04-13-2015 10:51 PM
Hi
Environment:
EV 10.0.1.1171
exchange 2010
Archive location centera (no collection enabled)
Problem:
Usage report shows items waiting for backup.
SQL query returns rows equal amount waiting to be backed up.
use vault_db
select * from JournalArchive where BackupComplete = '0'
Checked centera both queues empty - nothing waiting to be saved or replicated
dtraced StorageFileWatch - what should I be looking for it?
What else can I check in order to see if these items can be cleared from the watchfile...? And how to do it?
Not sure if correct but the sql query above also says value RecordCreationDate = from 2014 dates... Just got the environment under my management so don't know if it's been like this for a long time... looks that way to me though...
Sani B.
Solved! Go to Solution.
09-21-2015 03:31 AM
Hello Sani.
I don't recommend deleting the items from the watchfile.
Is it possible for you to verify the archive bit of items that are backed up? Is that 'cleared', or do the items still have the atrribut' 'A' set?
If A(rchive) is still set, you can basically do two things:
Set vaultstore in backup mode. open dos-prompt, do a 'attrib -a *.* /s' (clears archive attribute throughout all subfolders for all items). When done, clear backup mode.
Or, change partiton backup tab to 'use triggerfile'. Set backupmode. create the 'IgnoreArchiveBitTrigger.txt' file in the root of the partition. File can be empty. Clear backupmode. You should see the txt file change to .old extention.
For both, you should see the storagefilewatch process grabbing cpu-time. this means it is processing the items.
If the above does not work, or if the items have the archivebit cleared alright, you might want to work with support on it.
04-13-2015 11:19 PM
Actually correction here - they used to store archives to NTFS volumes so there is a chance that the stuck items if the creation date is correct are from those days and has nothing to do with centera... How do I verify that? The NTFS volumes have not been migrated to centera but left intact and messages can still be retrieved from those locations.
Sani B.
04-14-2015 04:01 AM
you can check the properties of the vault store partitions that are NTFS from the EV console to see how many "unsecured items" you have and if they're set to detect backups based on archive bit or trigger file and take action accordingly.
09-21-2015 03:19 AM
Finally can return to this problem - so the choise with the NTFS backups has been
- Use the archive attribute
What is the action I should take with this knowledge...?
I checked that the NTFS location get's backed up every night so the items are backed up - how do I move along the watch file... or is it okay to just delete these items from the watch file?
Sani B.
09-21-2015 03:31 AM
Hello Sani.
I don't recommend deleting the items from the watchfile.
Is it possible for you to verify the archive bit of items that are backed up? Is that 'cleared', or do the items still have the atrribut' 'A' set?
If A(rchive) is still set, you can basically do two things:
Set vaultstore in backup mode. open dos-prompt, do a 'attrib -a *.* /s' (clears archive attribute throughout all subfolders for all items). When done, clear backup mode.
Or, change partiton backup tab to 'use triggerfile'. Set backupmode. create the 'IgnoreArchiveBitTrigger.txt' file in the root of the partition. File can be empty. Clear backupmode. You should see the txt file change to .old extention.
For both, you should see the storagefilewatch process grabbing cpu-time. this means it is processing the items.
If the above does not work, or if the items have the archivebit cleared alright, you might want to work with support on it.
09-21-2015 05:07 AM
Hi GertjanA!
Thanks for the reply!
I tried the later suggestion on one partition that has few of the items in the watch file - didn't do anything so I'll open a support case for this as you suggested.
Sani B.
09-21-2015 11:34 AM
maybe let us know what support does to resolve the issue? if it's different than the solution you marked it would help to know in case someone else has the same issue and comes looking here for help.