Hi
We have 3 Management Servers - one in each of our 2 sites (one AD Forest) with AD Auto Assignment rules so that each site's SCOM 2012 agents are assigned to their local MS, with failover manually configured in these rules to only failover to the opposite agent managing MS. The 3rd MS has been removed from all Resource Pools and is used as a dedicated Console server (Web Console, Full Console, plus all remote consoles point to this one to lessen the load of the agent managing Management Servers - not entirely sure if this is supported/recommended however we were conscious of load prior to standing this 3rd MS up, and since doing so have seen an improvement in console user experience).
My actual question is:
I have read that putting Management Servers into Maintenance Mode in SCOM 2012 is now possible http://windowspowered.wordpress.com/2011/09/09/management-server-maintenance-mode-in-scom-2012/), however I see conflicting comments on various Microsoft articles stating not to (http://support.microsoft.com/kb/2704170/en-nz), due to workflows stopping, heartbeat failures from agents etc.
Putting our non-agent managing MS into Maint Mode is not a problem. We don't see any alerts when we do this.
Putting our agent managing MS's into Maint Mode causes Heartbeat failure alerts for all the agents reporting into the said MS. These Heartbeat alerts continue to exist until the MS comes out of Maint Mode (even if the MS is back up and running). I therefore presume that Maint Mode is still stopping workflows that the agents expect to be available due to their primary assignment configuration.
The behaviour I see is about 2 mins after rebooting the MS, the agents all alert for Heartbeat failures, then about 20 secs later the alert resolves. This is when they successfully failover to the alternate MS (I see this happen in their Event Log).
About 7-8 mins later (when the MS has restarted), another set of heartbeat alerts occur for the same agents, and I see they are now aware the server is back online, attempt to fail back, then due to Maint Mode, hit a wall and failover again.
This behaviour shows me that this is not a good idea to put these MS's into Maint Mode, however every month when we automate our SCCM Patching reboots, we don't want these MS's to alert when they reboot. Not putting them into Maint Mode will still cause all their agents to alert with Heartbeat failures in this scenario!
What does everyone here do in regards to rebooting Management Servers without causing an influx of alerts?
Thanks in advance
Steve