Backing Up issues, watchfile table not cleared
Created: 25 Jun 2010 | Updated: 10 Oct 2010 | 8 comments
This issue has been solved. See solution.
Hey Folks, I am a bit stuck with a customer's Problem of having elements awaiting Backup.
I Checked the following things
- Archiveable Attribute is cleared by the backup Software
- Backup Mode is set and Cleared by the Backup Software (Powershell commands)
- I see that EV is checking for secured Items --> Watch file partition scan restarted for all the partitions on a VaultStore as it is exiting the backup mode.
- this is happening on all vault stores
- I manually set and cleared the Backup mode
- I restarted all services
- One Vault store cleared the watchfile table at some time for some reason
I have some elements in a closed partition that are marked in the watchfile table, could that be that they are not checked?
EV System is EV 8.0 SP4 on a Windows 2003 machine, separate SQL 2005
Any Ideas appreciated
Thanks a lot
Holger
Discussion Filed Under:
Comments 8 Comments • Jump to latest comment
Hello Holger,
I assume you are backing up the closed partitions also?
What are the settings in the backup tab for the partitions? Archive Bit or triggerfile? (i used triggerfile)
If triggerfile, make sure the file's date and time are of today (ie a create new instead of a copy)
I had an issue where in 1 partition it was not removing items, and it turned out that the Storage File Watch has a memory leak (check the documents, it is acknowledged, no news on this yet). Support helped me, and I set the registry keys Critical and Warning on the adminservice (reg.pdf page 24 and 29). That allowed the 'items awaiting backup' to drop slowly. I had (not totally sure of numbers) over 2 million items for a store, it dropped to the normal level in 3 weeks (!).
I used the 'manual' way of clearing this partition:
Set the specific store in backupmode.
created new triggerfile
cleared backupmode
watched taskmanager to see when Storage File Watch failed
Set the specific store in backupmode.
created new triggerfile
cleared backupmode
etc.
Once the critical and warning were set, SFW ran without problems. (but, it took a long time to proces, I did do several steps as indicated above)
Thank you, Gertjan, MCSE, MCITP,MCTS, SCS, STS
Company: www.t2.nl
www.quadrotech-it.com
www.symantec.com/vision
http://seer.entsupport.symantec.com/docs/287955.htm
Thank you, Gertjan, MCSE, MCITP,MCTS, SCS, STS
Company: www.t2.nl
www.quadrotech-it.com
www.symantec.com/vision
Hi gertjan, thanks für the links, I'll check with the customer and let you know!
Cheers
Holger
----------------------------------------------------------------------------- Please grant me the ability to punch people in the face over standard TCP/IP
Hi Gertjan,
yes all the vault stores, even the closed ones are backed up.
we are using the Archiveable attribute, which is correctly cleared by the backup software.
I don't see any errors in the Log, maybe I'll have to set up a DTRACE, have to check for what to trace :/
Thanks for the Input, we'll check the issue...maybe even involve symantec support.
cheers
Holger
----------------------------------------------------------------------------- Please grant me the ability to punch people in the face over standard TCP/IP
Hi Gertjan,
yes all the vault stores, even the closed ones are backed up.
we are using the Archiveable attribute, which is correctly cleared by the backup software.
I don't see any errors in the Log, maybe I'll have to set up a DTRACE, have to check for what to trace :/
Thanks for the Input, we'll check the issue...maybe even involve symantec support.
cheers
Holger
----------------------------------------------------------------------------- Please grant me the ability to punch people in the face over standard TCP/IP
Would it be possible to get a DTrace of StorageFileWatch and then set storage in to backup mode and then put it in to read/write mode
give it a few mins then try uploading the saveset.
normally it means that theres other items that those savesets rely on that need to be backed up first before it can be considered secured
For instance if i send you an email with 2 x 1mb attachments, in EV2007 and below it would be a single DVS file that contains the email, the attachments and the parts being added to the index
now it would be
12345abcde.dvs <--- email
abcd12345.dvssp <---- attachment 1
abcd12345.dvscc <----- converted content for the first attachment
efghi6789.dvssp <----- attachment 2
efghi6789.dvscc <----- converted content for the second attachment
And with the new OSIS and sharing between multiple vault stores, those parts need to be backed up in all vault stores and partitions, so you need 5 pieces to be backed u to secure one email item
the odd thing is that every now and then a few items are cleared out, right now I am setting up a package for the symantec support with the customer. I'll also post the DTRACE here.
thanks
----------------------------------------------------------------------------- Please grant me the ability to punch people in the face over standard TCP/IP
it was quite a while ago but it worked out like gertjan said.
it took some attempts to clear everything out --> create trigger file many times after we changed it to "use trigger file"
Thanks all for the help
----------------------------------------------------------------------------- Please grant me the ability to punch people in the face over standard TCP/IP
Would you like to reply?
Login or Register to post your comment.