Video Screencast Help
Symantec to Separate Into Two Focused, Industry-Leading Technology Companies. Learn more.

Question around IndexCommited

Created: 22 Nov 2012 • Updated: 05 Dec 2012 | 10 comments
Bruce Crankshaw2's picture
This issue has been solved. See solution.

Hi All

I have a customer on EV 10 sp2 and we are seeing some strange behavior around the Indexing

The users are not seeing consistent Indexes in AE and search. I can see  in the JournalArchive table  the IndexCommited column that there are thousands of items with a value of 0. This tells me that these items have not been added to the Index so it explains the user Index inconsistency. Yet the EV monitoring isn't picking this up. Before the upgrade it would as per the article below

http://www.symantec.com/business/support/index?page=content&id=TECH50268&actp=search&viewlocale=en_US&searchid=1353652979929

 

My question is why would the EV monitoring not reflect these missing items in the Index. We do have a call open with support but they are looking at underlying Index issues and not this specific anomaly

 

Thanks

 

 

Discussion Filed Under:

Comments 10 CommentsJump to latest comment

JesusWept3's picture

I guess the question is, how many items are awaiting to be indexed?
And I'm not entirely sure it does report on them anymore and it's in the system status section of the VAC instead

Does it show in the system status?

Bruce Crankshaw2's picture

Hi Alex

The System Status in the VAC is what I am referring to when I said "monitoring". I should have explained properly.

 

So the issue being EV says system is healthy , green light; but in fact we have Items not committed to Indexing from SQL

Chau Tran's picture

Hi Bruce,

Sounds like a bug with the EVAC to me!!! :-)

In your IndexLocations (physical locations) check and see when was the last item created/modified? Does this match up with info in the SQL?

Regards,

Chau Tran

ASC, STS, SCS and MCITP

DeadEyedJacks's picture

Hi Bruce,

Do you have any index administration tasks running or queued up?  Particularly Upgrades?

This could explain delays in adding items to indexes being processed.

Hope this helps

Authorised Symantec Consultant on Archiving and eDiscovery ASC, STS, SCS, SSE+

Microsoft, NetApp and VMware certified professional MCTS, MCSE, MCSA, NCDA, NCIE-BR, VCP, VTSP

Bruce Crankshaw2's picture

Hi Jack

You have raised some  good points which I need to confirm with the customer

The customer in the past has upgraded some Indexes  to 64 bit and they plan to do more. But I need to clarify something  about why an Index Upgrade would impact the IndexCommited column

  • The customer doesn't put Services into Backup Mode
  • By design new items  always get indexed  to the 64 bit User Index so why would an upgrade of the old 32 bit Index effect this?

 

Thanks

DeadEyedJacks's picture

My experience has been that the system prioritises the index upgrade over deleting or inserting items into the new 64bit index.  Others may know for sure, but my impression has been that the new item inserts don't occur at all until the index upgrade completes.

 

Authorised Symantec Consultant on Archiving and eDiscovery ASC, STS, SCS, SSE+

Microsoft, NetApp and VMware certified professional MCTS, MCSE, MCSA, NCDA, NCIE-BR, VCP, VTSP

DeadEyedJacks's picture

 

Possibly relevant Technotes.

http://www.symantec.com/business/support/index?pag...

http://www.symantec.com/business/support/index?pag...

Authorised Symantec Consultant on Archiving and eDiscovery ASC, STS, SCS, SSE+

Microsoft, NetApp and VMware certified professional MCTS, MCSE, MCSA, NCDA, NCIE-BR, VCP, VTSP

Bruce Crankshaw2's picture

Hi Jack

This is an important point so hopefully we can get confirmation around this.

Thanks for the input around this anyway.

 

JesusWept3's picture

OK so here's what i would do

1. get a SELECT COUNT(*) FROM JournalArchive WHERE indexCommited = 0
2. Go to the Vault admin console,right click your site name and go to properties
3. Go to the Monitoring tab, second item down should be "Archived Items Awaiting Index"
4. What is the "Threshold" set to, is it higher than the COUNT(*) you got from step one?
5. Is it even Checked to monitor for items awaiting index?
6. If you right click "Status" under your "Directory on ...." , click "re-run all checks"
7. Does the alert then come up ?

One obscure thing it could be that i remember in the back of my mind is that you may need to update statistics in SQL, as i seem to remember a case or two when it would complain that it had X amount of items awaiting backup when it was really a lot lower and it was reporting stale information

So in that case i would do this and run the following SQL Statements

USE EVVSYourVaultStore_1
ALTER INDEX ALL ON JournalArchive REBUILD
GO

UPDATE STATISTICS JournalArchive
GO

DBCC FREEPROCCACHE
GO

DBCC UPDATEUSAGE (EVVSYourVaultStore_1, "JournalArchive")
GO

SOLUTION
Bruce Crankshaw2's picture

Thanks Alex, those are some good suggestions