[Linux-HA] all or none failover

[Lars Ellenberg]

On Mon, Jul 27, 2009 at 11:31:20AM -0700, Cantwell, Bryan wrote:
> Is there some configuration that I am missing then? My services start just fine, but then nothing seems to care if one of them dies. Unless the box can't be reached, nothing happens if httpd service is stopped, for instance .
> 
> Attached are my config files for review of a possible mistake.
> Thanks

> ha.cf:
> debugfile /tmp/log/ha-debug
> logfile	/tmp/ha-log
> keepalive 2
> deadtime 20
> warntime 10
> initdead 60
> udpport	694
> bcast	eth0 eth1		# Linux
> node	master slave
> ping 192.168.56.1
> respawn hacluster /usr/lib/heartbeat/ipfail
> debug 1
> auto_failback off
> use_logd yes
> crm off
  ^^^^^^^

so you do not use crm.

> haresources:
> master 192.168.56.107 firescope_server httpd

> cib.xml:

you do realize that the cib is not used if you do not use CRM?

you do realize that heartbeat v1
does not do resource monitoring _at all_?


-- 
: Lars Ellenberg
: LINBIT | Your Way to High Availability
: DRBD/HA support and consulting http://www.linbit.com

DRBD® and LINBIT® are registered trademarks of LINBIT, Austria.