[DRBD-user] Using DRBD environment to take contact and discover problems

Antonio Fernández Pérez antoniofernandez at fabergames.com
Tue Oct 15 18:32:43 CEST 2013

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

Hi everybody,

Recently I have configured DRBD+Pacemaker+Corosync cluster using LVM.  All
works fine.
Now I'm working in discovering and fixing possible issues. I have caused a
SPLIT BRAIN (Crossover ethernet cable disconnected). I have read about
typical operations that I have to follow to solve it: Choosing a victim,
making secondary all and then making in the other node, primary all.
In some cases the cluster, after split brain problem, has not master node
so, both nodes appears in crm_mon as slave but with ip resource started.
In other cases, after split brain, Corosync solves automatically the SPLIT
BRAIN and discards the data from a node (I don't know what is the algorithm
to choose a victim automatically) and make primary the other node.

Anybody could tell me if this is normal behaviour? I have to optimize my
configuration to get something ...
And another question ... in a environment with two servers, with two
ethernets and one of them as crossover link between nodes. What would the
most typical problems that could happen?

Thanks in advance.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linbit.com/pipermail/drbd-user/attachments/20131015/0168e396/attachment.htm>

More information about the drbd-user mailing list