[Pacemaker] Pacemaker 1.1.6 order possible bug ?
Andrew Beekhof
andrew at beekhof.net
Tue Sep 4 21:51:35 UTC 2012
On Mon, Sep 3, 2012 at 3:41 PM, Tomáš Vavřička <vavricka at ttc.cz> wrote:
> Hello,
>
> Sorry If I send same question twice, but message did not appeared on mailing
> list.
>
> I have a problem with orders in pacemaker 1.1.6 and corosync 1.4.1.
>
> Order below is working for failover, but it is not working when one cluster
> node starts up (drbd stays in Slave state and ms_toponet is started before
> DRBD gets promoted).
>
> order o_start inf: ms_drbd_postgres:promote postgres:start
> ms_toponet:promote monitor_cluster:start
>
> Order below is not working for failover (it kills slave toponet app and
> start it again) but it is working correctly when cluster starts up.
>
> order o_start inf: ms_drbd_postgres:promote postgres:start ms_toponet:start
> ms_toponet:promote monitor_cluster:start
I would recommend breaking this into "basic" constraints.
The shell syntax for constraint sets has been a source of confusion for a while.
order o1 inf: ms_drbd_postgres:promote postgres:start
order o2 inf: postgres:start ms_toponet:start
order o3 inf: ms_toponet:start ms_toponet:promote
order o4 inf: ms_toponet:promote monitor_cluster:start
If you still have problems with the expanded form, let us know.
>
> I want to the pacemaker to act as in 1.0.12 version.
> * when toponet master app is killed, move postgres resource to other node
> and promote ms_toponet and ms_drbd_postgres to Master
> * when one node is starting promote DRBD to master is is UpToDate
>
> Am I doing something wrong?
>
> It looks to me pacemaker ignores some orders (pacemaker should wait for DRBD
> promotion when starting toponet app, but toponet app is started right after
> DRBD start (slave)). I tried to solve this by different orders with
> combination symmetrical=false, split orders, different orders for start and
> stop, but no success at all (seems to me like completely ignoring
> symmetrical=false directive).
>
> Pacemaker 1.1.7 is not working for me, because it has broken on-fail
> directive.
>
> crm_mon output:
>
> ============
> Last updated: Fri Aug 31 14:51:11 2012
> Last change: Fri Aug 31 14:50:27 2012 by hacluster via crmd on toponet30
> Stack: openais
> Current DC: toponet30 - partition WITHOUT quorum
> Version: 1.1.6-b988976485d15cb702c9307df55512d323831a5e
> 2 Nodes configured, 2 expected votes
> 10 Resources configured.
> ============
>
> Online: [ toponet30 toponet31 ]
>
> st_primary (stonith:external/xen0): Started toponet30
> st_secondary (stonith:external/xen0): Started toponet31
> Master/Slave Set: ms_drbd_postgres
> Masters: [ toponet30 ]
> Slaves: [ toponet31 ]
> Resource Group: postgres
> pg_fs (ocf::heartbeat:Filesystem): Started toponet30
> PGIP (ocf::heartbeat:IPaddr2): Started toponet30
> postgresql (ocf::heartbeat:pgsql): Started toponet30
> monitor_cluster (ocf::heartbeat:monitor_cluster): Started toponet30
> Master/Slave Set: ms_toponet
> Masters: [ toponet30 ]
> Slaves: [ toponet31 ]
>
> configuration:
>
> node toponet30
> node toponet31
> primitive PGIP ocf:heartbeat:IPaddr2 \
> params ip="192.168.100.3" cidr_netmask="29" \
> op monitor interval="5s"
> primitive drbd_postgres ocf:linbit:drbd \
> params drbd_resource="postgres" \
> op start interval="0" timeout="240s" \
> op stop interval="0" timeout="120s" \
> op monitor interval="5s" role="Master" timeout="10s" \
> op monitor interval="10s" role="Slave" timeout="20s"
> primitive monitor_cluster ocf:heartbeat:monitor_cluster \
> op monitor interval="30s" \
> op start interval="0" timeout="30s" \
> meta target-role="Started"
> primitive pg_fs ocf:heartbeat:Filesystem \
> params device="/dev/drbd0" directory="/var/lib/pgsql" fstype="ext3"
> primitive postgresql ocf:heartbeat:pgsql \
> op start interval="0" timeout="80s" \
> op stop interval="0" timeout="60s" \
> op monitor interval="10s" timeout="10s" depth="0"
> primitive st_primary stonith:external/xen0 \
> op start interval="0" timeout="60s" \
> params hostlist="toponet31:/etc/xen/vm/toponet31"
> dom0="172.16.103.54"
> primitive st_secondary stonith:external/xen0 \
> op start interval="0" timeout="60s" \
> params hostlist="toponet30:/etc/xen/vm/toponet30"
> dom0="172.16.103.54"
> primitive toponet ocf:heartbeat:toponet \
> op start interval="0" timeout="180s" \
> op stop interval="0" timeout="60s" \
> op monitor interval="10s" role="Master" timeout="20s"
> on-fail="standby" \
> op monitor interval="20s" role="Slave" timeout="40s" \
> op promote interval="0" timeout="120s" \
> op demote interval="0" timeout="120s"
> group postgres pg_fs PGIP postgresql
> ms ms_drbd_postgres drbd_postgres \
> meta master-max="1" master-node-max="1" clone-max="2"
> clone-node-max="1" notify="true" target-role="Master"
> ms ms_toponet toponet \
> meta master-max="1" master-node-max="1" clone-max="2"
> clone-node-max="1" target-role="Master"
> location loc_st_pri st_primary -inf: toponet31
> location loc_st_sec st_secondary -inf: toponet30
> location master-prefer-node1 postgres 100: toponet30
> colocation pg_on_drbd inf: monitor_cluster ms_toponet:Master postgres
> ms_drbd_postgres:Master
> order o_start inf: ms_drbd_postgres:start ms_drbd_postgres:promote
> postgres:start ms_toponet:start ms_toponet:promote monitor_cluster:start
> property $id="cib-bootstrap-options" \
> dc-version="1.1.6-b988976485d15cb702c9307df55512d323831a5e" \
> cluster-infrastructure="openais" \
> expected-quorum-votes="2" \
> no-quorum-policy="ignore" \
> stonith-enabled="true"
> rsc_defaults $id="rsc-options" \
> resource-stickiness="5000"
>
> _______________________________________________
> Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://bugs.clusterlabs.org
More information about the Pacemaker
mailing list