We're managing a two-node v1 cluster.  We programmatically configure
the ha.cf and haresources, then fire up heartbeat on both ends.

What I'm running into is seems to be a wierd timing issue.  As
heartbeat is fired up on both nodes, what we intend to act as a
slave for our resources claims it sees a shutdown notice from the
primary.  But, the primary doesn't seem to generate one.

This using heartbeat 2.1.3 under RHEL5 update 2.

Our ha.cf file:

  logfacility   local5
  ucast eth0 172.20.100.79
  ucast eth0 172.20.100.63
  traditional_compression       false
  auto_failback on
  node  172-20-100-79.sentillioninc.net
  node  172-20-100-63.sentillioninc.net

Our haresources file:

  172-20-100-79.sentillioninc.net       IPaddr::172.20.150.245  mdnsd_vip

We aggregate our syslogs; here's a key snippet.

  Jul 22 08:48:14 172-20-100-63 heartbeat: [5705]: info: Link
  172-20-100-79.sentillioninc.net:eth0 up.

  Jul 22 08:48:14 172-20-100-63 heartbeat: [5705]: info: Status
  update for node 172-20-100-79.sentillioninc.net: status up

  Jul 22 08:48:14 172-20-100-63 heartbeat: [5705]: info: Received
  shutdown notice from '172-20-100-79.sentillioninc.net'.

XXX --^

  Jul 22 08:48:14 172-20-100-63 heartbeat: [5705]: info: Resources
  being acquired from 172-20-100-79.sentillioninc.net.

  Jul 22 08:48:14 172-20-100-63 heartbeat: [5717]: info: acquire
  all HA resources (standby).

Attached is a complete log of heartbeat under these conditions [with
minor annotations] (if attachments are honored by the mailing list
software).

Hopefully someone has some insight on this.  I'm happy to provide
other information if anyone's curious...

-- 
Brian Reichert                          <[EMAIL PROTECTED]>
55 Crystal Ave. #286                    Daytime number: (603) 434-6842
Derry NH 03038-1725 USA                 BSD admin/developer at large    

Attachment: ha-log.gz
Description: Binary data

_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Reply via email to