I am also seeing all sorts of quirky behaviour with process modelling. I have processes that are alerting when they are up and processes not alerting when they are down. If you lock a process so that it doesn't disappear from a device if a remodel occurs when the process IS actually down, sometimes it then seems to "get stuck" and never again shows as up, even though it is running.
I have revisited this area of Zenoss several times over the last 9 months and it has always been pain but often too quirky to document reliably. What DOES zCountProcs do? I have instances of processes where this is et to true and the performance graphs show process counts but I also see count graphs where I have zProcessCount set false. I have a process definition with monitor parameters set to false and it is obviously including instances where my regex is matching something in the parameter rather than the command name. I have had a bit of a dig around the code and cannot find the module(s) that actually check monitored processes for devices against the status that the hrSWRun collector plugin returns (standalone snmpwalk's deliver the expected results). I have run zenmodeler in debug for a monitored process and can see ZenModeler:snmpidx returning correct data for the device but I can't see what then checks this data again processes that should be monitored - anyone got any clues??? Cheers, Jane -------------------- m2f -------------------- Read this topic online here: http://forums.zenoss.com/viewtopic.php?p=33063#33063 -------------------- m2f -------------------- _______________________________________________ zenoss-users mailing list [email protected] http://lists.zenoss.org/mailman/listinfo/zenoss-users
