Time required to delete vaults/vaultstore

[Col_G]

I have just removed a couple of legacy vaults and the vaultstore that contained them from ev8 sp2... they are showing as being marked for deleteion in the admin console. and in the OM console the number of incomplete delete operations for that vaultstore is going up and has been for the last 24 hours.. (currently on 625k items)

At what point will it actually start deleting these items? i would like to remove the vaultstore and the associated database to reclaim the space and "tidy up" EV before re-implementing FSA archiving.

Cheers
Colin G

[Col_G]

Bit of an update, incomplete delete operations is now on 1.5mil 

Ive D-traced the storage service and that does seem to be doing its job.. it seems the indexing is holding it up.. looking on the server there is very very little activity which is making me think the indexing isnt working.

Ive tried restarting the indexing service to no effect.

Any ideas?

[MichelZ]

Anything in the eventlog? Try restarting the storage service.

[GertjanA]

Hi Col.

Verify that you have space on the index volume.
Verify no other ' thing'  is locking the volume/folders (think AV-scans, backup?)
Verify EV is not in backup mode for that indexlocation
If necessary, run a trace from within the console (select ev-server with issue, click Tools, Advanced (refresh if necessary)
you should get a ' trace' container underneath services and tasks.
Select new, create trace for Search and Index issues.
Check to see if something strange shows there.
check also taskmanager for index proces etc.

Be patient ;)

[Col_G]

The only things in the eventlog are whinging about mismatched sql collation (again) for a different vaultstore:
I will get this fixed (for the third time) - do you think this could be crashing the indexing service even though its a different vaultstore with an issue?.. there is no messages about the vault im trying to delete and according to VAC the indexes are intact..

Event Type: Error

Event Source: Enterprise Vault 

Event Category: Storage Online 

Event ID: 13360

Date: 08/04/2010

Time: 11:41:46

User: N/A

Computer:

Description:

An error was detected while accessing the Vault Database 'EVEVSMailstore' (Internal reference: .\ADODataAccess.cpp (CADODataAccess::ExecuteSQLCommand) [lines {1388,1390,1405,1428}] built Jul  3 19:52:25 2009): 

Description:  

Cannot resolve the collation conflict between "SQL_Latin1_General_CP1_CI_AS" and "Latin1_General_CI_AS" in the equal to operation.

SQL Command: 

 usps_GetJournalEntriesForIncrSync

Additional Microsoft supplied information:

Source:       Microsoft OLE DB Provider for SQL Server 

Number:       0x80040e14 

SQL State:    42000 

Native Error: 00000468 

HRESULT 0x80040e14

For more information, see Help and Support Center at http://evevent.symantec.com/rosetta/showevent.asp

Event Type: Error

Event Source: Enterprise Vault 

Event Category: Storage Online 

Event ID: 6796

Date: 08/04/2010

Time: 11:41:46

User: N/A

Computer:

Description:

A COM exception has been raised. 

<0x80040e14> 

Internal reference 

.\VaultStoreDB.cpp (CVaultStoreDB::GetJournalEntriesForIncrSyncRecords) [lines {12691,12702,12705,12706,12707,12708,12709,12710,12711,12718}] built Jul  3 20:53:39 2009 

An exception is raised when a process encounters an unexpected fault. 

For more information, see Help and Support Center at http://evevent.symantec.com/rosetta/showevent.asp

[Col_G]

Hi

Theres 25g free on a 100g index vol
AV has been disabled - no change
indexes are not in backup mode
Processes show one index process - IndexBroker.exe with no cpu usage

Ill set a trace running now.

Cheers

[GertjanA]

Hi Col,

My guess is your collation is an issue.
If you run deploymentscanner, does it report on more databases having this?

Get this fixed first. You might get in more trouble than desired. Maybe you should consider logging a call with support.

[Col_G]

Hi...

Strangely the deployment scanner didnt report any issues!

Yes i have had massive issues with collation in the past.. i thought id corrected it, but it looks like the two vaultstore DBs are still using the old SQL2000 collation as their "default" so incorrect collations have probably crept in when i added auditing and FSA reporting a couple of weeks ago.

I have put a change in to run the scripts to correct this tonight... so we shall see if it helps.

Cheers

[Col_G · Accepted solution]

Just thought id update this.

It seems it has finally gone and deleted the vaults without any input from me.... im still going to fix the collation on the other vault at the weekend which will get indexing functioning on that again... it looks like the actual performance of the deletion was very variable.. sometimes it would do 100gig in 2 hours at other times it was less than 10gig an hour.

Im also going to setup a SQL maint plan for the dbs as after looking at some other threads it needs to be done as any table with a index attached is actually too fragmented for the index to be any use!.

Thanks