separating index services and storage

hello. thanks for the time. we are upgrading from EV 9 to EV 10 and we want to separate the index from the storage services. we have 2000 users and I was reading if you have more than 4000 then thats when its worth separating and creating an index storage group. If we create one now ...it might not be as beneficial since the server wont be heavily taxed. we also have a separate server for DA and File Share.

the reason that the business wants to separate the index and storage is so that when a search is done is it not done against the index as well as the vault store? or do i have this incorrect? the problem is when legal does a search they get duplicate results and were told that separating this out will relieve that. And my thinking is that since the journal archive mailbox and user mailboxes is on the same vaultstore thats why they are getting duplicates? should we separate that out?

Does this need to be done during the change location wizard during the setup? or can i create an index server group after 10 is installed? Lastly is there any more information on SharePoint archiving besides what is given in the installation guide? any help would be appreciated. thanks.

Enterprise Vault

Information Governance

Upgrade

Jeff_Shotton
13 years ago
Completely agree with Tony.

In relation to the index server groups point, you never stated what the hardware of the existing system is. Index server groups are useful for abstracting indexing from already taxed systems. However it is a one way process, and in your situation, where you appear to have journaling on the same machine, you would likely see much better results by simply creating a new EV server and moving the journaling piece to that. There is a performance penalty for indexing on a remote machine, and if you can avoid this by moving the journaling piece this is likely more efficient.

This would involve creating a new vault store and journal archive in that vault store, but then those indexes would be created on the second EV server rather than the original one which is what you were trying to achieve with your indexserver group.

Regards,

Jeff

Rob_Wilcox1
Level 6
13 years ago
Educating users how to do searches will likely stop the duplicates. If they are using DA and selecting everything in the vault store, and that vault store has user and journal archives, then they are likely to get duplicates, I think.

So you could use move archive to move your user archives to a new vault store.

there isn't anything in the setup or upgrade which will allow you to do that automatically.

you can create a group after installation.

regarding sharepoint have you read the setup guide, as well as the install guide? There are also some videos on these forums which talk about sharepoint. I even think I saw a best practices guide.
TonySterling
Moderator
13 years ago
Seperating the Indexing from Storage will not prevent duplicates. When DA is searching it is searching the Indexs, not storage, so whoever told you that must have misunderstood.

As Rob said, most likely your Legal users are searching all archives. In DA you can configure what archives are searched, so instead of moving all your archives around I would recommend that. Additionally, a little training for the Legal users might be in order. :)

regards,
Tony
TonySterling
Moderator
13 years ago
Oh, for SharePoint have a read of this. https://www-secure.symantec.com/connect/blogs/major-step-forward-ev-sharepoint It has some good info and links.
Jeff_Shotton
Level 6
13 years ago
Completely agree with Tony.

In relation to the index server groups point, you never stated what the hardware of the existing system is. Index server groups are useful for abstracting indexing from already taxed systems. However it is a one way process, and in your situation, where you appear to have journaling on the same machine, you would likely see much better results by simply creating a new EV server and moving the journaling piece to that. There is a performance penalty for indexing on a remote machine, and if you can avoid this by moving the journaling piece this is likely more efficient.

This would involve creating a new vault store and journal archive in that vault store, but then those indexes would be created on the second EV server rather than the original one which is what you were trying to achieve with your indexserver group.

Regards,

Jeff
goof717
Level 5
13 years ago
thank you all for the quick response. i will go back to the business with the info provided here. what intrigues me is that i can separate the journaling and archiving that is currently shared by a single server to two different servers. the new servers are VMs. 64bit 2008R2. 16GB of memory. that is brilliant. thanks jeff.
goof717
Level 5
13 years ago
follow up question... if we separate out the journal and archive mail to different vault stores can users search both stores and only get a single result? and not duplicates for mail that resides in both archives? makes sense?
TonySterling
Moderator
13 years ago
No, anytime you are searching both journal archives and mailbox archives it is quite possible to get duplicates or more.

If your journal archive has all the mail from the date range you are searching you should only be searching it. That is why you set up journaling, to be able to capture all mail in\out of your organization.

DA does have some de-duplicating abilities now which should help in review and export, but you still will get the duplicate hits in the results.

See

About the deduplication feature in Discovery Accelerator

Article:HOWTO74920 | Created: 2012-03-30 | Updated: 2012-09-22 | Article URL http://www.symantec.com/docs/HOWTO74920

and

White paper: Deduplication in Compliance Accelerator and Discovery Accelerator

Article:DOC3621 | Created: 2011-02-10 | Updated: 2012-03-27 | Article URL http://www.symantec.com/docs/DOC3621
Rob_Wilcox1
Level 6
13 years ago
Typically your users (end-users) wouldn't be searching your journal archive right?
TonySterling
Moderator
13 years ago
I believe he is talking about Legal searching and getting duplicates, from the first post:

the problem is when legal does a search they get duplicate results and were told that separating this out will relieve that
goof717
Level 5
13 years ago
correct...legal searches. Tony that's a great point. Again I appreciate the time and knowledge from you all. Thank you.