[OpenIndiana-discuss] failmode

[Maurilio Longo]

Hi all,

yesterday morning one of the disks on the root pool of a pc of mine which is
used as iscsi target via comstar had a read/write failure, this morning at
00:10 GMT+1 a scrub started (from cron) on the same rpool (I think it went
nowhere).

Since the rpool hosts the boot image (/tftpboot/undionly.kpxe) which is sent
to the clients no client was able to boot anymore.

Rpool disks are consumer grade disks, so it is partly my fault, but I'd like
to understand if changing failmode from "wait" to "continue" could have made
things smoother and the system able to keep working (I had to boostrap it to
regain access to root pool), I was no more able to log into it.

BTW, during boot a message is spit out which makes you think a disk has been
detached, but it is not.

System is using oi_147 and so far worked flawlessly.

TIA.

Maurilio.

ps. I attach the fmdump log and zpool status.


-- 
 __________
|  |  | |__| Maurilio Longo
|_|_|_|____| farmaconsult s.r.l.