[Pacemaker] location ping rules

s.oreilly s.oreilly at linnovations.co.uk
Thu Nov 7 10:34:28 UTC 2013


Having some trouble getting a location rule to work.

Here is my current config:


Resources:
 Master: master_drbd
  Meta Attrs: master-max=1 master-node-max=1 clone-max=2 clone-node-max=1 notify=true
  Resource: drbd_mysql (class=ocf provider=linbit type=drbd)
   Attributes: drbd_resource=clusterdb_res
   Operations: monitor interval=15s (drbd_mysql-interval-15s)
 Group: mysql_grp
  Resource: MySQLfs (class=ocf provider=heartbeat type=Filesystem)
   Attributes: device=/dev/drbd0 directory=/var/lib/mysql_drbd fstype=xfs
  Resource: ClusterIP (class=ocf provider=heartbeat type=IPaddr2)
   Attributes: ip=10.*.*.* cidr_netmask=24 nic=bond0
   Operations: monitor interval=30s (ClusterIP-interval-30s)
  Resource: mysql (class=ocf provider=heartbeat type=mysql)
   Attributes: config=/var/lib/mysql_drbd/my.cnf socket=/var/lib/mysql/mysql.sock
pid=/var/lib/mysql/mysql.pid datadir=/var/lib/mysql_drbd/data
binary=/usr/bin/mysqld_safe
   Operations: monitor interval=20s timeout=30s (mysql-timeout-30s)
 Clone: ping-clone
  Meta Attrs: globally-unique=false
  Resource: ping (class=ocf provider=pacemaker type=ping)
   Attributes: host_list=10.*.*.* multiplier=1000 timeout=60s
   Operations: monitor interval=15s (ping-interval-15s)

Stonith Devices:
Fencing Levels:

Location Constraints:
Ordering Constraints:
  promote master_drbd then start mysql_grp (Mandatory)
(id:order-master_drbd-mysql_grp-mandatory)
Colocation Constraints:
  mysql_grp with master_drbd (INFINITY) (with-rsc-role:Master)
(id:colocation-mysql_grp-master_drbd-INFINITY)

I am trying to add a location rule to fail over if the network goes down on the
primary node with the following rule:

pcs constraint location master_drbd rule -INFINITY: not_defined ping or ping lte 0

This is just breaking the cluster though and nothing is started on either node.

Can anyone help?


Regards



Sean O'Reilly




More information about the Pacemaker mailing list