Hi! With multiple resources, I had the experience that monitor requests arrive as "big trains", i.e.: A log break, then a lot of monitoring requests, then another big break. It seems LRM cannot disentangle them properly. Maybe the first monito action should be triggered at "start+ random() % monitoring_interval" instead.
Here start-delay could also improve the situation by using "manual randomization". Regards, Ulrich >>> <alain.mou...@bull.net> schrieb am 21.11.2011 um 10:26 in Nachricht <ofb92ee3d1.29cb3dea-onc125794f.00319e96-c125794f.0032a...@bull.net>: > Hi > In the Pacemaker 1.1 Configuration Explained documentation, chapter 5.6, > we always can see "start-delay" attribute, > whereis my search on historical on Mailing List seems to tell that this > attribute is deprecated (or at least discouraged because > if it is needed, that would mean that the RA is somewhere wrong on start > action) > Based on all these remarks, what is today the real status of this option > "start-delay" ? > Thanks > Alain Moullé > _______________________________________________ > Linux-HA mailing list > Linux-HA@lists.linux-ha.org > http://lists.linux-ha.org/mailman/listinfo/linux-ha > See also: http://linux-ha.org/ReportingProblems > _______________________________________________ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
