how large are those POIs?  how many new messages a day are you trying to support?

the first jobs do tend to take quite a while,  the trick is how fast is the nightly update jobs taking to catch up?

You are getting a better rate than I'm seeing on a 2 PO system with lots of large messages such we are averaging 2 seconds per message, which for this show has been fine for a couple of overnight jobs that handle the load in about 2 hours.  This is a system that has been around a while and it on its third 'box' having started out on OES1/SLES9 with Retain 0.9x or there abouts.

Hi Andy,

PO1, First completion shows 5550701 messages processed. 5550701 archived. Still trying to complete second job.

PO2, First completion shows 7960235 messages processed, 6589493 archived. Still trying to complete second job.

HOLPO (Test PO) Nightly shows 152705 messages processed, 300 archived. Takes 5 hours to complete.

The Retain System Monitor never shows the server going above 20% CPU, 57% Free Memory, 60% Free Java Heap Memory.

It's like like the system has to scan each message over each which takes the time.  That's what I gather from HOLPO.

Maybe I need to create archive user groups and add more workers on the PO's?  

I am coming back to your original remark: 8 messages per second. That's not bad - maybe even better than average.

Andy is right when he mentioned that the first run will take a lot of time. You will see much more efficiency in future when only new messages will be archived.

Subsequent jobs need another settings as the initial, you will archive only new items. 8msg / s is pretty quick. It has no point to add another worker. The second job should take no more as couple hours. How much users do you have ?

Hi David,

PO1 has 1620 Users

PO2 has 1058 Users

With this speed it should take no more than couple hours. Did you checked job settings? 

What would be setup wrong in the job settings?

Take a look at the Profile(s) you are using,   under the Scope tab,  'Date Range to Scan' 
  sounds like you have it set to "All Items in Mailbox"  when now you want "New Items"

This is what I have configured for the profile and the thought process behind it.

Date Range Scan:

I selected "Number of days from job start start date and older"

Messages that are "31" days old and older.

"unchecked box" Use last archived item date as start date.

My thought process behind this is our GroupWise auto-empties trash in 30 days.  I don't care to archive what people trash.  Also If I wait to archive I can see what they do with the mail and they can too.  So every message doesn't show in their inbox, it gives them time to move the message into a folder.  What I don't have checked is Use last archived item date as start date.  I wonder if that would help?

Like I said I'm new to Retain so if this is the wrong thought process please let me know. 

Another thing I found was I disabled the workers on the PO servers and added them to the Retain server and now I'm getting 22m/sec on each worker.  That's positive!!!

Hmm ""unchecked box" Use last archived item date as start date."

If you do this, then Retain has to start from the very beginning. If you "last archived item as start date" then Retain will be a lot faster. This bit will be set only for successful archiving (each mailbox). So you will not lose any information.

Under "Advanced" you can define if folders should be omitted. I.e. trash!