system reboot because of hdd

[claudiu vasadi]

Hello ppl.


Straight to business.

FreeBSD 7.1-stable

2 hdd. 1 is ad2 and the other is ad6. ad2 is the BSD hdd, and ad6 is just
for data (movies, music, etc). ad2 is a 80GB Samsung P-ata133 and ad6 is a
WD 250GB S-ata2.

While running a process that was trying to create a 25GB file on a 30 GB
partition on the second hdd (ad2) I experienced ssh outage. Everything came
back to life after a short perioud of ~2 minutes. So, again I started the
process. This time, the outage was about 5 minutes. I was busy with
something else and did not run the process again. 2 minutes after that i get
a call from a customer that some thing is not working. so I check it and
surprize, the OS rebooted itself.

so, went to the logs and this is what i found out (/var/log/messages):

May 29 22:26:30 da1 kernel: ad6: TIMEOUT - WRITE_DMA48 retrying (1 retry
left) LBA=419468447
May 29 22:26:35 da1 kernel: ad6: TIMEOUT - WRITE_DMA48 retrying (0 retries
left) LBA=419468447
May 29 22:26:41 da1 kernel: ad6: FAILURE - WRITE_DMA48 timed out
LBA=419468447
May 29 22:26:41 da1 kernel: g_vfs_done():ad6s1f[WRITE(offset=19447808,
length=16384)]error = 5
May 29 22:26:35 da1 syslogd: kernel boot file is /boot/kernel/kernel
May 29 22:26:35 da1 kernel: ad6: FAILURE - device detached
May 29 22:26:35 da1 kernel: subdisk6: detached
May 29 22:26:35 da1 kernel: ad6: detached
May 29 22:26:35 da1 kernel: g_vfs_done():ad6s1f[WRITE(offset=36683776,
length=16384)]error = 6
May 29 22:26:35 da1 kernel: g_vfs_done():ad6s1f[WRITE(offset=16908288,
length=16384)]error = 6
May 29 22:26:35 da1 kernel: g_vfs_done():ad6s1f[WRITE(offset=36700160,
length=16384)]error = 6
May 29 22:26:35 da1 kernel: g_vfs_done():ad6s1f[WRITE(offset=114688,
length=16384)]error = 6
May 29 22:26:35 da1 kernel: panic: vinvalbuf: dirty bufs
May 29 22:26:35 da1 kernel: cpuid = 0
May 29 22:26:35 da1 kernel: Uptime: 45d22h15m29s
May 29 22:26:35 da1 kernel: Physical memory: 1003 MB
May 29 22:26:35 da1 kernel: Dumping 232 MB: 217 201 185 169 153 137 121 105
89 73 57 41 25 9




and (/var/log/all.log):


May 29 22:54:49 da1 fsck: /dev/ad6s1f: 6 files, 12 used, 17132271 free (31
frags, 2141530 blocks, 0.0% fragmentation)


exacly where the file was created. but it was 1 not 6 files that i wanted to
create but 1.




the process that I run is "dsmfmt" of TSM server for Sun. it creates a file
volume of a specific size for use in tsm server itself for defining storage
pool capacity.

so, I know that the hdd was to the limit. It could be a hardware issue I
know, but right now dnt have resources to try somewere else so I'm asking a
oppinion. Has anyone dealt with this situation before ? OS reboot because of
high hdd load ?