EV9 poor (journal) archiving performance

[goatboy]

Hi

Environment in EV9.0.2 and Exchange 2003.

We have ~2300 users being journaled to a single mailbox, this mailbox is targeted by journaling archiving in EV - single journaling task.

I ran the numbers over 3 months and we're getting an average of 6891 items being archived per hour by this task, obviously a way off the best practise numbers quoted in the EV9 performance guide (20k-60k). Servers are beefy, plenty of RAM, CPU, storage are (new) Centeras with collections enabled.

The mailbox targeted by the journal task always has a big backlog, we can't seem to archive quickly enough.

Questions:

1. Is it a good idea to up the "number of concurrent connections to exchange server" from 5 to 10 as per the performance guide? What about the "Maximum number of items per target per pass"  (currently set to 1000)

2. Is it beneficial to set up, say, an additional 4 journaling tasks, and split the 2300 users so they journal into 4 mailboxes instead of 1? Any downsides to doing this? I can't find any recommendation in the performance guide around number of journal tasks vs users.

3. Can you have separate journal tasks all archiving to the same archive, or are new archives needed for the new journal tasks?

Thanks

[goatboy]

Thanks Tony, I was just about to clarify that. And thanks Jesus, your advice is much appreciated! I'll run some disk queue stats, but for now, EV performance is acceptable.

[JesusWept3]

Huh, where in the doc does it say that? It contradicts that chart then :\

[TonySterling]

It is in the intro and yes it does contradict the chart!  :)

[goatboy]

Yesterday's stats - 640254 = 26677 per hour! Finally in a range (albeit the lower range) documented in the EV9 performance guide.