Re: mkraid problems (v0.90/2.2.10/RedHat)
My /etc/raidtab looks like this: raiddev /dev/md0 raid-level 1 nr-raid-disks 3 nr-spare-disks 0 chunk-size 4 device /dev/sda2 raid-disk 0 device /dev/sdb2 raid-disk 1 device /dev/sdc2 raid-disk 2 raiddev /dev/md1 raid-level 1 nr-raid-disks 3 nr-spare-disks 0 chunk-size 4 device /dev/sda4 raid-disk 0 device /dev/sdb4 raid-disk 1 device /dev/sdc4 raid-disk 2 raid-level 1 with 3 disks? You have to use raid-level 5 (or 4). -Felix -- Felix Egli | E-Mail: [EMAIL PROTECTED] Communication Systems Inc. AG | Phone: +41-1-926 61 42 GoldNet | Fax:+41-1-926 61 45 Grundstrasse 66, CH-8712 Stäfa, Switzerland
linux-raid 0.9 on SUSE-Linux 6.1
Hello everybody, I'm using SUSE Linux 6.1 with kernel 2.2.5 Shipped with SUSE is mdtools 0.42. So I was loadding the 0.90 rpm packet. I installed it an by calling any raid-tool I get a segmentation fault. Is there anybody who managed tho problem and could provide me any help ? Thx Barney
RE: raidtools compilation problem
I run Redhat 6.0/kernel 2.2.5-15 and installed raidtools-0.50b. I believe RH6.0 comes with raidtools-0.90, so there is no point trying to install raidtools-0.50b. Just setup your /etc/raidtab for the configuration you want, and then run mkraid. For rebooting, you may want to include raid support in your initrd. Do "mkinitrd --with raid0 /boot/initrd 2.2.5-15", replacing raid0 with whatever raid level you need. Then edit /etc/lilo.conf to make sure the initrd line is pointing to /boot/initrd. Now run lilo to make sure you use the new initrd from the next reboot. As long as you don't want / on raid, that's all you should need to do. See the Software-RAID HOWTO in the raidtools package or at http://ostenfeld.dk/~jakob/Software-RAID.HOWTO/ for more details. p.s. Where can I get the higher version of raidtools than 0.50b? http://www.country.kernel.org/pub/linux/daemons/raid/alpha Cheers, Bruno Prior [EMAIL PROTECTED]
RE: mkraid problems (v0.90/2.2.10/RedHat)
raid-level 1 with 3 disks? You have to use raid-level 5 (or 4). No you don't. You can have as many mirrors in a RAID-1 mirror set as you want. The setup described will protect against the simultaneous failure of two disks (i.e. the failure of a second disk before you are able to replace the first). But you could argue the case that this was an unnecessary level of paranoia. It depends on the circumstances. Cheers, Bruno Prior [EMAIL PROTECTED]
RE: mkraid problems (v0.90/2.2.10/RedHat)
I can't figure out why mkraid is aborting. No messages show up in the syslog, and I get what looks like a typical response from /proc/mdstat You are missing persistent-superblocks lines in your raidtab. I thought this shouldn't matter, as it should default to "persistent-superblocks 1", but mkraid is very sensitive to the exact order of items in raidtab, so maybe try putting it in to see if that helps. Also, did you follow the advice given recently on this list about using kernel versions 2.2.7 with raid? Check the archives (http://www.kernelnotes.org/lnxlists/linux-raid/) for the last couple of weeks for a lot of advice about the complications with using 2.2.8+ kernels. Cheers, Bruno Prior [EMAIL PROTECTED]
RE: linux-raid 0.9 on SUSE-Linux 6.1
I'm using SUSE Linux 6.1 with kernel 2.2.5 Shipped with SUSE is mdtools 0.42. So I was loadding the 0.90 rpm packet. I installed it an by calling any raid-tool I get a segmentation fault. mdtools 0.42 doesn't require a kernel patch to work, so I assume the SuSE kernel doesn't include this patch. You need to rebuild the kernel with the raid0145 patch which you can get from http://www.country.kernel.org/pub/linux/daemons/raid/alpha. You might be better off getting the matching raidtools from there as well, instead of using the rpm. Alternatively, you could use mdtools 0.42 or raidtools 0.50 and not bother with the patch. But you would probably be better off with the latest raid support. Cheers, Bruno Prior [EMAIL PROTECTED]
FAQ
It strikes me that this list desperately needs a FAQ. I'm off on holiday for the next two weeks, but unless someone else wants to volunteer, I'm willing to put one together when I get back. If people would like me to do this, I would welcome suggestions for questions to go in the FAQ. Cheers, Bruno Prior [EMAIL PROTECTED]
RE: Problem implementing raid-1 filesystems
Here's your problem: request_module[md-personality-3]: Root fs not mounted It looks like you are using a RedHat stock kernel, not one you built yourself. RedHat's kernels use modular raid support, not built-in. The modules sit in /lib/modules which is on root and, in your case therefore, on /dev/md8. So when the kernel tries to start the raid arrays, it looks for the module to provide raid support, but can't find it because it is on a raid device, which can't be started because the module to support it isn't available. Catch-22. There are two solutions. Build raid support into a new kernel (more stable). Or use an initrd that includes raid support. To do this, do "mkinitrd --with raid1 /boot/initrd 2.2.5-15". Then edit /etc/lilo.conf to add an initrd=/boot/initrd for the relevant image section. Now run lilo to make sure you use the new initrd from the next reboot. I think the messages in syslog are from when you mkraid'ed the arrays, not from the reboot (hence the time discrepancy). I still think this system is too "belt-and-braces" and an unnecessary waste of disk space, by the way. Cheers, Bruno Prior [EMAIL PROTECTED]
Re: FAQ
Bruno Prior wrote:
It strikes me that this list desperately needs a FAQ. I'm off on holiday for the
next two weeks, but unless someone else wants to volunteer, I'm willing to put
one together when I get back. If people would like me to do this, I would
welcome suggestions for questions to go in the FAQ.
Whoever volunteers: The first answer should summarize which version of
{md,raid}tools works with which kernel patched with{,out} patch XY.
Can't think of an question for that, though.
IMO it is very necessay to clear the fog that has laid itself across
raid-with-linux in the last weeks or so.
Marc
--
Marc Mutz [EMAIL PROTECTED]http://marc.mutz.com/
University of Bielefeld, Dep. of Mathematics / Dep. of Physics
PGP-keyID's: 0xd46ce9ab (RSA), 0x7ae55b9e (DSS/DH)
Resync Priority.
Quick question concerning resyncing / regeneration... I have 4 9.1 U2W SCSI drives each one with 6 partitions, in a setup as follows /dev/md0 (sda2, sdb2, sdc2, sdd2) /dev/md1 (sda3, sdb3, sdc3, sdd3) /dev/md2 (sda5, sdb5, sdc5, sdd5) /dev/md3 (sda6, sdb6, sdc6, sdd6) /dev/md4 (sda7, sdb7, sdc7, sdd7) plus the swap partitions.. sda1, sdb1, sdc1, sdd1... Is there anyway to set a preference on which md devices to rebuild first in the event of a failure? I have been testing with the odd powerfailure scenario and it seems for me that when the system comes back up, for me anyway, the system tries to resync /dev/md2, which for me is a massive database partition which takes over half-an-hour to regenerate. For me it would be more important to check the stability of the OS and make sure the system comes back online properly first, then goes back to recover the /dev/md2. Any ideas? Regards, Matt.C. -Systems Admin Alliance
Re: FAQ
On Fri, 9 Jul 1999, Marc Mutz wrote:
Bruno Prior wrote:
It strikes me that this list desperately needs a FAQ. I'm off on holiday for the
next two weeks, but unless someone else wants to volunteer, I'm willing to put
one together when I get back. If people would like me to do this, I would
welcome suggestions for questions to go in the FAQ.
Whoever volunteers: The first answer should summarize which version of
{md,raid}tools works with which kernel patched with{,out} patch XY.
Can't think of an question for that, though.
Hrm. "I have kernel X-Version and I'm trying to use raidtools Y-Version,
but it isn't working. What's up with that?" Maybe also include a few
more subquestions like "I'm using the version of raidtools that came with
my Linux distro, but it doesn't seem to be working"
IMO it is very necessay to clear the fog that has laid itself across
raid-with-linux in the last weeks or so.
Having inherited the responsibility of maintaining a kernel RPM after
2.2.8 and discovering that software raid stuff no longer worked, I'd
have to agree. ;-)
--
Kelley Spoon [EMAIL PROTECTED]
Sic Semper Tyrannis.
Re: Resync Priority.
Although your point is very valid, the stability of the filesystem is not in question although the array is running in "degraded" mode, the filesystem stability is assured by fsck rather than the resync process surely your data is more important than your OS. If a second disk were to fail before the resync was complete then only resynced data would be recoverable... so the OS is probably least important... although it should be small enough that resyncing it is trivial I'd be interested to know what answer you get... mabe it's based on the device numbers although the fact that it starts on MD2 suggests that this is not the case I think the most likley explanation is that it is related to the order in which the devices were either created or detected... if detected then it's probably the device which had the most recent timestamp in it's superblock which is probably the one which took longest to unmount. James On Fri, 9 Jul 1999, Matt Coene wrote: Quick question concerning resyncing / regeneration... I have 4 9.1 U2W SCSI drives each one with 6 partitions, in a setup as follows /dev/md0 (sda2, sdb2, sdc2, sdd2) /dev/md1 (sda3, sdb3, sdc3, sdd3) /dev/md2 (sda5, sdb5, sdc5, sdd5) /dev/md3 (sda6, sdb6, sdc6, sdd6) /dev/md4 (sda7, sdb7, sdc7, sdd7) plus the swap partitions.. sda1, sdb1, sdc1, sdd1... Is there anyway to set a preference on which md devices to rebuild first in the event of a failure? I have been testing with the odd powerfailure scenario and it seems for me that when the system comes back up, for me anyway, the system tries to resync /dev/md2, which for me is a massive database partition which takes over half-an-hour to regenerate. For me it would be more important to check the stability of the OS and make sure the system comes back online properly first, then goes back to recover the /dev/md2. Any ideas? Regards, Matt.C. -Systems Admin Alliance A.J. ([EMAIL PROTECTED]) Sometimes you're ahead, somtimes you're behind. The race is long, and in the end it's only with yourself.
Linux root mirror recipie (repost)
I posted this a while back, I think some people found it usefull... so here goes again Since I wrote this however I have decided that it makes sense if possible to use device numbers for raid that are the same as the partitions they are built from for ease of maintainance! So that a mirror built from sda2 and sdb2 would be md2 regardless of if that isn't the next available md number of course this isn't always going to be possible but it makes things simpler if you can. A.J. ([EMAIL PROTECTED]) Sometimes you're ahead, somtimes you're behind. The race is long, and in the end it's only with yourself. -- Forwarded message -- Date: Fri, 09 Jul 1999 12:49:10 +0100 From: A James Lewis [EMAIL PROTECTED] To: [EMAIL PROTECTED] Subject: Linux root mirror This guide assumes that there are two disks hda hdc, and that "/" is /dev/hda2 It also assumes that /boot (The kernel) is NOT on /. /boot requires special handling but I suggest simply maintaining a copy on both disks until LILO is raid aware. This example was done with Martin Bene's failed disk patch and raidtools snapshot 19990421 (+kernel patches) and Linux 2.2.6. All of which must be installed before you start! 1. Create "raidtab" file. (Note deliberately failed disk = mounted "/") raiddev /dev/md0 raid-level1 chunk-size64 nr-raid-disks 2 nr-spare-disks0 device/dev/hdc2 raid-disk 0 device/dev/hda2 failed-disk 1 2. run mkraid. DESTROYING the contents of /dev/md0 in 5 seconds, Ctrl-C if unsure! handling MD device /dev/md0 analyzing super-block disk 0: /dev/hdc2, 700560kB, raid superblock at 700480kB disk 1: /dev/hda2, failed 3. mke2fs the new array mke2fs 1.14, 9-Jan-1999 for EXT2 FS 0.5b, 95/08/09 Linux ext2 filesystem format Filesystem label= 175440 inodes, 700480 blocks 35024 blocks (5.00%) reserved for the super user First data block=1 Block size=1024 (log=0) Fragment size=1024 (log=0) 86 block groups 8192 blocks per group, 8192 fragments per group 2040 inodes per group Superblock backups stored on blocks: 8193, 16385, 24577, 32769, 40961, 49153, 57345, 65537, 73729, 81921, 90113, 98305, 106497, 114689, 122881, 131073, 139265, 147457, 155649, 163841, 172033, 180225, 188417, 196609, 204801, 212993, 221185, 229377, 237569, 245761, 253953, 262145, 270337, 278529, 286721, 294913, 303105, 311297, 319489, 327681, 335873, 344065, 352257, 360449, 368641, 376833, 385025, 393217, 401409, 409601, 417793, 425985, 434177, 442369, 450561, 458753, 466945, 475137, 483329, 491521, 499713, 507905, 516097, 524289, 532481, 540673, 548865, 557057, 565249, 573441, 581633, 589825, 598017, 606209, 614401, 622593, 630785, 638977, 647169, 655361, 663553, 671745, 679937, 688129, 696321 Writing inode tables: done Writing superblocks and filesystem accounting information: done 4. Change root to /dev/md0 in /etc/fstab. 5. Copy filesystem to new array, mounted on /mnt. moo:/# for i in `ls | grep -v mnt | grep -v proc` do cp -dpR $i /mnt done 6. Edit /etc/lilo.conf on old copy to spec "root = /dev/md0" (The new array) 7. Run lilo to reinstall the bootblock moo:/# lilo Added Linux * moo:/# 8. Use fdisk to mark the new partition as type "0xfd" so it can autostart. 9. Reboot 10. Pray, time passes. 11. run "df" to ensure that your new array is indeed mounted as "/" 12. edit lilo.conf (You have an old copy now!) to refer to your array. (You could have edited this before booting but I think it's better to do it now since you are making it look like your system does now, not what you think it might look like in the future You can predict the future if you like once you become an expert at this.) 13. run lilo again to reinstall the bootblock, ensure it doesn't error!! If you get an error return to step 11. 14. mkdir /proc /mnt (You excluded these earlier to avoid recursive copy) 15. mount /proc 16. raidhotadd /dev/md0 /dev/hda2 (Add your old root volume to the array... Like ODS/SDS metareplace. 17. cat /proc/mdstat and see your reconstruction in progress... moo:~# more /proc/mdstat Personalities : [raid0] [raid1] [raid5] read_ahead 1024 sectors md0 : active raid1 hda2[2] hdc2[0] 700480 blocks [2/1] [U_] recovery=2% finish=17.3min unused devices: none 18. Use fdisk to mark old "/" partition as type "0xfd" so it can also autostart. 19. Wait until resync complete and reboot. 20. Congratulations you now have a mirrored "/" Now think about swap and /boot...
Re: mkraid problems (v0.90/2.2.10/RedHat)
On Thu, Jul 08, 1999 at 01:00:42PM -0700, Zack Hobson wrote: Hello RAID hackers, RedHat 6.0 w/ 2.2.10 kernel (compiled with RAID-1 support) raidtools 0.90 compiled from distributed source (ie, non-RedHat) raidtools 0.90 does not work with stock kernels either you use old raidtools or you patch the kernel But... kernel patches do not work with kernels 2.2.7 (yes i know you can modify a couple of lines, but even if i do it on my own boxes, i will not suggest anybody to take the risk) besides that 2.2.10 has the infamous fs corruption bug. i'd suggest downgrading to the last 'stable' redhat kernel. Luca P.S. (for Felix) you can build an N-way mirror (n2) for added redundancy, also please quote only relevant parts of a message when replying. -- Luca Berra -- [EMAIL PROTECTED] Communications Media Services S.r.l.
Re: 2.2.9 Patch
MadHat wrote: I did all the fixes you have been talking about and made a newer patch. I just used it on a new system and it compiles and works for me. http://www.unspecific.com/pub/linux/raid/raid-0145-2.2.9-patch.gz Same patch works for 2.2.10 I will not guarantee anything. This is just what has been discussed. I am running this on a few of my boxes and will report any problems I see. -- Lee Heath (aka MadHat) "The purpose of life is not to see how long you can keep doing it. It's to see how well you can do it. There are things worse than death, and chief among them is living without honor." -- Woodchuck, DC-stuff list
[Fwd: backup/redundancy solutions]
"Jonathan F. Dill" wrote: Gordon Henderson wrote: So does no-one apart from me use Amanda? I've been using it for many years or different systems and it's never let me down. Emails me every day with a report and to remind me to change tapes if needed. I have 2 main servers here, one running Solaris, the other Linux, both with DLT tapes (The Solaris has a DLT4000 - 40GB compressed, the Linux I use AMANDA--that is the "nightly backup" that I mentioned in my other post. I have about 70 SGI workstations with about 500 GB disk space of which about 300 GB is actually used. The capacity of the 10h works out to about 126 GB using AMANDA gzip compression and the 8505-XL hardware compression turned off (I tried with hw compression turned on and I actually got LESS data on each tape since the data was already gzipped). AMANDA takes care of figuring out what level backup to do when, and once you've got the cycle set up, which tape to reuse when, so I can actually keep all 300 GB backed up nightly on the "puny" 10h, with level-0 backups at least once a week. Recently, I acquired an Exabyte Mammoth to add to the setup, but I haven't received the tapes that I ordered for it yet. The AMANDA check program runs once a day via a cron job and sends me e-mail if there are any problems that I need to fix before the run, otherwise it's completely automated. There is no bullshit about trying to figure out what time to schedule what backup, or having to type in any commands to do it, because all of that is figured out and optimized by the AMANDA scheduler. -- "Jonathan F. Dill" ([EMAIL PROTECTED]) CARB Systems and Network Administrator Home Page: http://www.umbi.umd.edu/~dill -- "Jonathan F. Dill" ([EMAIL PROTECTED]) CARB Systems and Network Administrator Home Page: http://www.umbi.umd.edu/~dill
Re: FAQ
These questions are from the point of view of 0.90 or higher (i.e. RH 6.0). - How do you recover a RAID1 or a RAID5 with a bad disk when you have no spares, i.e. how do you hotremove and hotadd? Please go through it step by step because many paths seem to lead to hangs. - How do you recover a RAID1 or a RAID5 when you do have a spare? Does it or can it work automatically? - How do you keep your raidtab file sync'd with your actual RAID when persistent-superblock is 1? Is there a translator for the numerical values, i.e. for parity-algorithm, found in /proc/mdstat? Thanks, Larry Dickson Land-5 Corporation At 12:06 PM 7/9/99 +0100, you wrote: It strikes me that this list desperately needs a FAQ. I'm off on holiday for the next two weeks, but unless someone else wants to volunteer, I'm willing to put one together when I get back. If people would like me to do this, I would welcome suggestions for questions to go in the FAQ. Cheers, Bruno Prior [EMAIL PROTECTED]
Re: linux-raid 0.9 on SUSE-Linux 6.1
I installed the raid 0.90 tools and installed and patch the 2.2.6 kernel. Though I did it from source .. don't know if that matters. Corse .. getting Yast to believe it's root fs was /dev/md0 is a completely different storry. Schackel, Fa. Integrata, ZRZ DA wrote: Hello everybody, I'm using SUSE Linux 6.1 with kernel 2.2.5 Shipped with SUSE is mdtools 0.42. So I was loadding the 0.90 rpm packet. I installed it an by calling any raid-tool I get a segmentation fault. Is there anybody who managed tho problem and could provide me any help ? Thx Barney
Sftwr Raid w/Redhat 6.0
I'm trying to setup Raid 1 in sftwr using Redhat 6.0 w/2 identical EIDE drives (13GB) as hda hdc, but can't get it to work. Is there any definitive doc somewhere that describes how to do this w/the current kernel? Thanks . . . jim
Re: Sftwr Raid w/Redhat 6.0
hi jim I'm trying to setup Raid 1 in sftwr using Redhat 6.0 w/2 identical EIDE drives (13GB) as hda hdc, but can't get it to work. Is there any definitive doc somewhere that describes how to do this w/the current kernel? http://ostenfeld.dk/~jakob/Software-RAID.HOWTO I use linux-2.2.6 or 2.2.10 for raid0/1... w/ rh-6.0 have fun alvin http://www.Linux-Consulting.com/Raid/Docs - have running debian-2.2, rh-6.0, slack-4.0, suse-6.1, cald-2.2, pht, turbocluster soon
raid-problems after sync
hello after a reboot (caused by a power fail) my raid was checked with ckraid and brought back into sync, but e2fsck sais, that the md-device-partition has zero length?? the problem is, that my /usr /home and /var on the md-device resist i used the md-tools from debian 2.0 (with a self build 2.2.6-ac2 (the crash happened before switch to 2.2.7 with raidtoos 0.90 from april) any hints? peter
Re: A view from the other side...
On Thu, 8 Jul 1999, Brian Leeper wrote: If the drives are the same size, the following command works very well to copy a partition table from one to the other: dd if=/dev/sda of=/dev/sdb bs=1024 count=5 I am curious to know if this can help in creating a RAID1 mirror from an existing single disk setup? I would like to mirror the drive on my server without having to reinstall anything. About the only way I know of to do this, which I haven't tried and don't know if it works, is to create the mirror in degraded mode and the copy the data from your existing drive to it, then add your existing drive to the mirror and resync. I'm sure someone else has a better idea about this than I do.. Brian
