Well today was the day. Symantec came on site and we upgraded our systems from 7.0 SP2 to 2007 today. So far things look like they went well. We'll be doing another health check tomorrow and some training/knowledge transfer. The engineer found a few reg keys we were missing for better performance & GUI interaction, got our PDF conversion working, and setus up for "rolling indexes" for users to aid us in cutting down service interruption if we ever have to rebuild a corrupt index rebuilding.
Things I noticed so far.
1. Some backend servers want a reboot, some do not. Why? 13 out of 16 backend servers did not need reboots. No noticeable commonality between those that did.
2. Some confusion on the /PAE switch. We have /PAE enabled (but no /3GB) on our EV servers. One EV server didn't complaint about it, but the other 5 did and would not let us continue until we removed it from boot.ini. The KB article it points you to only mentions /3GB and nothign about /PAE. This should be rectified in some way. Either give official guidance on /PAE, or do not force us to remove it. I re-enabled it after installing 2007.
3. Our front-end servers appeared to upgrade correctly, but testing OWA found that deleting any item from OWA no longer worked. This included non-archived normal items. The only way to get it to work was to uninstall EV, copy the original OWA control files back to their proper directory, delete the "originals copied by EV" directory, reboot, then reinstall EV 2007 from scratch. No clue yet what happened here, don't really care now that it is fixed. Something to keep an eye on for the rest of you though.
4. I'm not quite sure I like the execution of "recover deleted items" yet. I'll ask the engineer in the morning, but it doesn't look like it'll be self-serving for the users. You have to look at the properties of the user's Archive, go to a deleted items tab, and click recover deleted items. You can only recover all items in a single batch, not one by one like one normally can with Outlook's recovery deleted items feature. Perhaps if we give the users the ability to see their archive properties it will be available to them. I'll find out tomorrow.
5. The 7.0 licenses are compatible with 7.5. Even though we were promised the upgrade notices would go out by 8/6, we have not received ours yet.
6. The upgrade proceedure is identical to the 7.0 SP2 upgrade we did two or so weeks ago.
7. The 2007 user extension .msi is another megabyte or so bigger (if anyone cares).
8. The broken "close" button in Archive Explorer seems to have been removed, or I just cannot find it anymore, lol.
9. You can finally set "Users can delete archived items" at the user level instead of at the site level. Woo hoo!!
10. Performance? Tough to say since we are only in pilot mode with 20 users archiving right now. The engineer says it should be far better than it was in 7.0. I don't really expect any issues once we roll it out to production, but we'll see.
11. Vault usage reports seemed to actually be working, they were slow and/or wouldn't work in 7.0. I'll keep trying them.
12. Nothing had to be done to the SQL 2005 SP2 server for the upgrade to work.
13. Clusters definitely make the process a little more intricate. You really have to keep tracks on each step for each node as its easy to miss one. I'd love to see a little more automation to these processes.
14. I really really really wish there was a way to automate the moving of users from one EV server to another. In our environment we are sometimes required to move hundreds or thousands of users from one Exchange server to another. The second Exchange server will not necessarily be served by the same Enterprise Vault server. The only way I see to deal with this for us to day is to a) disable user archiving, b) export their content back to Exchange, c) move the mailboxes, d) re-enable archiving, e) let archiving happen all over again. This is pretty much the epitome of inefficiency. It would be great for somehow if the provisioning task had some logic to say "Oh **bleep**. I know that user, they moved from here to there and another EV server handle that Exchange server. I better start migrating their content over there. Some of you are probably wondering "why in the hell does this matter?". We have to keep track of who's data is where from a legal standpoint and for simple things such as maintenance so we can notify the proper customers that their vaults will be unavailable if something must take place that takes the whole cluster down.
15. I'll know more tomorrow evening. :)
Message Edited by Brian Day on 08-13-200709:53 PM
Message Edited by Brian Day on 08-13-200709:57 PM
Message Edited by Brian Day on 08-13-200710:37 PM