[Pacemaker] Pacemaker / Postfix startup problem...

[Raoul Bhatia [IPAX]]

hi adam,

On 14.04.2011 18:10, Adam Reiss wrote:
> Hi Raoul,
>
> We're trying to setup a HA SMTP Relay, so having pacemaker stop/start
> the services as it passes the work over to the other machine, should
> Postfix fail...  Is there a better way to allow an HA SMTP relay?

when we're setting up a clustered postfix, we do not mess with the
default /etc/postfix/ config but use a different location on a drbd
backed deviced instead.

e.g. /data/mail/

this way, local mail deliverey (cron output!) works without any issue -
even if the clustered postfix is down (e.g. for maintenance) or simply
migrated to a different host.

> It's running under VMWare, having two different guests, on two different
> hosts...
>
> I've attached the output you've requested. :)
>
> There is no syslog file in /var/log .

mhm - your hb_report is incomplete too. i don't know centos - where does
centos' syslog write it's logfiles?

anyways, i've updated the postfix ocf ra to handle some configuration
cases and errors:

 
https://github.com/raoulbhatia/resource-agents/tree/master/heartbeat/postfix

depending on your system, you might need to apply the following patch:

-: ${OCF_FUNCTIONS_DIR=${OCF_ROOT}/lib/heartbeat}
-. ${OCF_FUNCTIONS_DIR}/ocf-shellfuncs
+: ${OCF_FUNCTIONS_DIR=${OCF_ROOT}/resource.d/heartbeat}
+. ${OCF_FUNCTIONS_DIR}/.ocf-shellfuncs


could you please give it a shot and report whats happening?

if it is still *not* working for you, i would need your current
configuration, a new hb_report and the system's logfiles.

thanks,
raoul