[ClusterLabs] Antw: Re: Antw: [EXT] Re: crm resource stop VirtualDomain ‑ but VirtualDomain shutdown start some minutes later

Ulrich Windl Ulrich.Windl at rz.uni-regensburg.de
Thu Feb 17 04:18:25 EST 2022

>>> "Lentes, Bernd" <bernd.lentes at helmholtz-muenchen.de> schrieb am 16.02.2022 um
17:44 in Nachricht
<1319350245.178868537.1645029860404.JavaMail.zimbra at helmholtz-muenchen.de>:
> ----- On Feb 16, 2022, at 1:01 PM, Ulrich Windl 
> Ulrich.Windl at rz.uni-regensburg.de wrote:
>> Bernd,
>> I guess the syslog(/journal of the DC has better logs.
> Unfortunately the journal didn't reveal something.

OK, maybe my corosync.conf has different logging:
logging {
        fileline:       off
        to_stderr:      no
        to_syslog:      yes
        debug:  off
        timestamp:      on
        logger_subsys {
                subsys: QUORUM
                debug:  off
        to_logfile:     no
        syslog_facility:        daemon

>> As I see it now, it seems stop of vm_pathway takes a few minutes, and no 
> other
>> action is started befor that is done.
>> I think I once said it "Clusters are not for the impatient", i.e.: Don't 
> start a
>> noew action when the previous action did not complete yet.
> Does that mean when i want to shutdown some VirtualDomains that i have to do 
> this one by one,
> always waiting for the complete shutdown before stopping the next one ?
> That could last very long.

Yes, as "stop" just sets the role of the resource; processing happens in the background.
The crm syntax does not have a "--wait", but you can list all names in one command!
That'll be more efficient in general.


> Bernd

More information about the Users mailing list