cancel
Showing results for 
Search instead for 
Did you mean: 

How can I clear the backup status if my vault partitions are on a SAN?

tfoster
Level 4

So I have a couple of "Critical" statuses the vault's throwing at me, and has for a long time now. We don't actively backup the vault data because it's all on a Compellant SAN and all my vault drives are LUN's. There's no reason to backup the vault data because that's a waste of space. If a drive fails, or multiple drives fail, you replace them and let them rebuild in the array.

 

How can I clear these flags that items haven't been backed up? Do I manually put a trigger file just to shut it up?

 

The warnings I'm referring to are,

"New items awaiting backup or replication"

"Check for success of backup and replication of newly-archived data"

11 REPLIES 11

A_J1
Level 6
Employee Accredited Certified

Hello tfoster,

 

From EV Version 8.0 onwards EV does not secure the items in SQL unless and untill you backup your partition.

You can use the following scripts and schedule them in Task schedular.

We use Triggerfile to let EV know that the backup has been done.

Your script should be as follows.

SET BackupMode

C:\Windows\SysWOW64\WindowsPowerShell\v1.0\PowerShell -psconsolefile "C:\Program Files (x86)\Enterprise Vault\EVShell.psc1" -command "& {Set-VaultStoreBackupMode -Name 'Name of the Vaultstore Group' -EVServerName evserver1 -EVObjectType VaultStoreGroup}"

C:\Windows\SysWOW64\WindowsPowerShell\v1.0\PowerShell -psconsolefile "C:\Program Files (x86)\Enterprise Vault\EVShell.psc1" -command "& {Set-IndexLocationBackupMode -EVServerName evserver1 -EVSiteName 'EV Lab'}"

del "C:\EVStorage\Express Vault Store\Express Vault Store Ptn1\ignorearchivebittrigger.old"

Note: You have to specify the above command for all of your partition

(The above statement in bold will deleted the "ignorearchivebittrigger.old" file from the partition)

CLEAR BackupMode

C:\Windows\SysWOW64\WindowsPowerShell\v1.0\PowerShell -psconsolefile "C:\Program Files (x86)\Enterprise Vault\EVShell.psc1" -command "& {Clear-VaultStoreBackupMode -Name 'Name of the Vaultstore Group' -EVServerName evserver1 -EVObjectType VaultStoreGroup}"

C:\Windows\SysWOW64\WindowsPowerShell\v1.0\PowerShell -psconsolefile "C:\Program Files (x86)\Enterprise Vault\EVShell.psc1" -command "& {Clear-IndexLocationBackupMode -EVServerName evserver1 -EVSiteName 'EV Lab'}"

ECHO > "E:\Enterprise Vault Stores\MailArchiving Ptn1\ignorearchivebittrigger.txt"

Note: You have to specify the above command for all of your partition

(The above statement in bold will create the "ignorearchivebittrigger.txt" file in the partition)

You can use the above PowerShell commands to create a batch file and schedule the Set-Backupmode before the Backup Schedule and Clear-Backupmode script after the backup is complete.

 

Also you can refer to the following articles:

How to use a trigger file for devices that do not support the Archive bit when using Enterprise Vault (EV)

http://www.symantec.com/docs/TECH35610

Generating PowerShell backup commands for your environment

http://www.symantec.com/docs/HOWTO37806

Download the Generating PowerShell Backup Commands guide below:

http://www.symantec.com/docs/TECH75694

 

I hope this helps !!!

 

plaudone
Level 5
Employee Accredited Certified

It is recommended that the data in the vault store locations, index locations, and SQL DB's be backed up in order to recover from a disaster situation.  If the vault store data is lost there is no method of recovering user data.  I have seen many times that drive data cannot be recovered from a failure and backups are the only method for recovering all of the user data.  

Best Practices for Enterprise Vault Backup- http://www.symantec.com/docs/TECH172279

The backup mode process from a database standpoint ensures that all items are marked as secured so that they can be purged from the temporary tables.  This is an indication that the data has been secured by some process outside of EV (Backup, Replication, Snapshot). 

 

 

tfoster
Level 4

Maybe I'm confused, but why are "backups" needed If the vault stores are on a SAN? If you operate your Vault on a VM that is on a SAN, and the vaultstore partitions are LUN's on that SAN, the SAN itself is basically it's own redundant solution, why should I need to back up the vaultstores?

 

We can do snapshots and have 2 weeks worth of replays, but the data is basically as secure as you can possibly imagine because it would take the entire SAN to die for the data to be lost.

 

I can understand backing up if you're using DAS(local drives), but everything is on a SAN..

plaudone
Level 5
Employee Accredited Certified

It would still be recommended to have the data backed up to recover from a disaster situation like the loss of a data center.  If all the data is in one location and there is an issue with that location due to hardware, natural disaster and such the data may not be recoverable if not backed up and stored in a remote location for recovery.  

A_J1
Level 6
Employee Accredited Certified

Hello tfoster,

 

As i mentioned earlier with Enterprise Vault Version 8.0 onwards the storage model has been designed in such a way that unless until you backup your vaultstore partition the records from watchfile table will not clear.

To further explain :

Once you archive an item any entry is made in several tables in VaultStore and VaultstoreGroup Databases(Also called as fingerprint databases). One of the table referred as Watchfile table in Vaultstore Databases has a column called "items secured" This column is updated once you have backed-up your item in the partition. 

And the updation of the column is done by the StorageFileWatch process which scan the vaultstore partition to confirm each item present in the watchfile table is secured or not. So depending upon the Backup method (ArchiveBit or Trigger File) set on the partition it will update the watch file table.

Now this StorageFileWatch process triggers only when storageservice is restarted or you set/clear enterprise vault from backup-mode.

Worst Case if you don't do you backups:

EV will consider that the backup is not done and the records from watchfile will keep on increasing which will log an event in the VAC under Status tab. If its keeps on increasing then it could cause performance issue with EV,Database fragmentation etc..

So if you are sure that the data present in the partition is secured. Then you don't have to per from backups but you have to run the scripts as suggested in my above comment and let EV consider that the backup is done and it needs to clear all the records from the watchfile table.

 

I hope this helps !!!

 

 

A_J1
Level 6
Employee Accredited Certified

Hello tfoster,

 

Let me know if you have any further queries.

A_J1
Level 6
Employee Accredited Certified

Hello tfoster,

 

do you have any further queries ???

WiTSend
Level 6
Partner

Having been through a SAN failure, I can tell you that having a periodic back-up or SAN replication to a secondary storage platform is highly desirable. 

tfoster
Level 4

Hi, yes we have SAN replication between 2 data centers over a Gig-E link. So even if one SAN died, we'd failover everything to the other data center.

 

A_J, I'll follow the steps to manually run the scripts. Thanks!

FreKac2
Level 6
Partner Accredited Certified

Just to add one thing on the disaster scenario.

That is that replication will not protect you from logical errors.

E.g. file system corruption or other data integrity errors, those will be replicated as well to the second site.

 

As with everything you have to weigh the risk and costs of any solution.

If the data is critical for the company then doing backups would be at the top of my list to cover.

The important thing is to make a choise based on the severity of a data loss, not really the likelyhood of it occuring.

Just because something is rare doesn't mean it don't happen and Murphy is always there to prove someone wrong.

So if it's really important, back it up, if not YMMV.

 

tfoster
Level 4

I agree.. it's not like we don't do backups.. we have 2 weeks worth of Compellent Replays and I would be able to restore to any of those..