Problem with EV detecting backup

[Vilian_Alexandr]

Hello guys,

I have a problem with EV, it's not being able to properly detect wheather an EV store partition is backed up. EV is set to look for a trigger file, and I have verified that this file is properly modified every time. But EV states that "There is a build-up of savesets that Enterprise Vault cannot confirm have been backed up or replicated." Also, another thing I found is that if you restart the EV Storage service, and then "Run now" the archiving task, it start processing. This is very furstrating, as EV is in clustered config, and every morning I have to manually migrate the services from one node to the other in order to force the processing of the remaining savesets. And the "How to fix" help topic that pops up pretty much says "well, check that the backup is running, and if it is - pray that this time archiving works"

Any ideas?

[LCT]

What version of EV are you running? I sespect it's 9.0.3, if it is then it's more likely to be this issue.

http://www.symantec.com/business/support/index?page=content&id=TECH180841

If you are not using 9.0.3 then you need to dtrace the post process to find out why it is not detecting backup.

[Vilian_Alexandr]

EV version is 9.0.2, but the topic you suggested seems relevant, I will check it and eventually dtrace if the solution given does not apply to my situation. Thanks.

[Vilian_Alexandr]

The above mentioned method did not help... and I saw a new problem today - may events 8391, 8390 and 6469 - hundreds of them. I requested from the DB team to run diagnostics on all related DBs but nothing. I asked the BKP team if the last few backups were successfull - they were. So I'd have to say that the problem would be with the EV itself. Can you please tell me how to set a dtrace (with the proper parameters, I mean)?

[TonySterling]

The trigger file needs to be recreated, not just re-named.  How are you putting EV in and out of backup mode.

For the errors you are now seeing, can you post the text of each error?  The text can vary and it is important to know the details.

[LCT]

ok...technote belows outline how to setup dtrace etc.

http://www.symantec.com/business/support/index?page=content&id=TECH38122

You need to dtrace:

StorageFileWatch.exe

StorageServer.exe

StorageCrawler.exe

w3wp.exe

StorageOnlinOpns.exe

If you can, enable these before the backup is just about to finish, once backup is finished it will trigger the IgnoreArchiveBitTrigger process. hopefully this will capture what is happening.

If not, Enable dtrace, delete IgnoreArchiveBitTrigger.old and create a new IgnoreArchiveBitTrigger.txt and the restart EV Storage Service.

[Vilian_Alexandr]

Event ID 8390 Source Storage Online

The EnterpriseVault.DirectoryConnection object reported an error.

The SQL database server detected a deadlock while accessing the Vault Database '%2'

Description: %1

Source: %3

Number: %4

SQL State: %5

Native Error: %6

HRESULT: %7

SQL Command:%8

=============================

Event ID 8391 Source Directory Service

The EnterpriseVault.DirectoryService object reported an error.

The SQL database server detected a deadlock while accessing the Vault Database '%2'

Description: %1　

Source: %3

Number: %4

SQL State: %5

Native Error: %6

HRESULT: %7

SQL Command: %8

 ============================

[Vilian_Alexandr]

Just a side note - I found in the KBs here SQL queries to check the status of all the DBs and wheather they were backed up or not and not a single error was in there... For some reason EV, however, cannot verify the status of neigher the partition nor the DBs.

[TonySterling]

Please detail for us your backup procedures.

When and how do you put EV in backup mode?

What do you use to backup the Partitions, SQL, and Indexes?

How is EV brought out of backup mode?

[Vilian_Alexandr]

Daily bakups at:

Monday – Friday   4:45 AM and 23:45 PM
Saturday 4:45 AM and 23:45 PM 
Sunday 4:45 am and 23:45 PM

EV is put into and brought out of backup via pre- and post-backup scripts, digging to find the exact scripts

Backup is done by HP Data Protector, using the IgnoreArchiveBitTrigger.txt file

[Vilian_Alexandr · Accepted solution]

guys, we managed to convince our client to upgradeto 9.0.4 last week, and i wasn't able to see any further issues... And another thing that i believe is related - at the time this error first appeard, there were active migrations from a legacy archiving solution to the SEV server. These migrations used quite alot of the client's bandwidth, so perhaps the channel left was too thin for SEV to squeeze all its tasks and this caused delays.