Video Screencast Help

Design Considerations for Exchange Mailbox Archiving

Created: 30 Jul 2013 | 7 comments

I am getting ready to deploy Exchange Mailbox archiving and would appreciate it if you all could let me know if you see any problems in, or have any suggestions about, the following design for user interaction with EV;

 

BACKGROUND

 

I work for a small governement organization with < 150 mailboxes, About 6 months ago I finished a pretty massive project that involved restoring 10 years of GroupWise backups, exporting the content to PSTs, and then importing those into an EV shared archive for legal disovery purposes. At the same time we switched over to Exchange and began journal archiving with EV, which has been working very well ever since. I originally designed the EV system to accomodate mailbox archiving and it is now to move forward. Since we are so small, we are using a standalone EV server with a separate, deidicated SQL server. Both servers are physical. I have followed the Best Practices when designing the system and I'm pretty sure that the resources are underlying design are adequate for our needs from a performace and capacity perspective. The environment, including backups, is currently 100% stable. The only real unknown is seeing how performance holds up as users begin to hit their archives, especially if we also have DA case in progress.

 

Our desktops are 100% VDI, with a few laptops that are used to connect to virtual desktops through a VPN. Users also use OWA to connect to their email remotely through a reverse proxy. All user's use Outlook 2010 in on-line mode only, so we will not be using virtual vault at all. We have some people who have been here for a long time and who have very large mailboxes. This has been exacerbated by the fact that the State of Arizona has essentially no electronic records retention policy and many people have been told by our AG to keep everything forever. (The AG's office actually suggested that people "print every piece of email and file it" for retention purposes.) Legal retention is no longer an issue with journal archiving, so the primary purpose of mailbox archiving in our envirnoment is to get control over the size of mailboxes and to start expiring messages.

 

USER ENVIRONMENT

 

I want to keep everything for our user's as simple as possible. I am currently planing on deploying EV to them as follows:

  • Run the add-in in light mode with no Virtual Vault.
  • Manual Store, Restore, and Delete enabled.
  • Archive Explorer and Search Vaults enabled and running inside Outlook 2010.
  • Use a single retention policy for all mailboxes (currently 5 years, although I will try to reduce this).
  • Enable shortcut deletion after 18 months (remember this is an evironment where people have never had to delete anything).
  • Archive based on age only greater than 1 year to start. Switching to age + quota if we can get the MB size under control to the point where we can implement quotas.
  • Provide our users with some control over retention by creating folders with EVPM, assigning different retention policies to them, and configuring EV to modify the retention policy when a shortcut is moved. I am planning to run the EVPM job on a schedule and run it against the group used to provision mailboxes.

I realize that the initial archive runs may take quite a while before the system has caught up so I plan to archive nightly to start and perhaps switch to weekly schedule down the road since we have <10000 items going through the journal mailbox every week. I am going to be leaving this position in a few months and want to create a system that is as simple to use, manage, and maintain as I can make it.

 

I would apreciate any comments.

 

Thank you.

 

Operating Systems:

Comments 7 CommentsJump to latest comment

Rob.Wilcox's picture

Sounds like you have most things covered in your ideas/designs there.

Sticking with the 'simpler' search and archive explorer will mean that your users when they use OWA will still be able to access their archived content in the same way that they do when they're using Outlook.

GabeV's picture

You might want to consider collections enabled in the partitions to speed up the backup process. Other than that, I think you have everything covered. The next person in charge will be really happy.

“Success is not final, failure is not fatal: it is the courage to continue that counts.”–Winston Churchill

Pradeep_Papnai's picture

Overall the strategy for archiving single exchange 2k10 server consist of approximately 150 mailbox + journal looks good.

I would say, take regular backup & run SQL maintenance regularly. You also think to have dedicated box for DA server. Disable Archive Explorer / search for user if they use OWA so you don't need to publish EV server externally but archiving/retrieval will still work.   

Keep monitor EV services, other resources and if any error / warning comes then fix it as soon as possible before problem takes bigger shape.

 (EV monitoring feature can be used http://www.symantec.com/docs/HOWTO36998 Enterprise Vault Operations Manager, http://www.symantec.com/docs/HOWTO27843 Accessing Enterprise Vault Operations Manager )

Please also have a look performance guide http://www.symantec.com/docs/DOC4553 for EV 10.0.

Additionally it would be worth to have health check of environment & take inputs for best design for EV archiving from Symantec Partners/Symantec professional-support team.

PeterWendell's picture

I really appreciate you taking the time to respond. I'm glad that you guys didn't see any real problems.

 

I already have collections enabled on all the partitions. This became necessary early on when I saw how long full backups were taking on the partition containing the shared archive that holds all of the imported GroupWise email. It's only 5 million items, but still...

 

I have a SQL maintenace plan created and tested according to the Best Practices documentation, and just need to work with my backup guy to find a spot in the schedule to run it. Is it really necessary to shutdown the EV services before running the SQL maintenance, or is enabling backup mode sufficient?

 

I've been monitoring the EV system daily since launch and by now I have a really good idea of how it is working. The monitor is very helpful. I've noticed from doing this and reading this forum, that EV is mostly very reliable, but problems can go undetected if you are not actively looking after it and go from being minor to being a nightmare.

 

Currently DA is installed on the EV server. We do not have a lot of legal discovery going on consistently. In fact it hasn't been used in earnest since I created this system. (We get sued consistently, but not regularly.) From what I've been able to determine so far in testing, it looks like that at the worst I will need to schedule the DA searches and analytics processing to run during off-hours. This should not present a problem for us. If we see that the peformance isn't holding up for the mailbox users while a case is in progress, my successor will need to make changes.

 

Thanks again for your input.

GabeV's picture

One more thing that you need to take in consideration is the optimize settings for EV, CA, DA and SQL Server:

http://www.symantec.com/docs/TECH56172

I hope this helps.

“Success is not final, failure is not fatal: it is the courage to continue that counts.”–Winston Churchill

Pradeep_Papnai's picture

HI Peter,

As per TN http://www.symantec.com/docs/TECH74666, Symantec recommends regular backups of all EV, CA, and DA SQL databases and During backups, We should place the EV services in read-only or backup mode, and stop CA/DA services. Other EV services are not require to be stop.

Regards

EV-C

PeterWendell's picture

Thanks for the replies.

EV is currently being backed up by the Netbackup EV agent according to the Best Practices, which takes care of the stores, indexes, and databases, except for DA. The DA data bases will be backed up through A SQL maintenance plan.

 

From reading over the links (I had already applied the optimization settings) it does sound like all the services will need to be stopped while the SQL maintenace (shrink, index rebuild and statistics update) is run. I'll get this queued up this weekend for a test.

 

Thanks again.