Ask the Expert: Avamar Plugin for Microsoft Applications

Hello Everyone,

Welcome to Ask the Expert conversation. My name is Pankaj, i'll be your host for this event.

This is an opportunity for you to ask all sorts of Avamar Specific Windows Module questions(Includes File System, SharePoint, Exchange, MS SQL, Hyper-V and BMR)

Just a note : Remember, no question is a silly question

Besides, the session, I'll try to post a daily backup tips and latest KB's/Fixes for our Avamar Users.

Welcome everyone, this ATE discussion has began. Feel free to post your questions and our SMEs will be sure to answer them before the end of our event. Let's keep this discussion respectful and informative. Thanks!

Today's KB : Exchange VSS Backup is failing with error: avtar Error <7042>: Failed to create snapview, as one of the elements required for path "XXXXXXXXX.edb" not found

Today, let's talk about some our new "Demand Paging cache and Growing Size of the caches".

For the file cache, each entry is 64 bytes, or 6.4 MB per page and each champion is 20 bytes or 200 KB per page.

On disk, the champions are stored with the entries that give a size of about 6.7 MB for each page of the file cache. There is some overhead, but it is a very small percentage of the total size, and usually not relevant.

In RAM, the champions are stored separately. The page sizes are about 6.4 MB for the file cache. The champions are brought into RAM in a single array that contains all of the champions for all of the applicable backups in the cache.

As an example, the file cache required for 40 M files is approximately 616 pages: 40 M / 65,000 (estimated) or 4.13 GB on disk (616 * 6.7 MB).

These values are representative for very large backups. Auto-tuning alters the number of pages stored in memory based on the size of previous backups and other factors, including the actual memory available to the process. The values may be considerably smaller or larger, based on previous backups requirements.

Hope that helps in understanding why your paging cache might juice up a little more of your disk space, but remember, it is more effective as compared to monolithic cache file.

Hi Pankaj,

Could you please explain about Disaster Recovery of exchange backups ? What should be taken care to ensure DR of exchange server ?

Regards,

Pawan

Hi Pawan,

I'll answer it in two parts :

1) Preparing for Disaster Recovery from EMC Avamar perspective

2) Performing the Disaster Recovery (Microsoft and EMC Avamar)

I. Preparing for Disaster Recovery

1. Ensure that you have onsite and offsite copies of the installation disks for the operating system, Exchange Server, and any software updates.

2. Perform full Windows server backups by using the Avamar Client for Windows. Include System State and all critical and noncritical disks in the backups. The EMC Avamar for Windows Server User Guide provides full backup instructions.

3. Back up all Exchange databases by using the Avamar Plug-in for Exchange VSS.

II. Performing Disaster Recovery

Microsoft :

For Exchange DR, it is important to note that almost all of the configuration settings for Mailbox and Client Access servers are stored in Active Directory.

Exchange includes a Setup parameter for recovering lost servers. This parameter, /m:RecoverServer, is used to rebuild and re-create a lost server by using the settings and configuration information stored in Active Directory.

However, be aware that there are several settings which are not restored, such as changes to local web.config and other configuration files. In addition, custom registry entries are not restored.

We recommend that you use a reliable change management process to track and recreate these changes.

For detailed steps about how to perform a server recovery of a lost Exchange 2013 server, see Recover an Exchange Server. For detailed steps about how to recover a lost server that's a member of a database availability group (DAG), see Recover a database availability group member server.

EMC Avamar

Procedure

1. Restore the server from the Windows client backup as described in the EMC Avamar for Windows Server User Guide.

2. Complete the recommended Microsoft Exchange Server disaster recovery procedure by using the steps in one of the following articles on the Microsoft TechNet website:

- Exchange Server 2013 and 2010 -- "Recover an Exchange Server" at

http://technet.microsoft.com/en-us/library/dd876880.aspx

Exchange Server 2007 -- "Disaster Recovery" at

http://technet.microsoft.com/enus/library/aa998848(EXCHG.80).aspx

3. Use the Exchange VSS plug-in to restore the Exchange databases.

4. Restore the Exchange databases by using one of the following methods, depending on the environment:

     - Restore the databases from the Exchange VSS plug-in backups.

     - Bring the Exchange server back online in the DAG or cluster, and allow the mailbox databases to synchronize from the current versions on the other servers in the DAG or cluster.

     - Use a combination of both methods. Reseed the mailbox databases with the Exchange VSS plug-in backups, and then synchronize the newly-restored Exchange server with the current version on the other servers in the DAG or cluster.

5. When you are sure that the restore of the environment is correct and complete, perform a full backup of the newly restored environment.

Since, we are talking about DR's, let's consider another situation : You have deleted a Database and have re-created the database in Exchange but due to some reasons, you have a requirement to restore the database that was backed up before the re-creation happened.

In that situation, the restore might fail with the below error message :

avexvss Error <13015>: VSS writer did not find anything to restore

avexvss Info <12994>: Restore completed successfully.

This is a known issue and the error happens because of the fact that the BCD component file that was backed up before the database got rebuilt has the old database Guid information and since, usually guid information changes with the database creation, Plugin is not able to find the target database in records.

In order to resolve the issue, there are Hotfixes available for various builds or you can simply upgrade to Avamar 7.1 since, the fix has been checked into the release.

If required, please contact EMC Support for the Hotfix.

Pankaj Pande,

Awesome event, thank you! Can you give us a quick run down as if we were reading a cheat sheet about the clients here:

Windows SQL makes perfect sense to me and is pretty straight forward (in green). It's the others boxed in blue and orange, that I think would benefit from a high level explanation of what it does and where you would use it. [I know this is an outdated screenshot, but was in a time crunch and was all I had on hand ]

Appreciate it,

-Ed

Hello Ed,

Apologies for the delay, but I must say, its a great query.

Please find the information about the plugins.

Thanks,
Pankaj Pande

Pankaj,

This is perfect!

Thank you,

-Ed

Hi Pankaj,

Would you be willing to discuss the --parallel and  --parallel-type features of the Avamar Windows client for file system backups? How customers can back up multiple drive letters simultaneously and the criteria for doing so?

-- Joe

Hi Joe,

During a backup an avtar stream is started per volume or target.  (Please note that there will be no multiple avtar processes showing during the backup since, Avtar is a 64-bit Multi-Threaded process) 

The sub-processes are traversing the volumes/targets in parallel which increases data and number of files we can read at the same time from the disk.  Multiple streams will use the same cache loaded into RAM. It has also been verified that paging cache can be used with parallel flag. 

Parallel streams work on a single connection to the server, thus only one server session is being utilized.

NOTE: Use of parallel flag is not suitable for all of the clients. It is recommended that customers work with support to determine if the client is will benefit from using this flag.  Because only eligible servers will use this flag, it is recommended to turn parallel processing in avtar.cmd on a client unless the client has its own dataset in GUI

To turn on parallelism the following flag needs to be set: (By default this flag is set to false)

--parallel=true

In order to enable multiple streams, please set the number of maximum streams n for the client.  By default the value is set to 1, which means no more than one stream will be used. 

--max-streams=n

There are two parallel types that can be used:

By target - creates one stream per each target as specified in a dataset (It is recommend for Unix and Linux clients).  In order to use this type, it is recommended that the targets in a dataset are defined so that the client will benefit from the parallelism setting.

By volume - creates one stream per each backed up volume, to max-streams value, regardless of dataset. This option is recommended for Windows systems.

To set the parallel type use the following:

--parallel-type=by-target

or

--parallel-type=by-volume

If not set, the --parallel-type defaults to by-target.

For more information, please follow the KB : support.emc.com/kb/184634

Hope this answers your query.

Thanks,
Pankaj Pande

This Ask the Expert event has concluded. Many thanks to those who participating with their questions, but especial thanks to our SMEs who loan their time to answer questions.