[Linux-HA] Can't get my resource group to run

[Ask Bjørn Hansen]

On Nov 27, 2006, at 1:28 AM, Andreas Kurz wrote:

[....]
> The most interesting information can be found in the log on the DC,  
> on a
> node that is not the DC there is much less information in the logfile.

Ah, yes -- I see now a few interesting lines among all the cib*  
clutter  :-)

[...]
> Try to add that two parameters to your crm_config, then heartbeat  
> should
> do this job for you automatically:
>
> <nvpair id="stop_orphan_actions" name="stop_orphan_actions"  
> value="true"/>
> <nvpair id="stop_orphan_resources" name="stop_orphan_resources"
> value="true"/>

Yup, that did the trick!

It'd be nice if the gui had those options ...

[...]
> Hmm ... is that "drbd" resource written by yourself, or was the script
> that comes with older drbd installations called "drbd" and not
> "drbddisk" like I have one on my installations?

Ah, indeed!

With "drbd" it was using /etc/init.d/drbd which is, of course,  
stopping the drbd driver rather than just toggling primary/secondary  
status (as the drbddisk script does).   The drbddisk script defaults  
to "all" for start/stop, but that doesn't work for "status" which  
makes it not work if I don't specify the parameter - thanks for the  
hint!

After changing that I noticed
Nov 27 01:50:12 app1 lrmd: [8743]: ERROR: (raexeclsb.c:execra:266)  
execv failed for /etc/init.d/drbddisk: No such file or directory

in the log, which get me to figure out that I should change the  
"class" from "lsb" to "heartbeat" to use the script in /etc/ha.d/ 
resource.d/ (as documented here http://linux-ha.org/v2/Examples/Simple )

(Sorry about the verbosity; documented it for the list archives).


   -  ask

-- 
http://develooper.com/ - http://askask.com/