[Pacemaker] DRBD not demoted when moving resources

[Саша Александров]

Hi, all!

I set up the latest clustering sftware versions from CentOS 6.5 repo
(pacemaker 1.1.10 + pcs + drbd 8.4.4) on a two-node cluster.

pcs constraint show
Location Constraints:
Ordering Constraints:
  promote ms_u01 then start u01_fs
Colocation Constraints:
  u01_fs with ms_u01 (with-rsc-role:Master)

Now, when I reboot or shut down 'active' server - filesystem gets promoted
on the secondry, and filesystem gets mounted - all as expected. However,
when I move resource group with pcs resource move - resources are stopped
(correct) and FS unmounted, but DRBD resource stays Primary, after some
timeout I get failed action "monitor with role:Master" on the "target"
server and eventually resource is kept out of the target server.

I tried to add additional ordering constraint, like this:

pcs constraint show
Location Constraints:
  Resource: OraGroup
    Enabled on: server2 (role: Started)
Ordering Constraints:
  promote ms_u01 then start u01_fs
  stop u01_fs then demote ms_u01
Colocation Constraints:
  u01_fs with ms_u01 (with-rsc-role:Master)

This did not change anything...

Am I missing somethig?


The previous OS (it was OpenSuSE) and software (pacemaker 1.1.8 + crmsh)
worked as expected:

primitive u01_drbd ocf:linbit:drbd \
        params drbd_resource="u01" \
        op start interval="0" timeout="240" \
        op stop interval="0" timeout="100" \
        op monitor interval="20" role="Slave" timeout="60" depth="0" \
        op monitor interval="10" role="Master" timeout="60" depth="0"
primitive u01_fs ocf:heartbeat:Filesystem \
        params device="/dev/drbd2" directory="/u01" fstype="ext3" \
        op start interval="0" timeout="60" \
        op stop interval="0" timeout="60"
primitive u01_ip ocf:heartbeat:IPaddr2 \
        params ip="********" nic="eth0" cidr_netmask="255.255.255.248"
primitive u01_ip_local ocf:heartbeat:IPaddr2 \
        params ip="10.10.10.253" nic="eth1" cidr_netmask="255.255.255.0"
primitive u01_lsnr ocf:heartbeat:oralsnr \
        params sid="orcl" \
        op start interval="0" timeout="120" \
        op stop interval="0" timeout="120" \
        op monitor interval="30" timeout="60"
primitive u01_orcl ocf:heartbeat:oracle \
        params sid="orcl" \
        op start interval="0" timeout="300" \
        op stop interval="0" timeout="300" \
        op monitor interval="30" timeout="180"

group U01Group u01_ip u01_ip_local u01_fs u02_fs u01_lsnr u01_orcl \

        meta target-role="Started"
ms ms_u01 u01_drbd \
        meta master-max="1" master-node-max="1" clone-max="2"
clone-node-max="1" notify="true" target-role="Master"
location cli-prefer-U01Group U01Group \
        rule $id="cli-prefer-rule-U01Group" inf: #uname eq server2
colocation u0_on_drbd inf: u01_ip u01_ip_local ms_u01:Master u01_fs
ms_u02:Master u02_fs u01_lsnr u01_orcl
order order_u01 inf: u01_ip u01_ip_local ms_u01:promote u01_fs:start
ms_u02:promote u02_fs:start u01_lsnr u01_orcl




-- 
Best regards,
Alexandr A. Alexandrov
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.clusterlabs.org/pipermail/pacemaker/attachments/20140403/edb2bda8/attachment-0002.html>