EV 10.0.4 CH3 Mailbox Archive Server memory 98% utilized

The EV Archive server has 7000 user and shared mailbox archives on it.  We have 8 Exchange 2013 servers each with about 1000 user mailboxes and shared mailboxes.

How many users do you have on this EV server? You might have outgrown the server and its starting to run out of resources.

And you are targeting all these exchange server swith one EV server? Are you also journaling? All this could be the reason for your issues

your server is simply running out of resources. could be a combination of mapi, iis, msmq, sql, storage, network connections, etc. two main issues i see:

1. 8 Exchange servers is too many for a single EV server to archive

2. 7000 users is beyond what i would recommend for 1 EV server

I have 8 archive tasks setup from one EV server.  The archiving runs from 6pm to 12am.  In that time it hits all mailboxes.  The Exchange Journaling is done on a different EV server.

The server is sized pretty good.  16cpu and 32gb of Ram.  Hitachi SAN Storage.  I have 23 mount point partitions at 1TB each. 

If the archiving is running at night why am I out of resources during the day time when its just taking iis requests.

Next steps would be to ensure AV exclusions are correctly set per this article.

https://www.veritas.com/support/en_US/article.000032085

the truth is we see this all the time. cpu and ram are great but there are other factors too. if veritas or an authorized partner designed your EV architecture like this then i'd go back to them and let them know it's not working. what typically happens is the system is designed to accomodate a certain environment and then the environment changes but the system isnt redesigned to accomodate because everything seems to work fine until a random issue like this pops up.

as far as best practices go, you're pretty far outside of them and i would recommend adding a second EV mailbox archiving server, splitting the 8 archiving tasks to 4 on each server, and eventually moving the related archives (3500 or so assuming even distribution in Exchange) to the second EV mailbox server.

another option would be to increase your archiving window and stagger your 8 archiving tasks' schedules. also look at the number of connections per exchange server per task (in the task properties) and make sure it's not above 5.

lastly, if you upgrade to EV 11 and Windows 2012 you can take advantage of Outlook 2013 (instead of 2007) and an OS with better resource management.

by the way, 23 mount points and all that HBA traffic are also eating up resources outside of cpu and ram.

We were running into severe resource (memory) issues when doing searches as well.  Upgrade of the SQL server and higher performance SAN solved our problems.

Might want to look into SQL fragmentation and statistics then if upgrading is not an option:

https://www.veritas.com/support/en_US/article.TECH168905