0

I have clustering nfs server with corosync and pacemaker.

I Installed environment successfully then i fount a problem while testing.

enter image description here

That screen is captured after add resources.

nfs1 server is working well, and all resources are watched by pacemaker.

The problem occur after stop nfs-service.

If i input commend "systemctl stop nfs", nfs service is stopped. Then cluster is move to nfs2 automatically(This is ok)

enter image description here

Then i input commend "pcs cluster standby bp-nfs2", as a result cluster server is moved to bp-nfs1, and all resources was be satrted without nfsserver.

enter image description here

  1. Even if i started again nfs-service, pacemaker's nfs resource is still stopped.

  2. I want that pacemaker make nfs-service run when pacemaker start.

This is nfs resource create commend.

pcs resource create nfsserver ocf:heartbeat:nfsserver \ nfs_shared_infodir="/mnt/sharedisk/" \ --group resource-group

Somebody know about this issue, please teach me.

If you can't under stand my English, I'm sorry. Thank you.

Gs.
  • 540
  • 1
  • 3
  • 21

1 Answers1

0

Solved myself.

Reason.

If resource is stopped by trouble, the resource have FailAction. We can see what resource have failAction with pcs status commend.

And Pacemaker never start to watch resource which have failaction, It is pacemaker's specification.

Solution.

Clear failaction manually with commend "pcs resource cleanup [resource name]".

If you want to cleanup automatically, input this commend "pcs resource defaults failure-timeout=60s".

When occur trouble your resource, your activated node move to another node(failover).

Then start up watching pacemaker resources. then cleaned up failaction automatically after 60 sec.

Gs.
  • 540
  • 1
  • 3
  • 21