[DRBD-user] read only ? internal journal full ?

Bruno Depero bruno.depero at digicar.com
Mon Jan 17 09:43:36 CET 2005

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


A smal problem I have.
Primary server with 4x300Gb sata raid5 (md) disks with gigalan.
Secondary scsi-160 4 hd's with hw raid always gigalan.
Yesterday was the third time, in one month, the main drbd device becomed 
read-only.
Both pc has 2.6.9 kernel and has ups. Upsmon reports no failures on 
power supply.
I guess it becomes read-only when it's journal become full and no way to 
replicate it to the secondary pc.
Ping always works, the two pc are always connected.
If I unmount the drbd device, stop and restart drbd (unloading drbd 
module) no way to restart it. Reports that there are no internal usable log.
If I do a full restart of the primary pc, when loading drbd it reports 
there is no connection with the secondary and starts the timeout 
counter, or I have to type yes to abort it.
I connect via ssh to the primary pc and I can ping the secondary without 
any problem, either I can connect from the primary to the secondary via 
ssh. Firewall is disabled.
I can connect via ssh to both primary and secondary pc without any 
problem but drbd 0.7.7 reports that there is no connection ?!?
It's a database server, so isn't good when on week-ends drbd lose 
connection and device become read-only !
Some suggestions ?

Thank you in advance, best regards.

Bruno Depero




More information about the drbd-user mailing list