We are running two Exchange 2013 CU7 CAS/MBX servers. On both servers we see once a day a CPU spike caused by the MSExchangeMailboxAssistants:
We are seeing this spike on both servers (both servers have 2 active databases en 2 passive databases). The problem is that during this spike users are experiencing performance problems with their Outlook client.
As far as I understand the Mailbox Assistance is kind of black box process which can be configured and should only run when it doesn't affect the performance. However, for us this theory doesn't match real life.
When I look in the eventviewer I see the following related events (about 300 on each server):
| Information | 3/17/2015 18:31 | MSExchangeMailboxAssistants | 25001 | -25 | The people relevance feeder for mailbox /o=Organization/ou=Exchange Administrative Group (RXDIAEHF18SPDLT)/cn=Recipients/cn=J.Doe picked 3 sent items to process. Number of sent items successfully processed 3 and number of sent items failed to be processed 0. Total time taken to process this mailbox 408.9071. |
Can someone explain what kind of maintenance is performed and how I should interpret the events? The time, is it 408 milliseconds or seconds? Is there a way I can limit or tune the MSExchangeMailboxAssistants process?
Other recent topics
