Hi
I have troubles with my softwareraid. it keeps corrupting itself. i mean,
it's corrupt, i do "e2fsck /dev/md0", say yes to everything, filesystem
seems ok, work on it, filesystem is corrupt again...
the raid was fine during the last months running on rh7.3 and 8.0. now
i upgraded to rh9 and then the troubles started.
the upgrade went relatively smooth, just anaconda couldn't correctly figure
out my boot-drive to install grub. it suggested /dev/hda but it should be
/dev/sda. i don't know, if it touched hda (i did a nfs-install booting from
floppy, that's probably what got anaconda confused. it didn't see the higher
bootpriority of scsi over ide).
i then fixed grub manually to boot from sda.
during the upgrade my raid was disabled (deleted from fstab) because
anaconda couldn't mount it and aborted the upgrade.
after installation i installed all updates to rh9.
cat /proc/version
Linux version 2.4.20-18.8 ([EMAIL PROTECTED]) (gcc version 3.2 20020903 (Red Hat Linux
8.0 3.2-7)) #1 Thu May 29 08:57:39 EDT 2003
the system is a 1GHz Athlon on a Asus A7V (i think). the system and bootdrive
is a scsi-disk connected to a ncr-controller. the raid is on 4 ide-disks
connected to the onboard via and promise controllers.
cat /etc/raidtab:
raiddev /dev/md0
raid-level 5
nr-raid-disks 4
nr-spare-disks 0
chunk-size 32
persistent-superblock 1
parity-algorithm left-symmetric
device /dev/hda1
raid-disk 0
device /dev/hdc1
raid-disk 1
device /dev/hde1
raid-disk 2
device /dev/hdg1
raid-disk 3
when i brought the raid back up, it ran in degraded mode. /dev/hde1 was
kicked from the raid. so i cheched the drive with "hdparm -t -T" and
"badblocks /dev/hde", but it seemd fine. so i did "raidhotadd /dev/md0
/dev/hde1".
after waiting for reconstruction:
cat /proc/mdstat
Personalities : [raid5]
read_ahead 1024 sectors
md0 : active raid5 hdg1[3] hde1[2] hdc1[1] hda1[0]
480238656 blocks level 5, 32k chunk, algorithm 2 [4/4] [UUUU]
unused devices: <none>
then i ran "e2fsck /dev/md0" which took a long time and tons of errors,
i had tons of stuff in lost+found and also most of the other files
were corrupt (could see bits from different files mixed together :(
of course i have no backup, but the data is not so valuable. it just takes
time to reencode all my cds again...
i tried to sort files out, but then the raid became corrupt agaiain, after
another fsck i tried again, but again the raid became corrupt.
that's where i am now.
i could start from scratch. ditch everything and recreate raid...
then i noticed this:
top -bn1
14:00:02 up 1 day, 11:45, 4 users, load average: 1.09, 1.11, 1.09
90 processes: 88 sleeping, 2 running, 0 zombie, 0 stopped
CPU states: 3.0% user 2.8% system 94.0% nice 0.0% iowait 0.0% idle
Mem: 514652k av, 506864k used, 7788k free, 0k shrd, 218388k buff
383288k actv, 0k in_d, 10192k in_c
Swap: 1049592k av, 22824k used, 1026768k free 240800k cached
PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME CPU COMMAND
[..]
9 root 18446744073709551615 -20 0 0 0 SW< 0.0 0.0 0:00 0
mdrecoveryd
[..]
5694 root 18446744073709551615 -20 0 0 0 SW< 0.0 0.0 1:00 0
raid5d
17801 root 9 0 1144 984 904 S 0.0 0.1 0:00 0 sshd
[..]
note the priority of mdrecoveryd and raid5d! (the sshd is just there to show
the normal priority of all the other processes).
what's wrong here?
does anybody successfully run a raid5 on kernel 2.4.20-18.8?
do you need more information?
how sould i proceed?
thanks for reading so far :)
bye
dworz
--
redhat-list mailing list
unsubscribe mailto:[EMAIL PROTECTED]
https://www.redhat.com/mailman/listinfo/redhat-list