Monitoring backup question

[Kenneth_Hansen]

Hi,

I have a EV environment  8.0 SP3. Backing up EV using HP Data protector and SQL agent, al EV DB's a selected.
Enterprise Vault logs keep reminding with a warning that X number of save sets is not backed up.
Data Protector report all backup ok. 

Anyone seen this?

[GertjanA]

Is the backup program changing the archivebit?

if not, use the triggerfile. This is described in the manual (administering I believe)

The process I use is the following:
Set ev-server in backupmode, using the powershell command. This is started by the dataprotector job (use an account to run the service under, and grant the account at least " storage admin" rights in EV Authorization Manager.
backup EV (indexes, Stores, msmq etc)
run a postbackup command (by the backupjob) which writes the IgnoreArchiveBitTrigger.txt file in the store-locations.
Then in same command clear backup mode.

That will kick off the Storage File Watch process that does a lot, amongst others changing pending archivved icon to archived.

Gertjan

[JesusWept3]

Kenneth,
is the number the same or does it grow each day?
It could be that its just certain files that aren't being backed up for one reason or another.

What you can do is if you go to the Vault Store database that its reporting against, then open the Watchfile table.
This has a list of all the items that it is awaiting backup from, what you can then do is some random samples, check on disk that the items have had the archivebit removed  and that they are on backup as well.

You can use the IgnoreArchiveBitTrigger.txt as well like gertjan says, just remember to do this *after* the backup has complete and not before, also the ignorearchivebittrigger.txt file needs to be created, not copied, otherwise this will start to cause issues in the future. and lastly remember on the partition properties to set up the backup details.

to create an ignorearchivebittrigger.txt file, easiest way to do it is add a line like this to your backup batch file

type nul > E:\Enterprise Vault Stores\YourVaultStore Ptn1\IgnoreArchiveBitTrigger.txt


but honestly, you may find that its just the type of backup you're running isn't clearing the Archive bits, so for instance in BE and NBU both do FULL backups, these remove the archive bit, and i *think* incremental also removes the archive bit, but a differential does not (i'm a bit fuzzy on which does and does not, so that might be reversed)

you may find that when you next do a full backup, that number will go down, or at least the contents of the watchfile table will be different

[Kenneth_Hansen]

When backing up EV stores, we do use power shell, and this seems to work just fine. Short cuts a being made and so on. 
I have to verify with the back adm, if backup is set to clear archive bit or not. 
I might be very wrong here, but I thought that the warning message in event log was referring to SQL backup not beeing ok??

[MichelZ]

Kenneth

The message you are seeing is not SQL backup, it is actual SaveSet Backup (which is measured by either Archive Bit reset and/or trigger file).
There IS a sql eventlog backup message, but it more reads like "your sql database has not been backed up for x days"

Regards
Michel

[JesusWept3 · Accepted solution]

Yup, so in my environment  because i'm a lazy admin i get the following

Event ID 41008 = Savesets not backed up
Event ID 41011 = Vault Store Database not backed up
Event ID 41012 = Vault Store Database Transaction Logs not backed up
Event ID 41203 = Fingerprint Database not backed up etc


Event Type: Warning
Event Source: Enterprise Vault
Event Category: Monitoring
Event ID: 41008
Date:  1/20/2010
Time:  10:25:15 AM
User:  N/A
Computer: AT-081L18J9GNNX
Description:
There are 8337 savesets that have not yet been backed up or replicated for Vault Store 'EVUsers'.
Review your procedures and make any changes needed to ensure that backups happen in a timely fashion.
If you are using an EMC Centera, check the size of the Centera replication queue.


Event Type: Warning
Event Source: Enterprise Vault
Event Category: Monitoring
Event ID: 41011
Date:  1/20/2010
Time:  10:13:39 AM
User:  N/A
Computer: AT-081L18J9GNNX
Description:
The SQL database for Vault Store 'EVUsers' has not been backed up for 43 days. The information in this database is at risk until the database has been backed up.
Review your SQL database backup procedures and make any changes needed to ensure that backups happen in a timely fashion.

Event Type: Warning
Event Source: Enterprise Vault
Event Category: Monitoring
Event ID: 41012
Date:  1/20/2010
Time:  10:09:49 AM
User:  N/A
Computer: AT-081L18J9GNNX
Description:
The SQL database transaction log for Vault Store 'EVUsers' has not been backed up for 43 days. The information in this database is at risk until the database has been backed up.
Review your SQL database backup procedures and make any changes needed to ensure that backups happen in a timely fashion.

Event Type: Warning
Event Source: Enterprise Vault
Event Category: Monitoring
Event ID: 41203
Date:  1/20/2010
Time:  10:06:17 AM
User:  N/A
Computer: AT-081L18J9GNNX
Description:
The SQL database for Vault Store Group 'EVVSGDefaultUpgradeGroup_1_1' has not been backed up for 43 days. The information in this database is at risk until the database has been backed up.
Review your SQL database backup procedures and make any changes needed to ensure that backups happen in a timely fashion.



In your case, you either are backing up most DVS files correctly as they are turning in to shortcuts
OR you have it set to Immediately After Archive, and in EV2007, EV never kept a record of files waiting to be backed up before, so you never saw the message. You would only see that in an After Backup scenario

However in EV8 that changed, regardless of After Backup or Immediately After Archive, EV will always keep a list of items to be backed up in the WatchFile table, unless you set IgnoreBackup to True or use an IgnoreArchiveBitTrigger.txt file