[Linux-HA] Problems configuring pingd to failover ldirectord
beekhof at gmail.com
Mon Sep 24 06:02:56 MDT 2007
On 9/24/07, Peter Farrell <peter.d.farrell at gmail.com> wrote:
> Hi all.
> Active / Passive set up.
> 2 nodes, one resource (ldirectord) balancing traffic for IP addresses
> on 2 web servers.
> 2 nics [ eth0 'dmz facing' - eth1 'crossover cable on 10.0.0.1/2]
> *see attached cib.xml / ha.cf
> What I want to happen:
> > Resource prefers to run on dmz1
> > When ping node loses connection - fail to dmz2
> > When ping node resumes connection - fail back to dmz1
> What happens:
> Both nodes running.
> > ifdown eth0 on dmz1 (ping node loses connection)
> dmz2 picks up like a champ, resource migrated.
> > ifup eth0 (ping node resumes connection)
> nothing happens.
when you get it into this state, attach the CIB (with status) here
(the one below seems to be from some time after)
> *at this point - I try to force the issue and;
> > ifdown eth0 on dmz2
not a good idea:
* you'll get a monitor failure
* monitor failures dont automatically mean the resource gets moved
* then you'll then get a start failure (which i see did happen)
* start failures mean the resource gets moved, but also mean the
resource can never go back to dmz2 until you use crm_resource -C
try crm_resource -M instead
> nothing happens - resource doesn't run on either node.
something happened on dmz1 that makes heartbeat think it's not allowed
to run ldirectord there. logs might shed more light on what/why.
> If I 'service heartbeat restart' on dmz1 - ldirectord restarts and
> resource becomes manged again.
> When I run crm_verify -L, I get nothing to std out.
> When run -V is says resource can't be run on other node.
> -Peter Farrell
> -Cardiff, Wales
> Linux-HA mailing list
> Linux-HA at lists.linux-ha.org
> See also: http://linux-ha.org/ReportingProblems
More information about the Linux-HA