Clustered postgres-8 resource fails to start with no error other than "Starting Service postgres-8:<name> > Failed" during relocation when rebooting a node in RHEL 5 or 6
Issue
- After issuing a soft reboot on the node running a service with
postgres-8
resource, it fails to start on the other node. It doesn't give any useful failure messages:
Sep 25 16:21:15 node2 clurgmgrd[4862]: <notice> Member 1 shutting down
Sep 25 16:21:20 node2 clurgmgrd[4862]: <notice> Starting stopped service service:database
[...]
Sep 25 16:21:23 node2 clurgmgrd: [4862]: <info> Starting Service postgres-8:posgresql
Sep 25 16:21:26 node2 clurgmgrd: [4862]: <err> Starting Service postgres-8:posgresql > Failed
Sep 25 16:21:26 node2 clurgmgrd[4862]: <notice> start on postgres-8 "posgresql" returned 1 (generic error)
Sep 25 16:21:26 node2 clurgmgrd[4862]: <warning> #68: Failed to start service:NSDG_Postgresql; return value: 1
-
When I reboot the active node which is running the service, service successfully relocates on passive node. After reboot, first cluster node comes back and joins the cluster as normal. Now, when I tried to reboot the node 2 which is running the service now, it fails to relocate the service back to node 1 which is acting as passive now.
-
Cluster service is failing to relocate from one node to other node of cluster with
postgres-8
resource
Environment
- Red Hat Enterprise Linux (RHEL) 5 and 6 with the High Availability Add On
rgmanager
- One or more services with a
postgres-8
resource in/etc/cluster/cluster.conf
- RHEL 6:
resource-agents
releases prior to3.9.5-12.el6
- RHEL 5:
rgmanager
releases prior to2.0.52-54.el6
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.