Hello.
1. kernel 2.4.20pre2
hardware Intel LanceWood L440GX with onboard AIC controller
4 hdd, "Vendor: IBM Model: DDYS-T18350M Rev: S93E
2. software raid5:
Personalities : [raid1] [raid5] [multipath]
read_ahead 1024 sectors
md0 : active raid5 sdd2[3] sdc2[2] sdb2[1] sda2[0]
52146624 blocks level 5, 64k chunk, algorithm 2 [4/4] [UUUU]
unused devices: <none>
3. log:
...
Aug 21 16:37:29 ruby kernel: 09:00: rw=0, want=140379796, limit=52146624
Aug 21 16:37:29 ruby kernel: attempt to access beyond end of device
Aug 21 16:37:29 ruby kernel: 09:00: rw=0, want=1938395668, limit=52146624
Aug 21 16:37:29 ruby kernel: attempt to access beyond end of device
Aug 21 16:37:29 ruby kernel: 09:00: rw=0, want=1971900992, limit=52146624
Aug 21 16:37:29 ruby kernel: attempt to access beyond end of device
Aug 21 16:37:29 ruby kernel: 09:00: rw=0, want=1971900992, limit=52146624
Aug 21 16:37:29 ruby kernel: attempt to access beyond end of device
Aug 21 16:37:29 ruby kernel: 09:00: rw=0, want=140379796, limit=52146624
Aug 24 16:30:49 ruby kernel: attempt to access beyond end of device
Aug 24 16:30:49 ruby kernel: 09:00: rw=0, want=140379796, limit=52146624
Aug 24 16:30:49 ruby kernel: attempt to access beyond end of device
Aug 24 16:30:49 ruby kernel: 09:00: rw=0, want=1938395668, limit=52146624
Aug 24 16:30:49 ruby kernel: attempt to access beyond end of device
Aug 24 16:30:49 ruby kernel: 09:00: rw=0, want=1971900992, limit=52146624
Aug 24 16:30:49 ruby kernel: attempt to access beyond end of device
Aug 24 16:30:49 ruby kernel: 09:00: rw=0, want=1971900992, limit=52146624
Aug 24 16:30:49 ruby kernel: attempt to access beyond end of device
Aug 24 16:30:49 ruby kernel: 09:00: rw=0, want=140379796, limit=52146624
...
4. after uptime 8 day /dev/md0 - go "read-only"
(tune2fs enable on /dv/md0 "Errors behavior:Remount read-only")
and e2fsck found many error (empty lost+found)
Question: why such can occur and what to make to avoid repetition?
--
bye.
Andrey Nekrasov, SpyLOG.