On Jul 8, 2013, at 12:59 PM, Andreas Kurz <[email protected]> wrote:
> On 2013-07-08 19:40, Jeff Frost wrote: >> We're testing out the pgsql master slave streaming replication resource >> agent that's found here: >> >> https://github.com/ClusterLabs/resource-agents/blob/master/heartbeat/pgsql >> >> and using the example 2-node configuration found here >> https://github.com/t-matsuo/resource-agents/wiki/Resource-Agent-for-PostgreSQL-9.1-streaming-replication >> as a template, we came up with the following configuration: >> >> >> node node1 >> node node2 >> primitive pgsql ocf:heartbeat:pgsql \ >> params pgctl="/usr/pgsql-9.2/bin/pg_ctl" psql="/usr/pgsql-9.2/bin/psql" >> pgdata="/var/lib/pgsql/9.2/data/" start_opt="-p 5432" rep_mode="async" >> node_list="node1 node2" repuser="replicauser" restore_command="rsync -aq >> /var/lib/pgsql/wal_archive/%f %p" master_ip="192.168.253.104" >> stop_escalate="0" \ >> op start interval="0s" role="Master" timeout="60s" on-fail="block" > > Looks like you are missing the monitor operations ... as described in > the example you are referring. In the monitoring operation such > master-slave agents recalculate their master-score and refresh e.g. in > this RA various node-attributes. > > And you should follow the described procedures to correctly start-up the > cluster. Interesting. When I set it up, I did it with pcs like so: pcs -f pgsql_cfg resource create pgsql ocf:heartbeat:pgsql \ params \ pgctl="/usr/pgsql-9.2/bin/pg_ctl" \ psql="/usr/pgsql-9.2/bin/psql" \ pgdata="/var/lib/pgsql/9.2/data/" \ start_opt="-p 5432" \ rep_mode="async" \ node_list="node1 node2" \ repuser="replicauser" \ restore_command="rsync -aq /var/lib/pgsql/wal_archive/%f %p" \ primary_conninfo_opt="keepalives_idle=60 keepalives_interval=5 keepalives_count=5" \ master_ip="192.168.253.104" \ stop_escalate="0" \ op start timeout="60s" interval="0s" on-fail="restart" \ op monitor timeout="60s" interval="10s" on-fail="restart" \ op monitor timeout="60s" interval="9s" on-fail="restart" role="Master" \ op promote timeout="60s" interval="0s" on-fail="restart" \ op demote timeout="60s" interval="0s" on-fail="stop" \ op stop timeout="60s" interval="0s" on-fail="block" \ op notify timeout="60s" interval="0s" But when I pull the info out with crm, it appears as in my original post: primitive pgsql ocf:heartbeat:pgsql \ params pgctl="/usr/pgsql-9.2/bin/pg_ctl" psql="/usr/pgsql-9.2/bin/psql" pgdata="/var/lib/pgsql/9.2/data/" start_opt="-p 5432" rep_mode="async" node_list="node1 node2" repuser="replicauser" restore_command="rsync -aq /var/lib/pgsql/wal_archive/%f %p" master_ip="192.168.253.104" stop_escalate="0" \ op start interval="0s" role="Master" timeout="60s" on-fail="block" and definitely missing the monitor operations as you point out. And sure enough, using crm to configure it causes it work as expected. I guess that's what I get for trying to use pcs. Any idea how it should be done with pcs? _______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
