-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Monday 2006-01-02 at 11:20 -0000, Con Hennessy wrote:
Ok, it seems you have a raid device directly formatted as reiserfs.
/dev/md0 /raid0 reiserfs acl,user_xattr 1 2
Can you mount/umount it manually? No, I still get the error : # mount /raid0 mount: wrong fs type, bad option, bad superblock on /dev/md0, missing codepage or other error (could this be the IDE device where you in fact use ide-scsi so that sr0 or sda or so is needed?) In some cases useful info is found in syslog - try dmesg | tail or so And dmesg shows me : ReiserFS: md0: warning: sh-2006: read_super_block: bread failed (dev md0, block 2, size 4096) ReiserFS: md0: warning: sh-2006: read_super_block: bread failed (dev md0, block 16, size 4096) ReiserFS: md0: warning: sh-2021: reiserfs_fill_super: can not find reiserfs on md0
It seems there is no fs. Perhaps you need to recreate the reiserfs on /dev/md0 - I trust you had no data on it yet. [...] No, keep reading.
If there are problems, run a fsck on it while umounted. I can do this for both partitions involved in the raid, and no problems were detected.
No, you have to do that on /dev/md0 directly, not the components. If you do on the components, and something is written, they get out of sync, the array is broken.
I then tried to mount one of the partitions but it failed : # mkdir /tmp/xx # mount /dev/hda6 /tmp/xx mount: /dev/hda6 already mounted or /tmp/xx busy
You can not do that while /dev/md0 is active.
If the md0 has problems being activated during boot up, it could be that you need to activate some modules in initrd - but as your raid seems to be a data disk not needed for booting up, it doesn't matter, it will activate just a little bit later during boot.
Then, if it will not mount automatically during boot, you can configure it "noauto" in fstab, so that you can manually mount it later. I also tried this, but the kernel still insists on trying to mount it during boot. The only way I could find to disable that was to comment out the line in fstab :(
I can understand the kernel activating /dev/md0, but not mounting it. What the kernel would try to do is to fsck it automatically during boot, even if it doesn't want to mount it. If you want to avoid that step, set the last digit to '0' in the fstab line. I think, on a second read of your symptoms, that that is your problem.
I'm just guessing, but it seems you don't have problems with the raid per se, but with the reiserfs on it. Treat it as any other reiserfs problem.
I'm really not 100% sure about that. My /proc/mdstat shows me : Personalities : [raid1] md0 : inactive hda6[0] hdc6[1] 16653824 blocks
Inactive? Try "mdadm --detail /dev/md0", it gives more detailed info. If it is inactive, it can not be fscked, nor mounted. You will have to activate it first, something went wrong.
But when I look at the size of the partitions I see : /dev/hdc6 76975 93411 8284216+ fd Linux raid autodetect and /dev/hda6 2600 3641 8369833+ fd Linux raid autodetect
From my understanding of raid, the maximum blocks shown by /proc/mdstat should be 8284216 and not nearly twice that !
Unless it were raid 0... or different block size, they don't have to be the same. Copy here all messages related to the raid device during boot; for example, the output of: grep -i -e "raid\|md0\|md:" /var/log/boot.msg Mine shows (suse 9.3): <6>md: md driver 0.90.1 MAX_MD_DEVS=256, MD_SB_DISKS=27 <6>md: Autodetecting RAID arrays. <6>md: autorun ... <6>md: considering hdd13 ... <6>md: adding hdd13 ... <6>md: adding hdb11 ... <6>md: adding hda11 ... <6>md: created md0 <6>md: bind<hda11> <6>md: bind<hdb11> <6>md: bind<hdd13> <6>md: running: <hdd13><hdb11><hda11> <6>md: raid1 personality registered as nr 3 <6>raid1: raid set md0 active with 2 out of 2 mirrors <6>md: ... autorun DONE. <5>ReiserFS: md0: found reiserfs format "3.5" with standard journal <5>ReiserFS: md0: using ordered data mode <5>ReiserFS: md0: journal params: device md0, size 8192, journal first block 18, max trans len 1024, max batch 900, max commit age 30, max trans age 30 <5>ReiserFS: md0: checking transaction log (md0) <5>ReiserFS: md0: Using r5 hash to sort names <5>ReiserFS: md0: using 3.5.x disk format Have a read at /usr/share/doc/howto/en/txt/Software-RAID-HOWTO.gz - -- Cheers, Carlos Robinson -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.0 (GNU/Linux) Comment: Made with pgp4pine 1.76 iD8DBQFDuXz8tTMYHG2NR9URAkewAJwPbu0FV8Hi0LBecv1Ro9+LJwVKwwCffyaH 282CCOsMGLxq7hcX9mpysQY= =uCqy -----END PGP SIGNATURE-----