Upon checking, the CPU utilization is not that heavy. During peak period, it will only top out until 55% of CPU utilization. There are also times when the server will just crash.
As for the agent manager, it will not manage the scheduled agents accordingly. In fact, agents do not run on time or do not run at all. We enabled seven agent managers to manage around 80 scheduled and triggered agents.
First, is there any way or tools I might use to troubleshoot the slow performance of the Lotus Domino server -- other than analyzing the rip file? Second, why can't the agent managers manage all the agents accordingly? Are seven managers sufficient enough for the agent managers to manage 80 to 100 triggered and scheduled agents? Also, is there a specific maintenance exercise that can keep the agent manager in good shape and well-managed?
You didn't state what version you are on or the operating system, but it seems to me that you are running into the inherent issues of multi-threading in Lotus Domino, or the lack thereof.
Lotus Domino is not the cleanest at launching multiple threads for the same process. The way I have handled this in the past is to stagger those agents out over the 24-hour period on a dedicated agent server.
That way, you can generally get away with three or four agent managers, which is far more tolerable than seven. Finally, make sure your Lotus Domino server document is set up to allow amgr to use the CPU. The default value is 50% of CPU before it slows down amgr or skips agents.
FYI: The setting in the Lotus Domino server document that allows amgr to use the CPU only applies to legacy Lotus Domino servers, version R4.6 and earlier. I don't think there are many of those still around. (I think I would delete the second sentence of this one- and maybe just add 'legacy' before Lotus Domino…)
Do you have comments on this Ask the Expert Q&A? Let us know.
Related information from SearchDomino.com:
This was first published in March 2007