Hi,

On Tue, May 27, 2008 at 08:53:27AM -0700, Joe Bill wrote:
> Would it be possible to have an execution flow diagram
> of what calls to a resource's RA HB performs, and if
> it depends on the resource's state itself ?
> 
> Typically, I assume that HB *always* performs a
> 'monitor check-level 0' operation *before* performing
> a 'start' or a 'stop'.
> Is this correct ?

Not always. The cluster (crm) keeps track of the status. If it
doesn't know the status of a particular resource then it would do
a probe (i.e. a monitor op) to check it. For example, that
happens when a node starts. Or after the resource has been taken
out of the unmanaged mode.

Hence, it is very wrong to mess with the resource without letting
the cluster know. Though there's also an external program
(crm_resource) which may be used to ask the crm to refresh the
state of the resource (option -P).

> I also assume that HB performs a 'monitor check-level
> 0' *after* a successful "start" or "stop".
> Is this correct ?

No. If the RA exits with success (0) the action's considered to
have been successful. (Hmm, how does that sound :)

> Or, I assume that, if HB performs a 'monitor
> check-level 0', and that operation returns a "ERROR"
> status, HB automatically performs another 'monitor' of
> the same resource but with the next check-level known
> by HB, provided such a check-level (i.e. 10 and 20)
> have been previously declared (configured) for the
> resource's monitoring, until a monitoring operation
> returns a clear "STARTED" or "STOPPED" status.
> Is this correct ?

No.

> Or does HB perform "monitor" operations at *all* check
> levels regardless of the status returned, of course
> outside the periodic programming of monitoring
> functions ?

Various check levels are done only periodically and only when
they are defined (and enabled) for a resource.

> Also, outside of the periodic monitoring at those
> different check-levels configured, HB does not perform
> a higher check-level monitoring operation if one
> monitoring operation returns a "STARTED" or "STOPPED"
> status, and that status matches what HB believes is
> the current status for that resource.

No. All monitor operations are carried out when they are
scheduled. If any of them results in a failure (or reports that
the resource is not running) some action may be taken
immediately (typically failover or restart).

Thanks,

Dejan

> Is this correct ?
> Thank you for your replies.
> Is 
> 
> 
>       
> _______________________________________________
> Linux-HA mailing list
> [email protected]
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems
_______________________________________________
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