PCI bridge: Intel Corporation 5520/5500/X58 I/O Hub PCI Express Root Port 1 (rev 13) 00:03.0 PCI bridge: Intel Corporation 5520/5500/X58 I/O Hub PCI Express Root Port 3 (rev 13) 00:07.0 PCI bridge: Intel Corporation 5520/5500/X58 I/O Hub PCI Express Root Port 7 (rev 13) 00:10.0 PIC: Intel Corporation 5520/5500/X58 Physical and Link Layer Registers Port 0 (rev 13) 00:10.1 PIC: Intel Corporation 5520/5500/X58 Routing and Protocol Layer Registers Port 0 (rev 13) 00:14.0 PIC: Intel Corporation 5520/5500/X58 I/O Hub System Management Registers (rev 13) 00:14.1 PIC: Intel Corporation 5520/5500/X58 I/O Hub GPIO and Scratch Pad Registers (rev 13) 00:14.2 PIC: Intel Corporation 5520/5500/X58 I/O Hub Control Status and RAS Registers (rev 13) 00:14.3 PIC: Intel Corporation 5520/5500/X58 I/O Hub Throttle Registers (rev 13) 00:19.0 Ethernet controller: Intel Corporation 82567LM-2 Gigabit Network Connection 00:1a.0 USB Controller: Intel Corporation 82801JI (ICH10 Family) USB UHCI Controller #4 00:1a.1 USB Controller: Intel Corporation 82801JI (ICH10 Family) USB UHCI Controller #5 00:1a.2 USB Controller: Intel Corporation 82801JI (ICH10 Family) USB UHCI Controller #6 00:1a.7 USB Controller: Intel Corporation 82801JI (ICH10 Family) USB2 EHCI Controller
[gentoo-user] Re: How does grub assemble a RAID1 for / ??
On 04/04/2010 18:50, Mark Knecht wrote: [snip] No problem supplying it. I did the rebuild this morning but forced metadata to Type 1.0. No change as you suggested there wouldn't be. OK, here's: 1) lspci to read& lspci -k to see drivers both from the non-RAID kernel 2) The RAID kernel config 3) At the very end a diff between the kernel config in this email and the running one without RAID. (I.e. - the changes I made to attempt to mount / which is on RAID.) Note that the Marvell SATA controller is for two external eSATA ports that have nothing attached at this time. It's the Intel controller that's in play here. Thanks, Mark [snip lspci data] OK. [snip kconfig data] Something isn't right here. This .config appears to be severely stunted. It's missing lots of options that should be defined (whether active or not). Indeed, a typical .config might easily exceed 2000 lines. It's also missing the comment at the top which describes the kernel version in use. Is this really the entire .config file that is currently residing within /mnt/gentoo/usr/src/linux, having mounted the root filesystem and that was used to build the kernel you're attempting to boot with? If so, it's broken and you should just delete it entirely and start anew with make menuconfig. If not, then please present the file in full (no diffs, no obfuscation please). Aside from all of that, notable options that are going to be required to boot in your case are: CONFIG_MD_RAID1 CONFIG_SATA_AHCI CONFIG_BLK_DEV_SD CONFIG_BLK_DEV_SR CONFIG_MSDOS_PARTITION (normally implicit but worth mentioning) That, and the option corresponding with whichever filesystem you use. Also, make sure CONFIG_SYSFS_DEPRECATED_V2 is off or else udev will balk. You may use the forward slash key to search for option names in make menuconfig (never edit .config directly). All needed options should be enabled as <*>. Also, if you're not experienced with kernel configuration and need a skeleton .config with which to begin, I would suggest you take a look at http://kernel-seeds.org. Cheers, --Kerin
[gentoo-user] alsa-update not in sync with kernel version
Hi, I am running a vanilla kernel (linux-2.6.32.11), This kernel uses alsa-1.0.21. When doing the eix-sync-thingy, emerge always suggests to update to alsa-1.0.22. Is there a way that emerge/eix or whatever relizes the version of alsa which the kernel is runnig and only suggests updates which are in sync which such version? Best regards, mcc -- Please don't send me any Word- or Powerpoint-Attachments unless it's absolutely neccessary. - Send simply Text. See http://www.gnu.org/philosophy/no-word-attachments.html In a world without fences and walls nobody needs gates and windows.
Re: [gentoo-user] Checking sanity of system...
>Von: nor...@smash-net.org >Datum: 04.04.2010 11:37 >An: >Betreff: Re: [gentoo-user] Checking sanity of system... > >Am 04.04.2010 07:18, schrieb meino.cra...@gmx.de: >> Hi, >> >> this is no security issue in sense of attacks...it is related >> to the consistency of the system. >> >> Simple question (and may be complicate to answer... ;) ) >> >> How can I check, that my Gentoo system is uptodate, consistent >> and sane? >> >> Best regards, >> mcc >> >> >> >Hi, > >Every Update: >emerge --sync && emerge -uDN world >--> Read the Packagemessages for Instructions. >etc-update# Merge new Configfiles >revdep-rebuild# Identify broken libraries > >From time to time: >emerge --deplcean (-p) && revdep-rebuild# Delete old packages and >sort out the resulting broken packages >eclean distfiles# Delete the old source-packages in your distfile repo. > >Regards, >Norman > > > and aditionally from time to time: eix -u (packages which have at least one slotted version installed which is not the best version within that slot) and eix-test-obsolete -d (display missing packages or packages with obsolete entries in /etc/portage/package.* ) Check for viruses, lookout for SMART data, and filesystem inconsistencies, check temperatures, check log files, listen to fans... there are many levels where a system could fail. Urs
[gentoo-user] Moving the system from one disk to another
Hi, I have to move my whole system from one disk to another bigger one. I think of doing as follows: Boot a system via CD/DVD (knoppix for example). Mount small disk read-only Mount bigger disk read-write cd into mountpoint of the first one cp -a . ../ Seems to me slow but correct? Or? (I have to set the bootable flag of the correct partion additionally...) I dont want to have a booting system afterwards, which "runs" for -- say -- three month and suddenly hit a obscure bug due to my copy-commands, which only did it to 99.87% correct... ;) I would like to preserve as much as possible of the file/directoy times ,,, Or does a mystical command with s-tar a better job faster? Thank you very much in adance for any help! Best regards, mcc -- Please don't send me any Word- or Powerpoint-Attachments unless it's absolutely neccessary. - Send simply Text. See http://www.gnu.org/philosophy/no-word-attachments.html In a world without fences and walls nobody needs gates and windows.
[gentoo-user] Re: How does grub assemble a RAID1 for / ??
On 04/04/2010 19:43, Mark Knecht wrote: [snip] Aside from all of that, notable options that are going to be required to boot in your case are: CONFIG_MD_RAID1 CONFIG_SATA_AHCI CONFIG_BLK_DEV_SD CONFIG_BLK_DEV_SR CONFIG_MSDOS_PARTITION (normally implicit but worth mentioning) That, and the option corresponding with whichever filesystem you use. Also, make sure CONFIG_SYSFS_DEPRECATED_V2 is off or else udev will balk. You may use the forward slash key to search for option names in make menuconfig (never edit .config directly). All needed options should be enabled as<*>. Also, if you're not experienced with kernel configuration and need a skeleton .config with which to begin, I would suggest you take a look at http://kernel-seeds.org. Cheers, --Kerin Sorry. I was on a Windows box and it looks like putty cut it off. Booted into Linux and this looks more correct. (2424 lines in vi on that machine and 2424 lines in late on this machine so I think it's all there...) I looked at your suggestions above and they are all set to =y. CONFIG_SYSFS_DEPRECATED_V2 is 'not set'. - Mark # # Automatically generated make config: don't edit # Linux kernel version: 2.6.33-gentoo # Sun Apr 4 09:56:49 2010 # [snip long config] Ah, that looks better. Not that I've pored over every line, but at first glance everything seems to be in order. There are no obvious gotchas that I can see, so I'm somewhat puzzled. Here are a few random things that spring to mind though ... I would suggest switching off CONFIG_IDE. It may contend for control of your hardware with the AHCI driver, assuming that emulation/comptability mode is enabled in the BIOS. The device nodes may be unavailable at the time that they are needed, for some strange reason. If you mount the root filesystem from a livecd (with no bind mounts), try creating static nodes in dev/: mknod /dev/md0 b 9 0 mknod /dev/md1 b 9 1 mknod /dev/md2 b 9 2 mknod /dev/md3 b 9 3 Note that the preferred minor of your array can be determined by examining any component partition. For instance, "mdadm -E /dev/sdb3". Avoid manual specification of the RAID parameters. The kernel should be perfectly able to assemble the array by examining the superblocks of partitions of type "FD". Does it work if you specify "root=/dev/sdb3" or "root=/dev/sdc3"? With raid1, it's possible to mount just the component partition (although it will later result in a resync). The point is, it would at least confirm as to whether the underlying block devices are available to the kernel from the outset. Cheers, --Kerin
[gentoo-user] Re: Moving the system from one disk to another
On 04/04/2010 20:05, meino.cra...@gmx.de wrote: Hi, I have to move my whole system from one disk to another bigger one. I think of doing as follows: Boot a system via CD/DVD (knoppix for example). Mount small disk read-only Mount bigger disk read-write cd into mountpoint of the first one cp -a . ../ Seems to me slow but correct? Or? cp -a is fine although, in my experience, I've found that rsync is more reliable: rsync -av /mnt/oldrootfs/. /mnt/newrootfs/. Note that -a covers almost everything but you may need the following additional options depending on how your filesystem has been used: -H = preserve hard link -A = preserve ACLs -X = preserve extended attributes (in in doubt, recommended) Whichever way you go about it, ensure that no pseudo-filesystem or bind mounts are present within "/mnt/oldrootfs" at the time. Cheers, --Kerin
Re: [gentoo-user] Re: How does grub assemble a RAID1 for / ??
On Sun, Apr 4, 2010 at 12:29 PM, Kerin Millar wrote: > > Ah, that looks better. Not that I've pored over every line, but at first > glance everything seems to be in order. There are no obvious gotchas that I > can see, so I'm somewhat puzzled. As am I! > > Here are a few random things that spring to mind though ... > > I would suggest switching off CONFIG_IDE. It may contend for control of your > hardware with the AHCI driver, assuming that emulation/comptability mode is > enabled in the BIOS. > > The device nodes may be unavailable at the time that they are needed, for > some strange reason. If you mount the root filesystem from a livecd (with no > bind mounts), try creating static nodes in dev/: > > mknod /dev/md0 b 9 0 > mknod /dev/md1 b 9 1 > mknod /dev/md2 b 9 2 > mknod /dev/md3 b 9 3 So this suggests that because I did the install from a running Gentoo chroot that the mknod commands didn't stick? Somehow they were part of, or because part of, the host Gentoo non-RAID install? Interesting. I'll boot next ffrom the Live CD and try it. > > Note that the preferred minor of your array can be determined by examining > any component partition. For instance, "mdadm -E /dev/sdb3". That's what I thought from the install guide but it seemed others here had other opions. > > Avoid manual specification of the RAID parameters. The kernel should be > perfectly able to assemble the array by examining the superblocks of > partitions of type "FD". > > Does it work if you specify "root=/dev/sdb3" or "root=/dev/sdc3"? With > raid1, it's possible to mount just the component partition (although it will > later result in a resync). The point is, it would at least confirm as to > whether the underlying block devices are available to the kernel from the > outset. OK, this was interesting, and I suppose it depends on what you mean 'should work'. Clearly it gets much further, into the interactive portion of the boot with the green asterics on the right. When it finally gets to the md0 portion it says the the superblock does not correctly specify an ext2 filesystem and asks for a password or Control D. Control D reboots and entering the password results in a (none) prompt and seems to require a hard reset. None the less it got much further, but not to a usable state. Off to do the Live CD work. Back in 30 minutes. Cheers, Mark
Re: [gentoo-user] Moving the system from one disk to another
meino.cra...@gmx.de wrote: > > Hi, > > I have to move my whole system from one disk to another > bigger one. > > I think of doing as follows: > Boot a system via CD/DVD (knoppix for example). > Mount small disk read-only > Mount bigger disk read-write > > cd into mountpoint of the first one > cp -a . ../ > > Seems to me slow but correct? Or? > > (I have to set the bootable flag of the correct partion > additionally...) > > I dont want to have a booting system afterwards, which "runs" for -- > say -- three month and suddenly hit a obscure bug due to my > copy-commands, which only did it to 99.87% correct... ;) > > I would like to preserve as much as possible of the file/directoy > times ,,, > > Or does a mystical command with s-tar a better job faster? > > Thank you very much in adance for any help! I like rsync and be sure to say --numeric-ids or whatever to retain the same userids for permissions, but otherwise you are good to go. -- Your life is like a penny. You're going to lose it. The question is: How do you spend it? John Covici cov...@ccs.covici.com
Re: [gentoo-user] Re: Moving the system from one disk to another
On Sun, 04 Apr 2010 20:35:11 +0100, Kerin Millar wrote: > Whichever way you go about it, ensure that no pseudo-filesystem or bind > mounts are present within "/mnt/oldrootfs" at the time. Use the -x option with rsync to stop it descending into other filesystems. -- Neil Bothwick Did you hear about the dyslexic devil worshiper? He sold his soul to Santa! signature.asc Description: PGP signature
[gentoo-user] Re: How does grub assemble a RAID1 for / ??
On 04/04/2010 20:52, Mark Knecht wrote: So this suggests that because I did the install from a running Gentoo chroot that the mknod commands didn't stick? Somehow they were part of, or because part of, the host Gentoo non-RAID install? Interesting. I don't know what the handbook says these days, but I understand that it's lately taken to recommending a bind mount of /dev. If that's the case, then none of the static device nodes that are contained with the /dev directory within the stage3 tarball will actually make it on to the root filesystem because, at the time of unpacking, they will be redirected to the (volatile) tmpfs mount at /dev with respect to the livecd environment. From the point of view of the problem you're having, it may not even matter but it's one thing to try anyway. It's interesting that the component disks are available. In principle, there's no reason why the array should not be assembled. Here's an example of how it looks during the boot process on one of my systems: Command line: root=/dev/md2 Kernel command line: root=/dev/md2 quiet ata1: SATA max UDMA/133 cmd 0xcc30 ctl 0xcc28 bmdma 0xcc40 irq 23 ata2: SATA max UDMA/133 cmd 0xcc38 ctl 0xcc2c bmdma 0xcc48 irq 23 md: raid1 personality registered for level 1 md: Waiting for all devices to be available before autodetect md: If you don't use raid, use raid=noautodetect md: Autodetecting RAID arrays. md: Scanned 6 and added 6 devices. md: autorun ... md: considering sdb3 ... md: adding sdb3 ... md: sdb2 has different UUID to sdb3 md: sdb1 has different UUID to sdb3 md: adding sda3 ... md: sda2 has different UUID to sdb3 md: sda1 has different UUID to sdb3 md: created md2 md: bind md: bind md: running: md: kicking non-fresh sdb3 from array! md: unbind md: export_rdev(sdb3) raid1: raid set md2 active with 1 out of 2 mirrors md2: detected capacity change from 0 to 248798183424 md: considering sdb2 ... md: adding sdb2 ... md: sdb1 has different UUID to sdb2 md: adding sda2 ... md: sda1 has different UUID to sdb2 md: created md1 md: bind md: bind md: running: md: kicking non-fresh sdb2 from array! md: unbind md: export_rdev(sdb2) raid1: raid set md1 active with 1 out of 2 mirrors md1: detected capacity change from 0 to 1085669376 md: considering sdb1 ... md: adding sdb1 ... md: adding sda1 ... md: created md0 You might consider disabling the framebuffer temporarily, as it may impact upon your ability to see what's going on early during the boot process. Cheers, --Kerin
Re: [gentoo-user] alsa-update not in sync with kernel version
On Sun, 4 Apr 2010 20:42:48 +0200, meino.cra...@gmx.de wrote: > I am running a vanilla kernel (linux-2.6.32.11), This kernel uses > alsa-1.0.21. > > When doing the eix-sync-thingy, emerge always suggests to update > to alsa-1.0.22. The 1.0.22 userspace tools work fine with the drivers in 2.6.32. -- Neil Bothwick I have seen things you lusers would not believe. I've seen Sun monitors on fire off the side of the multimedia lab. I've seen NTU lights glitter in the dark near the Mail Gate. All these things will be lost in time, like the root partition last week. Time to die. signature.asc Description: PGP signature
[gentoo-user] Anyone ever emerged dev-libs/boost with FEATURES="test" and finished?
I'm running with full system FEATURES="test" on, and I have a couple of programs that depended on dev-libs/boost. The boost testsuite always fails in my computer due to insufficient disk space, I usually simply skip the test for boost and just go on with the merge. But today, I decided to let the testsuite run to completion; so in preparation for that, I plugged in an external harddisk and made it so that /var/tmp/portage points to an empty disk image in the external harddrive. This setup works ok, and the testsuite is still running, however I saw now that the disk image's is now taking ~18 GB (and counting) while "du -sh" on /var/tmp/portage counted ~13GB. So, the question is, has anyone successfully compiled and run FEATURES="test" on boost and knows how much space the tests eat up in the end? I am suspecting of the possibility that maybe a testsuite gets into an infinite loop while writing a file or something constantly eats up diskspace. Or is it just that boost has an outrageously too extensive testsuite and it will turn out ok if I just left it to run.
Re: [gentoo-user] Re: How does grub assemble a RAID1 for / ??
On Sun, Apr 4, 2010 at 1:06 PM, Kerin Millar wrote: > On 04/04/2010 20:52, Mark Knecht wrote: >> >> So this suggests that because I did the install from a running Gentoo >> chroot that the mknod commands didn't stick? Somehow they were part >> of, or because part of, the host Gentoo non-RAID install? Interesting. > > I don't know what the handbook says these days, but I understand that it's > lately taken to recommending a bind mount of /dev. If that's the case, then > none of the static device nodes that are contained with the /dev directory > within the stage3 tarball will actually make it on to the root filesystem > because, at the time of unpacking, they will be redirected to the (volatile) > tmpfs mount at /dev with respect to the livecd environment. > > From the point of view of the problem you're having, it may not even matter > but it's one thing to try anyway. > > It's interesting that the component disks are available. In principle, > there's no reason why the array should not be assembled. Here's an example > of how it looks during the boot process on one of my systems: > > Command line: root=/dev/md2 > Kernel command line: root=/dev/md2 quiet > ata1: SATA max UDMA/133 cmd 0xcc30 ctl 0xcc28 bmdma 0xcc40 irq 23 > ata2: SATA max UDMA/133 cmd 0xcc38 ctl 0xcc2c bmdma 0xcc48 irq 23 > > You might consider disabling the framebuffer temporarily, as it may impact > upon your ability to see what's going on early during the boot process. > > Cheers, > > --Kerin Hi Kerin, First, thanks for sticking with me on this. I really appreciate it. Second, I apologize for the length of the reply but it's still not working and I wanted to try and clearly show the steps I've taken. Maybe you or someone else will see the step I'm missing. OK, more and more it's looking to me like on the RAID I'm just not getting the /devmdX special file that has to be there to mount/bind/whatever the hardware to. What creates it, other than mknod? Can I do it by hand some other way? Read on... 1) Booted non-RAID and look in /dev for md*. md0 is there. I mount the RAID and look in the directory /mnt/gentoo/dev/ for anything named md*. There's nothing. 2) Booting with the Live CD I modprobe raid1, assemble the RAID, and then do the recommended chroot: livecd usr # cd / livecd / # mount -t proc proc /mnt/gentoo/proc livecd / # mount -o bind /dev /mnt/gentoo/dev livecd / # cp -L /etc/resolv.conf /mnt/gentoo/etc/ livecd / # chroot /mnt/gentoo /bin/bash livecd / # env-update && source /etc/profile 3) I look in /dev and still no md*. (There shouldn't be) 4) Because I used md3 to get into the RAID from the Live CD I create some new devices, but specifically don't do md3: livecd / # mknod /dev/md0 b 9 0 livecd / # mknod /dev/md1 b 9 1 livecd / # mknod /dev/md2 b 9 2 livecd / # mknod /dev/md4 b 9 4 livecd / # mknod /dev/md5 b 9 5 5) ls the directory and I see the ones I created plus md3 bound from the chroot above: livecd / # ls /dev/md* /dev/md0 /dev/md1 /dev/md2 /dev/md3 /dev/md4 /dev/md5 /dev/md: 3 6) I exit the chroot and look again at what's now in the Live CD /dev. As expected I still see all 6. livecd / # exit exit livecd ~ # ls /dev/md* /dev/md0 /dev/md1 /dev/md2 /dev/md3 /dev/md4 /dev/md5 /dev/md: 3 livecd ~ # df Filesystem 1K-blocks Used Available Use% Mounted on tmpfs 3052532 29600 3022932 1% / /dev/sr0117392117392 0 100% /mnt/cdrom /dev/loop0 87424 87424 0 100% /mnt/livecd udev 10240 216 10024 3% /dev tmpfs 3052532 6152 3046380 1% /mnt/livecd/lib64/firmware tmpfs 3052532 0 3052532 0% /mnt/livecd/usr/portage /dev/md3 30969528 2254188 27142180 8% /mnt/gentoo livecd ~ # 7) Try booting RAID a couple of times with different command line options. It still fails the same way. 8) Boot back to non-RAID and mount the RAID, this time as md0 just to be different. I look in /dev and see only md0, as expected: keeper ~ # df Filesystem 1K-blocks Used Available Use% Mounted on /dev/sda3 30969600 7244840 22151596 25% / udev 10240 228 10012 3% /dev shm3053512 0 3053512 0% /dev/shm /dev/md0 30969528 2254188 27142180 8% /mnt/gentoo keeper ~ # ls /dev/md* /dev/md0 /dev/md: 0 keeper ~ # 9) I look in /mnt/gentoo/dev to see if ANY md special files are there. There are none: keeper ~ # ls -al /mnt/gentoo/dev/md* ls: cannot access /mnt/gentoo/dev/md*: No such file or directory keeper ~ # All the other stuff is there, but not the special files I created while in the chroot. So, to me, this comes down to something like when I'm booting /dev/md0 or md3 isn't available so the kernel cannot mount the RAID anywhere. What creates the special file, at least by hand, is mknod, and for some reason I'm unable to make t
[gentoo-user] Re: How does grub assemble a RAID1 for / ??
On 04/04/2010 23:16, Mark Knecht wrote: First, thanks for sticking with me on this. I really appreciate it. Second, I apologize for the length of the reply but it's still not working and I wanted to try and clearly show the steps I've taken. Maybe you or someone else will see the step I'm missing. [snip] I don't know what to try next. OK, I think I now understand what's happening here. I regret that I did not recall this earlier but only the original RAID superblock format (version 0.90.00) is supported for automatic assembly! I have two servers that are set up in a similar way as your box, and they both use this format. It's possible that the docs may be out of date but /usr/src/linux/Documentation/md.txt says: "When md is compiled into the kernel (not as module), partitions of type 0xfd are scanned and automatically assembled into RAID arrays. This autodetection may be suppressed with the kernel parameter "raid=noautodetect". As of kernel 2.6.9, only drives with a type 0 superblock can be autodetected and run at boot time." Also, look at this: http://www.mail-archive.com/linux-r...@vger.kernel.org/msg06215.html To quote Neil Brown: "v0.90 can be used with 'in kernel autodetect' (i.e. partition type 0xfd). v1 cannot (I consider this an improvement :-)" Well, I can't say I agree with him there. Anyway, it seems that you're using the 1.1 superblock format. So, what options does this leave you with? a) Backup the root filesystem, and re-create the array with the regular superblock format. If necessary, coerce mdadm with -e 0 but it should be a default. b) Rely on userspace tools to assemble the array. This means either having the root filesystem off raid, or using an initrd/initramfs image. I'd got for the first option as it keeps things simple and the benefits of the v1 format are nebulous in practical terms. Cheers, --Kerin
Re: [gentoo-user] Re: How does grub assemble a RAID1 for / ??
On Sun, Apr 4, 2010 at 3:47 PM, Kerin Millar wrote: > On 04/04/2010 23:16, Mark Knecht wrote: >> >> First, thanks for sticking with me on this. I really appreciate it. >> Second, I apologize for the length of the reply but it's still not >> working and I wanted to try and clearly show the steps I've taken. >> Maybe you or someone else will see the step I'm missing. > > [snip] > >> I don't know what to try next. > > OK, I think I now understand what's happening here. I regret that I did not > recall this earlier but only the original RAID superblock format (version > 0.90.00) is supported for automatic assembly! I have two servers that are > set up in a similar way as your box, and they both use this format. It's > possible that the docs may be out of date but > /usr/src/linux/Documentation/md.txt says: > > "When md is compiled into the kernel (not as module), partitions of type > 0xfd are scanned and automatically assembled into RAID arrays. This > autodetection may be suppressed with the kernel parameter > "raid=noautodetect". As of kernel 2.6.9, only drives with a type 0 > superblock can be autodetected and run at boot time." > > Also, look at this: > > http://www.mail-archive.com/linux-r...@vger.kernel.org/msg06215.html > > To quote Neil Brown: > > "v0.90 can be used with 'in kernel autodetect' (i.e. partition type 0xfd). > v1 cannot (I consider this an improvement :-)" > > Well, I can't say I agree with him there. > > Anyway, it seems that you're using the 1.1 superblock format. So, what > options does this leave you with? > > a) Backup the root filesystem, and re-create the array with the regular > superblock format. If necessary, coerce mdadm with -e 0 but it should > be a default. > > b) Rely on userspace tools to assemble the array. This means either > having the root filesystem off raid, or using an initrd/initramfs > image. > > I'd got for the first option as it keeps things simple and the benefits of > the v1 format are nebulous in practical terms. > > Cheers, > > --Kerin > > > Hi, From my post this morning: "No problem supplying it. I did the rebuild this morning but forced metadata to Type 1.0. No change as you suggested there wouldn't be." I guess I didn't post it there but what I meant by that was the following: 1) If you don't specify metadata then you get the newest - I think that's currently ver. 1.2 or something. 2) I tried 1.0 this morning (shown below) which didn't fix it. (commands used are below) I will immediately try 0.90 as I have no problem with the limitations at this time: 0, 0.90 Use the original 0.90 format superblock. This format limits arrays to 28 component devices and limits component devices of levels 1 and greater to 2 terabytes. I should hopefully know in an hour or two how this worked. Thanks for the help! Cheers, Mark keeper ~ # mdadm --create /dev/md0 --level=1 --raid-devices=2 /dev/sdb3 /dev/sdc3 mdadm: Note: this array has metadata at the start and may not be suitable as a boot device. If you plan to store '/' or '/boot' on this device please ensure that your boot-loader understands md/v1.x metadata, or use --metadata=1.0 mdadm: Note: this array has metadata at the start and may not be suitable as a boot device. If you plan to store '/' or '/boot' on this device please ensure that your boot-loader understands md/v1.x metadata, or use --metadata=1.0 Continue creating array? n mdadm: create aborted. keeper ~ # mdadm --create /dev/md0 --level=1 --raid-devices=2 --metadata=1.0 /dev/sdb3 /dev/sdc3 mdadm: array /dev/md0 started. keeper ~ # cat /proc/mdstat Personalities : [raid1] md0 : active raid1 sdc3[1] sdb3[0] 31463228 blocks super 1.0 [2/2] [UU] [>] resync = 3.9% (1241664/31463228) finish=5.2min speed=95512K/sec unused devices: keeper ~ #
[gentoo-user] Re: How does grub assemble a RAID1 for / ??
On 05/04/2010 00:12, Mark Knecht wrote: 1) If you don't specify metadata then you get the newest - I think that's currently ver. 1.2 or something. Interesting. I suppose that might be a change in mdadm-3.0 (a version which I have yet to use to create any new arrays). However, that would contradict the man page which still says: "0, 0.90, default" 2) I tried 1.0 this morning (shown below) which didn't fix it. Right. Any version above in the 1 series (1, 1.0, 1.1, 1.2) will not work. I'm certain that reverting to the original format is going to resolve the issue and that we've just been barking up the wrong tree(s) hitherto. --Kerin
Re: [gentoo-user] Re: Moving the system from one disk to another
Dnia 2010-04-04, o godz. 21:04:03 Neil Bothwick napisał(a): > On Sun, 04 Apr 2010 20:35:11 +0100, Kerin Millar wrote: > > > Whichever way you go about it, ensure that no pseudo-filesystem or > > bind mounts are present within "/mnt/oldrootfs" at the time. > > Use the -x option with rsync to stop it descending into other > filesystems. > > AFAIK "mount --bind / /somewhere" and rsync'ing /somewhere/ instead of / would be more useful then "-x" option - stage1,2,3 has static /dev entries which should also be copied. Since udev mounts it with tmpfs, rsync with -x would skip those entries (static and from tmpfs). I suppose you can ignore static /dev if you use initrd. Since author of this thread wants to mount filesystem(s) of "the system" from livecd of some kind, there is no point in using any of ideas in this or previous email - there will be no other filesystems mounted. I often use that trick with /somewhere/ to backup live system from laptop to external drive. But it does not work well with innodb... man mount man rsync good luck -- Kacper Kopczyński
Re: [gentoo-user] Re: How does grub assemble a RAID1 for / ??
On Sun, Apr 4, 2010 at 4:51 PM, Kerin Millar wrote: > On 05/04/2010 00:12, Mark Knecht wrote: >> >> 1) If you don't specify metadata then you get the newest - I think >> that's currently ver. 1.2 or something. > > Interesting. I suppose that might be a change in mdadm-3.0 (a version which > I have yet to use to create any new arrays). However, that would contradict > the man page which still says: > > "0, 0.90, default" > >> 2) I tried 1.0 this morning (shown below) which didn't fix it. > > Right. Any version above in the 1 series (1, 1.0, 1.1, 1.2) will not work. > I'm certain that reverting to the original format is going to resolve the > issue and that we've just been barking up the wrong tree(s) hitherto. > > --Kerin I'm emerging gentoo-sources in the chroot now. One thing about this that still confuses me is where /dev/md3, or whatever, comes from when I boot if the the mknod command is never executed within the chrrot. (As per the install guide.) Not a big deal to proceed and see what happens. Maybe the kernel just creates it based on discovering the RAID? Or it makes it because I explicitly define it at the command line? As I say, no big deal to just push forward but that's still a question for me at this point. Cheers, Mark
[gentoo-user] Re: Moving the system from one disk to another
On 05/04/2010 00:51, Kacper Kopczyński wrote: Dnia 2010-04-04, o godz. 21:04:03 Neil Bothwick napisał(a): On Sun, 04 Apr 2010 20:35:11 +0100, Kerin Millar wrote: Whichever way you go about it, ensure that no pseudo-filesystem or bind mounts are present within "/mnt/oldrootfs" at the time. Use the -x option with rsync to stop it descending into other filesystems. AFAIK "mount --bind / /somewhere" and rsync'ing /somewhere/ instead of / would be more useful then "-x" option - stage1,2,3 has static /dev entries which should also be copied. Since udev mounts it with tmpfs, rsync with -x would skip those entries (static and from tmpfs). Well, no, because my response was based on the fact that the duplication will be carried out from an alternate environment provided by a CD/DVD, as Meino clearly stated in the original post. Thus, bind mounts, pseudo-filesystems and chroots need not come into the equation whatsoever. Indeed, it's the very same concern that you express which resulted in my recommendation to avoid such shenanigans and keep it simple. Ergo, just mount the root filesystem - nothing else - and copy it as-is. Static /dev entries would be copied without issue, as would everything else. It really couldn't be simpler. You post hinges on the notion that he would be performing the process while booted from the system he is duplicating, in which case your advice would, of course, be entirely sensible. Ergo, he would indeed be best advised to bind mount / to a temporary directory and use that as the source for the exact reasons that you mention. I personally would not recommend doing it under these circumstances but it can certainly be done (though I'd suggest dropping to runlevel 1 first). Cheers, --Kerin
[gentoo-user] Re: How does grub assemble a RAID1 for / ??
On 05/04/2010 00:58, Mark Knecht wrote: One thing about this that still confuses me is where /dev/md3, or whatever, comes from when I boot if the the mknod command is never executed within the chrrot. (As per the install guide.) Not a big deal to proceed and see what happens. Maybe the kernel just creates it based on discovering the RAID? Or it makes it because I explicitly define it at the command line? Well, it shouldn't matter. The md block device will be initialised by the kernel as soon as it loads, whereupon it will be resolved internally by its registered major/minor numbers for the purpose of mounting the root filesystem (that's the "9, 0" that you referred to earlier in the thread). Once the root filesystem is mounted read-only, it proceeds to load init. Later, the device node will need to be present - for instance, when fstab is parsed - but udev will have taken care of it by that time. That is, udev will manifest the device node in a tmpfs filesystem which is mounted at /dev and dynamically populated early on during the boot process. Cheers, --Kerin
Re: [gentoo-user] Re: How does grub assemble a RAID1 for / ??
On Sun, Apr 4, 2010 at 4:58 PM, Mark Knecht wrote: > On Sun, Apr 4, 2010 at 4:51 PM, Kerin Millar wrote: >> On 05/04/2010 00:12, Mark Knecht wrote: >>> >>> 1) If you don't specify metadata then you get the newest - I think >>> that's currently ver. 1.2 or something. >> >> Interesting. I suppose that might be a change in mdadm-3.0 (a version which >> I have yet to use to create any new arrays). However, that would contradict >> the man page which still says: >> >> "0, 0.90, default" >> >>> 2) I tried 1.0 this morning (shown below) which didn't fix it. >> >> Right. Any version above in the 1 series (1, 1.0, 1.1, 1.2) will not work. >> I'm certain that reverting to the original format is going to resolve the >> issue and that we've just been barking up the wrong tree(s) hitherto. >> >> --Kerin > > I'm emerging gentoo-sources in the chroot now. > > One thing about this that still confuses me is where /dev/md3, or > whatever, comes from when I boot if the the mknod command is never > executed within the chrrot. (As per the install guide.) Not a big deal > to proceed and see what happens. Maybe the kernel just creates it > based on discovering the RAID? Or it makes it because I explicitly > define it at the command line? > > As I say, no big deal to just push forward but that's still a question > for me at this point. > > Cheers, > Mark > OK, I'm up and running now. Using --metadata=0.90 when first creating the RAID was the solution. It seems that md0 is (I guess) there by default and then md3 gets created by the kernel extra command line stuff I guess: k2 ~ # ls /dev/md* /dev/md0 /dev/md3 /dev/md: 0 3 3_0 k2 ~ # localhost ~ # df Filesystem 1K-blocks Used Available Use% Mounted on /dev/md3 41294780 2258892 36938208 6% / udev 10240 232 10008 3% /dev shm3053480 0 3053480 0% /dev/shm Thanks for all the help today Kerin. I really appreciate it! Cheers, Mark
[gentoo-user] Re: How does grub assemble a RAID1 for / ??
On 04/04/2010 07:20 AM, Mark Knecht wrote: keeper ~ # diff /usr/src/linux/.config /mnt/gentoo/usr/src/linux/.config 4c4 < # Mon Mar 29 01:02:31 2010 --- # Sun Apr 4 06:28:53 2010 893,912c893,906 < CONFIG_MD_LINEAR=m < CONFIG_MD_RAID0=m Hi Mark, Interesting thread, and I'm learning a lot from following along. I know very little about raid so I can benefit from nearly any question. Just one suggestion about diffs, though. Most people who post and/or test patches will expect the 'unified' diff format, which is generated by using 'diff -u'. Much easier for humans to parse.
Re: [gentoo-user] Re: How does grub assemble a RAID1 for / ??
On Sun, Apr 4, 2010 at 5:47 PM, walt wrote: > On 04/04/2010 07:20 AM, Mark Knecht wrote: >> >> keeper ~ # diff /usr/src/linux/.config /mnt/gentoo/usr/src/linux/.config >> 4c4 >> < # Mon Mar 29 01:02:31 2010 >> --- >>> >>> # Sun Apr 4 06:28:53 2010 >> >> 893,912c893,906 >> < CONFIG_MD_LINEAR=m >> < CONFIG_MD_RAID0=m > > Hi Mark, > > Interesting thread, and I'm learning a lot from following along. I know > very little about raid so I can benefit from nearly any question. > > Just one suggestion about diffs, though. Most people who post and/or test > patches will expect the 'unified' diff format, which is generated by using > 'diff -u'. Much easier for humans to parse. > Walt, Good to know about the diff stuff. I hope I can remember that next time. Glad if this thread can help others, either now or in the future. I do intend to follow up with the mdadm guys about the actual metadata requirements when I get a chance. I'm just finishing emerge -DuN @world on the new RAID build. Fast with an 8 core processor and Enterprise class RAID drives. Can hardly wait for the big 12 core machine with two RAIDs in it to start running. Cheers, Mark
[gentoo-user] Re: alsa-update not in sync with kernel version
On 04/04/2010 11:42 AM, meino.cra...@gmx.de wrote: Hi, I am running a vanilla kernel (linux-2.6.32.11), This kernel uses alsa-1.0.21. When doing the eix-sync-thingy, emerge always suggests to update to alsa-1.0.22. Do you have media-sound/alsa-driver installed? If so, that is the cause of your problem -- just emerge -C alsa-driver. That package is only for people who are testing/debugging alsa-drivers (or using prehistoric kernels on new hardware).
[gentoo-user] Re: Moving the system from one disk to another
On 04/04/2010 01:04 PM, Neil Bothwick wrote: On Sun, 04 Apr 2010 20:35:11 +0100, Kerin Millar wrote: Whichever way you go about it, ensure that no pseudo-filesystem or bind mounts are present within "/mnt/oldrootfs" at the time. Use the -x option with rsync to stop it descending into other filesystems. This is directed to all of you gurus who replied to Meino's post: Meino's unstated assumption is that his new (larger) disk is already formatted (possibly partitioned?) before he copies the existing filesystem to it. IIUC the new disk will then be unbootable until grub or equivalent is installed on the new disk. Does this seem correct, or not? My instinct is to use dd to duplicate the entire old disk to the new (unformatted) disk and then use gparted to twiddle it from there. (But I do love a puzzle ;o)
[gentoo-user] Re: Moving the system from one disk to another
On 05/04/2010 02:34, walt wrote: On 04/04/2010 01:04 PM, Neil Bothwick wrote: On Sun, 04 Apr 2010 20:35:11 +0100, Kerin Millar wrote: Whichever way you go about it, ensure that no pseudo-filesystem or bind mounts are present within "/mnt/oldrootfs" at the time. Use the -x option with rsync to stop it descending into other filesystems. This is directed to all of you gurus who replied to Meino's post: Meino's unstated assumption is that his new (larger) disk is already formatted (possibly partitioned?) before he copies the existing filesystem to it. IIUC the new disk will then be unbootable until grub or equivalent is installed on the new disk. Does this seem correct, or not? Absolutely correct. Two commands from the grub shell and job done :) My instinct is to use dd to duplicate the entire old disk to the new (unformatted) disk and then use gparted to twiddle it from there. (But I do love a puzzle ;o) In general, I'm a proponent of copying filesystems, as opposed to copying entire block devices or disks. That's not to say that there aren't some situations where the latter approach makes sense. Note that it's possible to copy just the portion of the first sector that contains bootloader code as thus: dd if=/dev/sda of=/dev/sdb bs=446 count=1 NOTE: that's 446 as opposed to 512 as the latter would result in the partition table being copied too, which would be most undesirable. Cheers, --Kerin
[gentoo-user] ~~Hi~~
http://sites.google.com/site/gni8hy9ojm/icfa5w
Re: [gentoo-user] alsa-update not in sync with kernel version
Neil Bothwick [10-04-05 01:32]: > On Sun, 4 Apr 2010 20:42:48 +0200, meino.cra...@gmx.de wrote: > > > I am running a vanilla kernel (linux-2.6.32.11), This kernel uses > > alsa-1.0.21. > > > > When doing the eix-sync-thingy, emerge always suggests to update > > to alsa-1.0.22. > > The 1.0.22 userspace tools work fine with the drivers in 2.6.32. Nice to hear, Neil, but is this also true for other combinations of versions in the near or fat future? That was the reason of my question... mcc > -- > Neil Bothwick > > I have seen things you lusers would not believe. > I've seen Sun monitors on fire off the side of the multimedia lab. > I've seen NTU lights glitter in the dark near the Mail Gate. > All these things will be lost in time, like the root partition last > week. Time to die. -- Please don't send me any Word- or Powerpoint-Attachments unless it's absolutely neccessary. - Send simply Text. See http://www.gnu.org/philosophy/no-word-attachments.html In a world without fences and walls nobody needs gates and windows.
Re: [gentoo-user] Re: alsa-update not in sync with kernel version
walt [10-04-05 05:02]: > On 04/04/2010 11:42 AM, meino.cra...@gmx.de wrote: > > > >Hi, > > > >I am running a vanilla kernel (linux-2.6.32.11), This kernel uses > >alsa-1.0.21. > > > >When doing the eix-sync-thingy, emerge always suggests to update > >to alsa-1.0.22. > > Do you have media-sound/alsa-driver installed? If so, that is the > cause of your problem -- just emerge -C alsa-driver. That package > is only for people who are testing/debugging alsa-drivers (or using > prehistoric kernels on new hardware). > > Hi, what's about people who compile software which needs the alsa-stuff, for example libraries and headers? mcc -- Please don't send me any Word- or Powerpoint-Attachments unless it's absolutely neccessary. - Send simply Text. See http://www.gnu.org/philosophy/no-word-attachments.html In a world without fences and walls nobody needs gates and windows.
[gentoo-user] ZSH: Gentoo-completion...how to modifiy?
Hi, adding Gentoo/emerge related stuff to the completion system of the zsh is nice...but getting a dark blue color for parts of the prompt with that is not. Where can I tuirn what to modifiy the color or get back my previous prompt? Best regards, mcc -- Please don't send me any Word- or Powerpoint-Attachments unless it's absolutely neccessary. - Send simply Text. See http://www.gnu.org/philosophy/no-word-attachments.html In a world without fences and walls nobody needs gates and windows.
