Index Behavior in a previous copy restore.

[Jason_Bunn]

I had a very helpful individual explain to me how the indexes behave. I am trying to get some validation on this. We are planning on doing some testing ASAP.

Let's say you need to restore your indexes from a previous copy.

First of all, will it look at itself and say OK my last index was (x)?

Secondly, will they look at SQL to determine the last email that was archived?

Let's assume it found that the last email was (x+100).

Next, does the index determine that it is 100 messages behind?

Lastly, do the indexes start poling SQL and Centera to build the missing 100 indexes?

Any validation and or clarification would be great! If this is valid I have not seen any documentation from Symantec on this. Seems like this is very important information when designing DR and failover.

Also, is there a SQL querry that tells you how far behind your indexes are?

Thanks

[Michael_Bilsbor]

When we update an index we always determine how far behind the index is and pull the items out of the archive for them to be indexed.
You can use the ManageIndexVolume utility to kick this off as well.

So you can just restore the indexes and when items are next stored in the archive ie the next archving run, indexing will detect the backlog and update the index to get it up to date.

[Jason_Bunn]

Thank you sir!

We have also comfirmed this in test.

I appreciate your reponse.

[Micah_Wyenn]

i thought manageindexvolume was a 5.0 only tool?

micah

[TonySterling · Accepted solution]

The index behaviour is the same for 5 as it is in 6. For 6 the tool you use is IndexVolumeReplay. If you use the tool be sure to choose Update and not Rebuild. Rebuild starts from nil.Message was edited by:
Tony Sterling

[Micah_Wyenn]

Right on...just didn't want anybody looking for the manage tool if they're using 6.0 and sayin' , "Dammit, where's that tool" and then calling support.

micah