On Saturday 30 June 2007, Lars Marowsky-Bree wrote:
> On 2007-06-30T05:59:13, YAMAUCHI HIDEO <[EMAIL PROTECTED]> wrote:
> 
> > In this situation, is there the method how it is easy to return the resource
> > that failed in a start to an original node?
> 
> crm_resource -C can be used to cleanup the failed start. 
> 
> It's an on-going discussion between Andrew and myself whether start
> failures should not simply increase the failcount by a configurable
> amount as well - if this defaulted to INFINITY, the current behaviour
> would still apply, but finer granularity would also be possible.

Our customer did not understand why a fail count is only raised if
the resource was running and and failed and not also when it failed 
to start. 

The worst impact on the fail-start is that this state is not shown in 
crm_monitor. So if we have a resource group and one of those
resource failed to start they have to manually go through all
resources with crm_resource -C (there is no chance that someone
who is not used to the cluster environment to explain the output 
of ptest or crm_mon -VVVVVVV...).
_______________________________________________
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