cancel
Showing results for 
Search instead for 
Did you mean: 

Vault items "Awaiting Backup" not cleairing down

James_Slack
Level 6

Hi,

There is a web page I use to monitor items awaiting backup, the Vault Store Usage Reporter.

This is indicating we are now up to 6 million items awaiting backup - and growing!

We use the trigger file method, as out current backup solution doesn't work correctly with Vault to reset the Archive Bits.

I have a script in place I manually run, it deletes all the old archive trigger files and creates new ones. But something is not right, it doesn't clear down any more.

I checked that all partitions are looking for the trigger bit and the file is going to the correct location... all looks right to me and this has been working fine for over a year.

When I first started looking at this, I stumbled across a table in SQL that was a list of all the items awaiting backup, so I could see which partition was having backup issues. I can't find this now.

Can anyone please tell me which table I need to look at, or maybe it was a query.

I need to work out which items are not getting maked as backed up.

Thanks

James

 

 

1 ACCEPTED SOLUTION

Accepted Solutions

GabeV
Level 6
Employee Accredited

James,

Take a look at this technote:

SQL Index Fragmentation and Statistics
http://www.symantec.com/docs/TECH168905

Basically, you need to run the following SQL query:

DBCC SHOWCONTIG WITH ALL_INDEXES

The technote above has more details of what you need to look for in the output. Keep in mind that if you have high fragmentation, you need to run a SQL maintenance to fix it. That will improve the EV performance. If that's the case (i.e. there is high fragmentation in SQL) , you need to read this technote:

How to automatically back up and perform recommended maintenance for Enterprise Vault SQL databases
http://www.symantec.com/docs/TECH74666

Forward it to the team that is in charge of the SQL server maintenance to confirm if they are following the correct steps.

View solution in original post

14 REPLIES 14

RahulG
Level 6
Employee

You may have to fistly make sure the backups are successful and the archive bit on the File on the partitions are getting reset. You may refer the the following document

http://www.symantec.com/business/support/index?page=content&id=TECH69486

you can runthe attrib command to reset the archive bit on the files 

James_Slack
Level 6

Sorry, should have stated this more clearly and checked my terms. We use the "Check for trigger file" option as our backup software DOES NOT reset the arvice bit. 

This is working. I delete the .OLD, create a .TXT, restart the storage service and it changes them all to .OLD again...

But it still shows files awaiting backup...

 

James_Slack
Level 6

Nope, not that issue.

The .txt is being created with the current time it is run. This is being converted to .old OK but still nothing clears down.

I know there is a table in SQL you can look at to see what is waiting to be backed up. It maybe a table off all the data and you filter it by if it has been maked as backed up or not, I can't remember.

I will look at the DTRACE guide in that link, but I am pretty sure there is a much easier way to do this.

James_Slack
Level 6

Hi RahulG,

I went to have a go at the DTRACE following the details given in your link, but I guess it has changed in EV10.

There is no "Storage File Watch" option in EV 10. Any idea which DTRACE option I should run?

I am testing with "Items remain archive pending", but I don't think that's right (it may be, as you can tell, I'm no pro!)

Thanks

GabeV
Level 6
Employee Accredited

James,

In EV 10 you should have StorageFileWatch as part of the processes to dtrace. If you run a dtrace, you shiould able to type:

DT> set StorageFileWatch v y 

... to add the process to the dtrace list. If you confirmed that the .old trigger file has been deleted and that the .txt file is being created in the root of the partition, then you need to run the dtrace. Also, you don't need to restart the storage service every time you take a backup, You can just use the powershell scripts to set/clear backup mode in the Vault Stores and indexes.

Follow these steps: (1) Set backup mode, (2) take a new backup, (3) make sure that the .old file is removed, (4) then that the .txt file is created, (5) run a dtrace and (6) clear backup mode. You might need to let the dtrace running for a few minutes. You can post the dtrace here to review it but you might want to open a ticket with support to troubleshoot this issue further.

I hope this helps.

RahulG
Level 6
Employee

It should be there, make sure you are typing it right StorageFilewatch without any spaces.  I hope you are checking it on the server where you have the storage service .

In the  Watch File tables do you see  as ItemSecured = 'True or false'. 

Did you tried created a new trigger file and see if that help.

James_Slack
Level 6

Hi,

The DTRACE instructions as per the link above were throught the GUI, so I used the "Items remain archive pending" which did actually cover Storage File Watch.

As a backup has been done already and we can't get 6 million to clear down I did the following.

1) Start trace

2) Delete all .OLD files

3) Generate new .TXT files

4) Stop Services

5) Start Services

6) Watched the Vault Store Usage Reporter. After a while, files did start going down, but stopped at about the same as earlier, so I believe one partition is having issues.

I am now looking through the DTRACE log (80+MB) and will update if I find something that hints at which partition is having the issue.

I just checked, again, all 48 partitions .TXT files got changed to .OLD.

GabeV, We could never get it to work with the scripts. We always had to stop and start the Storage service. Will look at that again once I have this sorted. Thanks for the heads up.

GabeV
Level 6
Employee Accredited

Check the fragmentation level in the SQL Server databases. If you verified from the Vault Store Usage Report that the items waiting for backup start decreasing but then stopped, I'd check the SQL databases fragmentation, just to make sure that everything is OK on that side.

James_Slack
Level 6

Sorry GabeV, I have no idea how to do that. I have access to the SQL server. Is there a command I run or is this a more generic SQL thing?

We have an external company who support our SQL, so my SQL experience consists of running the odd query against Vault.

GabeV
Level 6
Employee Accredited

James,

Take a look at this technote:

SQL Index Fragmentation and Statistics
http://www.symantec.com/docs/TECH168905

Basically, you need to run the following SQL query:

DBCC SHOWCONTIG WITH ALL_INDEXES

The technote above has more details of what you need to look for in the output. Keep in mind that if you have high fragmentation, you need to run a SQL maintenance to fix it. That will improve the EV performance. If that's the case (i.e. there is high fragmentation in SQL) , you need to read this technote:

How to automatically back up and perform recommended maintenance for Enterprise Vault SQL databases
http://www.symantec.com/docs/TECH74666

Forward it to the team that is in charge of the SQL server maintenance to confirm if they are following the correct steps.

James_Slack
Level 6

Ah, is the table I was thinking of DBO.Watchfile?

Is there a way to query this for ItemSecured = no, or something?

James_Slack
Level 6

Think this is it: SELECT * FROM [EVMailboxVaultStore].[dbo].[WatchFile] where ItemSecured = 0

Running now, will see how many I get.

Will look at SQL health in a minute!

Thank you

James_Slack
Level 6

Eventually this just worked, needed to redo the delete and create of .txt / .old about 8 times to clear all 6mil files. Had actually changed nothing, but used my above SQL to detirmine the stores causing the issues and confirmed they were all fine.

I suspect this was ultimately down to performance due to the SQL side of things and will make that as solution. Though I expect that it will ultimately be a can of worms, now open.

Thanks both for your help, sorry I can't make you both as the solution.