Date: Tue, 15 Jun 2004 17:24:47 +0200 From: Joerg Wunsch <j@ida.interface-business.de> To: freebsd-scsi@freebsd.org Subject: Drive statistics broken after replacing a drive? Message-ID: <20040615172447.D47727@ida.interface-business.de>
next in thread | raw e-mail | index | archive | help
I've got a Sun machine here where I've setup a vinum RAID-5 volume for
/usr/ports, mainly in order to play a bit with the RAID-5 feature
where I've never had an occasion before.
So in order to get a feeling for what needs to be done to replace a
failed drive of a RAID-5 plex, I started a "cvs update" on that
/usr/ports directory, in order to get some heavy disk traffic. Then,
I pulled one of the drives. The kernel properly announces the drive
failure, vinum degrades the RAID-5 plex, the "cvs update" is
continuing normally (not unexpectedly ;-).
I then replaced the drive by another one, labelled it, and restarted
the appropriate vinum structures in order to re-synchronize the RAID
group. All worked out pretty well, the RAID-5 plex is up & running
again -- but "iostat" no longer updates any statistics for that drive.
While all the remaining disks participating in that RAID group display
heavy traffic in iostat, and all the drive LEDs flicker as well, the
disk's statistics for the new-born drive keep solid at 0:
# iostat -d -w1 da3 da4 da5 da6
da3 da4 da5 da6
KB/t tps MB/s KB/t tps MB/s KB/t tps MB/s KB/t tps MB/s
19.38 1 0.01 19.35 1 0.01 19.29 1 0.01 6.17 0 0.00
2.00 4 0.01 2.00 16 0.03 2.00 14 0.03 0.00 0 0.00
2.00 23 0.04 2.00 3 0.01 2.00 21 0.04 0.00 0 0.00
^C
(The first line is the accumulated statistics, it never changes again
for da6.)
--
J"org Wunsch Unix support engineer
joerg_wunsch@interface-systems.de http://www.interface-systems.de/~j/
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20040615172447.D47727>