cancel
Showing results for 
Search instead for 
Did you mean: 

EV 10.0 with NetApp

We have a EV Cluster pointing to a SQL Cluster where all the DB's reside. I read alot about Symantec supporting Netapp in EV10.0.

My question is, I do a SQL backup to the SAN. What other steps are required to do a proper backup with Netapp?

Our Filer=3140

NetApp release=8.0.2

Thanks,

Isaac

1 Solution

Accepted Solutions
Accepted Solution!

OK we need to take a step

OK we need to take a step back about this
Are you talking about your Vault Store Partition data (DVS* files) or are you talking about the SQL Database files (LDF/MDF files)?

Because the two are so completely different.
The Vault Store Partitions store the CAB, DVS, DVSSP, DVSCC and ARCH files that make up the actual vault data, its the form that Emails and files get stored as after being archived by Enterprise Vault with different file extensions based on Single Instance Storage and items that need to be used for the index.

The SQL Databases though are for the most part MDF (the actual database file) and the LDF (the transaction log files).

With Enterprise Vault Store Partitions, files are archived and typically have the archive bit removed by the backup software to indicate the item has been succesfully backed up, after the storage service comes out of backup mode, EV will go through a list of items it has archived and check to see whether the Archive Bit has been removed, if it has then it will show that the item has been backed up (Secured) and change a Pending Item in a users mailbox to a fully archived item.

If it is a CIFS share that does not support the archive bit, then you need the backup software to use a file named IgnoreArchiveBitTrigger.txt, when the storage service restarts it will go through and *assume* that all items that have been archived before the create date of this text file has been backed up.

Once it has gone through the contents of the WatchFile table in the vault store database, it will then rename the .txt file to .old, as an indicator that it has finished processing that partition.

However the IgnoreArchiveBitTrigger.txt can only be used for Enterprise Vault Store Partition data, it cannot be used for SQL databases, transaction logs or Enterprise Vault Index files.

For SQL you will have to work with your backup software to have it configured to back the database up, and truncate the transaction logs etc, which can also be done via maintance plans.

Your best bet is to talk to your SQL Database Admins and your backup admins to come up with a proper route forward to determine how best to backup your sql databases.

https://www.linkedin.com/in/alex-allen-turl-07370146

View solution in original post

13 Replies

Well when Symantec talks

Well when Symantec talks about supporting different devices such as EMC Devices like Celerras, Centeras etc and NetApps , they're talking about storing the Vault Store Partitions and the DVS* files

For the SQL Server databases and such and their supported platforms of where the MDF/LDF files is in the realms of Microsoft to decide.

Some more information about NetApp and SQL Server support can be found here:
http://www.netapp.com/us/solutions/applications/microsoft-sql/ms-sql-business-backup-recovery.html

https://www.linkedin.com/in/alex-allen-turl-07370146

A few questions

Are you using NetApp as a backup destination only or are you using it for Vault Store data (DVS data)?

Are you using CIFS shares? 

Are you using Snaplock?

Assuming you are using CIFS for vault store data, and it isn't a Snaplock, you can backup the Netapp filer like any network share it one major difference...  Netapp filers don't use the archive bit.  So many times I see people backup CIFS shares as a snapshot, since you can't really do Differentials or Incrementals.

Since it doesn't use the archive bit, you'll need to use the IgnoreArchiveBitTrigger in order to inform EV that the data has been backed up.

I hope this helps.

Yes, this is what I was

Yes, this is what I was looking for.

 

So, if I backup my SQL on a CIFS share, all I need to do is change the bitlock? Is there any configurations on the NetApp that I need to implement? Or all the Netapp does it backup the data?

 

Is this the correct procedure to set the lockbit ?http://www.symantec.com/business/support/index?page=content&pmv=print&impressions=&viewlocale=&id=TECH35610

If you are using a Netapp for

If you are using a Netapp for a backup destination for your SQL backups, that really is outside the realm of EV, as JW2 referenced.

If you are using the Netapp for the Vault Store file data (not SQL data), then yes, that article should help.

The article talks about

The article talks about creating files etc. If my netapp takes a snapshot of the EV Store LUN location, how does EV know when to rename the file to .old?? Is there any documentation that explains how to configure EV or Netapp??

Accepted Solution!

OK we need to take a step

OK we need to take a step back about this
Are you talking about your Vault Store Partition data (DVS* files) or are you talking about the SQL Database files (LDF/MDF files)?

Because the two are so completely different.
The Vault Store Partitions store the CAB, DVS, DVSSP, DVSCC and ARCH files that make up the actual vault data, its the form that Emails and files get stored as after being archived by Enterprise Vault with different file extensions based on Single Instance Storage and items that need to be used for the index.

The SQL Databases though are for the most part MDF (the actual database file) and the LDF (the transaction log files).

With Enterprise Vault Store Partitions, files are archived and typically have the archive bit removed by the backup software to indicate the item has been succesfully backed up, after the storage service comes out of backup mode, EV will go through a list of items it has archived and check to see whether the Archive Bit has been removed, if it has then it will show that the item has been backed up (Secured) and change a Pending Item in a users mailbox to a fully archived item.

If it is a CIFS share that does not support the archive bit, then you need the backup software to use a file named IgnoreArchiveBitTrigger.txt, when the storage service restarts it will go through and *assume* that all items that have been archived before the create date of this text file has been backed up.

Once it has gone through the contents of the WatchFile table in the vault store database, it will then rename the .txt file to .old, as an indicator that it has finished processing that partition.

However the IgnoreArchiveBitTrigger.txt can only be used for Enterprise Vault Store Partition data, it cannot be used for SQL databases, transaction logs or Enterprise Vault Index files.

For SQL you will have to work with your backup software to have it configured to back the database up, and truncate the transaction logs etc, which can also be done via maintance plans.

Your best bet is to talk to your SQL Database Admins and your backup admins to come up with a proper route forward to determine how best to backup your sql databases.

https://www.linkedin.com/in/alex-allen-turl-07370146

View solution in original post

The files i am referring to

The files i am referring to is the CIFS share for EV partition, NOT SQL.

ok sorry, what confused me is

ok sorry, what confused me is in the original post you were talking about doing SQL backups to the SAN
Other than that, just refer to mine and chris' comments regarding backups from a NetApp/CIFS device

https://www.linkedin.com/in/alex-allen-turl-07370146

Here are my thoughts on this

Here are my thoughts on this one (been running EV on NetApp for 4+ years now).

I wouldn't recommend putting any EV data on CIFS. Your performance is going to be terrible. EV 10 best practices do not recommend this setup for indexing data. Vault Store data can tolerate CIFS, but your storage aggregate and raid groups better have a lot of decent disks.  You really need to setup a iSCSI target, at worst, for your indexing data. I would review the following guides which explain some of the storage requirements/best practices for EV 10:

EV 10 Performance Guide

EV 10 SQL Best Practices

EV 10 Indexing Best Practices

 

For backing up the EV Vault Stores on NetApp storage, you will need to use a trigger file (remember to also set the Vault Store Partitions to scan for a trigger file) . Here is a download I created a long time ago that will show you how to do this properly for NetApp backups. This was created for EV 8, but still applies to EV 9 and 10. Also...if you have licesning for NetApp SQL SnapManager, make sure to use that for your SQL backups, if your SQL data is stored on the NetApp (and be sure to snapshot the other EV data (vault stores/indexing) at the same time as SQL backups so you have consistent data in the event you need to restore your data.

So if i create a parition,

So if i create a parition, the Storage type should be NetApp Device if it is a NetApp lun or a NTFS Volume? The data is using ISCSI, not CIFS.

just an FYI, he is referring

just an FYI, he is referring to CIFS which is a FileSystem as opposed to iSCSI which is a connection type

https://www.linkedin.com/in/alex-allen-turl-07370146

Yep, I am using ISCSI. Now my

Yep, I am using ISCSI. Now my only question is what type of storage is this defined as when one creates a partition, a NetApp Device or NTFS Volume?

Choose NetApp. What this does

Choose NetApp. I'm pretty sure this is to tell EV to use the correct block size when using NetApp storage (as well as give the other options to tell EV if your storage is using Dedupe, Compression, or SnapLock).