[DRBD-user] drbd 0.6.12 oops on RHEL3

Lars Ellenberg Lars.Ellenberg at linbit.com
Tue May 25 13:59:54 CEST 2004

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.


/ 2004-05-25 13:30:25 +0200
\ Wim Vandersmissen:
> 
> Hi,
> 
> I'm running drbd 0.6.12 on RHEL3 with kernel 2.4.21-15.ELsmp
> orochimaru is running drbd on /dev/md1 (100gig software raid5), the other host
> oldinverse is running drbd on /dev/sdb2 (100gig hardware raid5)
> 
> After a 5 hour syncing the 100gb (orochimaru primary) I formatted nb0, mounted
> it and copied a few gigabytes on it. All was well  
> 
> Then I unmounted it, made orochimaru secondary, and oldinverse primary, mounted 
> nb0 on oldinverse, and did a find on the mounted nb0. This resulted in 
> "kernel: drbd0: [kjournald/1882] sock_sendmsg time expired, ko = 4294967265"
> errors. And processes in D-state. So I ran "reboot", that gave me the attached
> oops (painfully manually copied 
> 
> Any ideas ?

for the oops: DRBD does not (yet) have a reboot notifier.
so it likely does strange things when the box gets rebooted with
configured DRBDs. it expects its devices to be unconfigured first.
I never tried to reboot a still configured DRBD, so I cannot tell what
will happen. Maybe this leads to the oops.

both 'D' state and "ko ... " (which are warnings, not errors) are just
some hints that your setup suffers from high latency.
find the bottleneck, and eliminate it.

	Lars Ellenberg



More information about the drbd-user mailing list