4 errors occur constantly in the Application Log on the KVS Vault Server

[Brian_Lynch]

Version 5 CP3 - Window Server 2000 w/ SP4

First Error:
A low level indexing operation has failed
Error: an internal error
Index Id: 1A85090AE31C5C5419EA59E9A2E5FEE831110000mtkes03.ams.ammd.com
Source: Enterprise Vault
Category: Index Server


Second Error:
Abnormal error occurred

Error: A low level indexing operation has failed Error: #1 Index Id: #2
Reference: CAVUpdator/OAVI/o(h)
Index: 1A85090AE31C5C5419EA59E9A2E5FEE831110000mtkes03.ams.ammd.com
Info: IndexExists ;
Params (ignored_thresh=1000; chars_before_wildcard=3; unlimited_wild_words=1; indexformat=-1; cache_threshold=268435456; options=39; charset=1; ntiers=0; nbuckets=0
Source: Enterprise Vault
Category: Index Server

Third Error:

Could not get the IndexServer sequence number: The Alta Vista index is not open
Internal reference: GetSequenceNumber
Source: Enterprise Vault
Category: Storage Replay Index


Fourth Error:
Error updating an index.


%3
Source: Enterprise Vault
Category: Storage Replay Index


Any ideas on how to fix this would be greatly appreciated. Like I stated in the Subject, these 4 errors occur a few times a minute so it fills up my application log quickly.

Thanks,
Brian Lynch

[TonySterling]

Is it the same index id all the time? If so, you probably have a corrupt index. You can rebuild it using the IndexReplay.htm found in the root of Enterprise Vault folder.

[Michael_Bilsbor]

Check the actual directory that the index is stored in.
In there is a file called dataset.
Does it have any content at all or perhaps just one line or lots of lines.
Also check avtrace.log see if that reports a problem opening a file.

Is the index stored on a NAS? (or basically via a UNC path?)
Have you got the /3gb switch set in boot.ini
Is the index location being scanned by AV software?

[Brian_Lynch]

Thanks to both of you for your replies. I've posted the results below:

A_WS,
Yes it seems to be the same index everytime. I opened the IndexReplay and tried to run it on that index and I get an error "The Selected Archive is unavailable. Please retry later". I've tried against other indexes and it works fine. Any other ideas?

The Dodo,
I don't see a file named dataset. When I try to open avtrace.log I get an error "Not enough memory available to complete this operation. Quit one or more applications to increase available memory, and then try again". I can open other people's indexes (avtrace.log) just fine. I'm not using NAS or a UNC for file location. The Directories are all local on the server. I'm also not using any virus protection on the server. Hopefully I'm making sense. Any other ideas?

Message was edited by:
Brian Lynch

[TonySterling · Accepted solution]

you can rebuild it manually:

Rebuild the Indexes:
1. Stop all Enterprise Vault services except Admin, Directory, Indexing, & Storage.
2. Once all MSMQ queues are finished processing, then stop the remaining services.
If they do not empty, you can purge using the steps in the "More Information" section below.
3. Browse to your Indexes.
4. Find the Index -You can look on the Advanced tab of the Properties of the archive
Example: 13B59F5821C8C764F9834C867530951111110000kvsvault.
5. Rename the index to xxx13B59F5821C8C764F9834C867530951111110000kvsvault (just add 3"x"s to the beginning).
6. Launch SQL, go to the Vault Store database EVVaultStore, and open the ArchivePoint table.
7. Look for ArchivePointId of 13B59F5821C8C764F9834C867530951111110000kvsvault and change the ReplayState to "1".
8. If the Replay state is already set to "1", then set it to "0" and purge the "vault storage replay" queue
(If you need procedures for purging the queue, refer to steps in "More Information")
9. After the queue has been purged, then reset the replay state to "1"
10. Restart all of your services, the Index will rebuild itself.
This will take some time.
11. You will see messages in the App log such as "Processing Transaction ###### of ######",
this way you can monitor the index rebuild process.
12. The time it takes for the rebuild process to finish depends on the size of the index, so it may last several hours.


MORE INFORMATION How to purge MSMQ Queues:
1. From the desktop of the message Queuing server, right click the "My Computer" icon and select "Manage"
2. Expand the following: Services and Applications>Message Queuing>Private Queues>Enterprise Vault Storage Replay
3. Right click on "Queue Messages" and select All Tasks>Purge

[Brian_Lynch]

Thank you A_WS, that did the trick. It is running great now without a problem.

Thank you also The Dodo for helping out.

[TonySterling]

:)

Happy to help!