[DRBD-user] Using drbd in ha cluster

Lionel Sausin ls at numerigraphe.com
Thu May 19 10:06:26 CEST 2011

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


Hi,
DRBD ships with a script called snapshot-resync-target-lvm.sh that does
just that, that can be called before/after resync.
In our config we use :
before-resync-target "/etc/scripts/snapshot-resync-target-lvm.sh -p 15
-- -c 16k";
after-resync-target /etc/scripts/unsnapshot-resync-target-lvm.sh;
Please make sure you get the latest version of the script, because the
one shipped in ubuntu 10.04 (DRBD v8.3.7) is buggy and goes into an
endless loop.
Also, only a recent kernel will let you merge the snapshot back into the
main volume.
Lionel Sausin.

Le -10/01/-28163 20:59, Lyre a écrit :
> Hi all:
>
> I'm considering using drbd in 2 node ha cluster.
>
> Consider such situation:
>
> Primary down while secondary (from StandAlone to SyncTarget) was syncing
> from priamry , so that the data on secondary was Inconsistent and can't be
> primary。 In a ha cluster, we would prefer to bring up the secondary as the
> new primary node in such condition. It would nice to have snapshot before
> sync start, so that we can recover it to a clean but partially outdate state
> when primary crash during syncing. Is there a way to do this?
>




More information about the drbd-user mailing list