Hot-removing SCSI disk freezes woody RAID system
Hi all,
Tried this on linux-raid without result.
I am stress-testing an LVM-on-RAID5 setup with Debian GNU/Linux 3.0
"woody". Kernel version is 2.4.20
To test redundancy I hot-remove one of the SCSI SCA disks on which the
array is built. I would have expected the array to degrade, but instead
the machine freezes.
A lot of error messages come up, ending with:
-- Begin messages ---
sym53c8xx_reset: pid=383206 reset_flags=1 serial_number_at_timeout=0
scsi0: device driver called scsi_done() for a synchronous reset.
SCSI disk error: host 0 channel 0 id 2 lun 0 return code = 2603007f
I/O error: dev 08:22, sector 31496
--- End messages ---
The machine then locks up. It looks like the kernel is still there:
The syslog still dumps to stdout, the TCP/IP stack responds to ping,
and the virtual terminals present a login prompt, but I can't log in
and nothing else works.
On reboot, the array and filesystems (I use reiserfs for everything
expcept /boot which is on ext2) come up normally (arrays come up in
degraded mode, of course). This is good news, but I'd rather the
machine did not freeze.
Anyone have any ideas as to what causes the crash and how to prevent
it?
--
Best regards,
| George Karaolides Planitis Communications Ltd.,
|
| tel: +357 22 45 65 00 Office 102, 11 Florinis Street,
|
| web: www.planitis.net Nicosia CY 1065,
|
| email: george.karaolides@planitis.net Republic of Cyprus.
|
Reply to: