[Pacemaker] Pacemaker and DRBD - Will Not Promote

[Ted Young]

Good day everyone,

 

I am trying to build a pair of servers (Debian Squeeze) as a pacemaker
testbed.  Each one hosts a single DRBD drive.  One server is the master and
the other is the slave.  Ultimately, I will set up LVM to bring the two
together to create a fault tolerant SAN; but I have not gotten that far yet.

 

Problem: when the master node is turned off the second node remains a slave
(it is not promoted).  The slave does show that the master is no longer
available.

 

Here is the result of ptest -Ls:

 

drbd: 0 promotion score on none: 0

drbd: 1 promotion score on san2: -1

 

For some reason pacemaker refuses to allow for promotion on the second node.

 

Version: I upgraded to version 1.1.5.  Oddly enough, this worked fine under
1.0.9 with the exception of the fact that no matter how hard I tried I could
not keep the master where it was when the failed node was rebooted (no
matter how high I set resource-stickiness, the node that came back up was
promoted back to master).  So, if anyone can help me solve that problem with
1.0.9, I would be happy to simply downgrade.

 

On both versions I used variations of the same configuration, all yielding
the same results:

 

node san1

node san2

 

primitive drbd ocf:linbit:drbd params drbd_resource=r0 op monitor
interval=60s

 

ms msdrbd drbd meta master-max= 1 master-node-max=1 clone-max=2
clone-node-max=1 notify=true

 

rsc_defaults resource-stickiness=100

 

So, as you can see, I am trying to get DRBD working before I move any
further.  I cannot imagine what mistake I must be making for behavior to be
so inconsistent with the dozens of tutorials online.  I would deeply
appreciate any help.

 

Ted Young

http://tedyoung.me

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.clusterlabs.org/pipermail/pacemaker/attachments/20111027/210c2138/attachment-0002.html>