How to limit proxy streams on Avamar 7.1

In Avamar 7.1, each proxy is now only registered once but can do eight streams.  This is great for management like access to datastores and groups as you no longer have to modify eight registered proxies.  If you want to limit the amount of streams in version 7, all you had to do was not give some registered proxies access to the datastores and that was it. This was useful in smaller environment where you want the redundancy of having two VM proxies but limit them to say four streams per proxy.

In this new release this has to be done by modifying the mcserver.xml file:

Just change the value from the default "8" to the amount of streams you want, in the case above, "4." Remember to restart the MCS service in order for this change to take place.

*Keep in mind that this is a global change and will affect ALL proxies.

Avamar GLR and MAPI

During GLR, if you encounter the error "backup browse result is empty" and you created the AvamarBackupUser manually because the creation tool did not work, you might have run into the problem of the MAPI for the AvamarBackupUser not having been updated.

To fix this follow these steps:

1. On the exchange server where the AvamarBackupUser mailbox is located, open a powershell and type the following command: get-mailboxdatabase "mailbox or username" | fl   (in this case the user name is AvamarBackupUser)
2. Find and copy the legacy DN name
3. Create a file called 'avmapi.cmd' in the var folder of the Avamar client install directory (usually c:\program files\avs\var)
4. Paste this in that file '--legacy_exch_dn="LEGACY DN NAME"'  without the single quotes
5. Save file

Now if MAPI was the issue the GLR should work without any errors.

More about the systems I was using:

• Avamar 7.0.1-61
• Exchange 2010 SP3 on Server 2008 R2
• No DAG

Disabling ILB quiescing from Avamar

If for some reason you have issues with image level backup and need to stop the quescing of VMs this is how you do it from Avamar.

Go to the dataset for the VM image backup and edit it, then go to the 'options' tab.

Select 'Windows VMware Image' from the drop down menu.

At the bottom click on the 'More' button.

In attribute enter: [avvcbimage]quiesce_fs

In Attribute Value enter: false

Click the + button to add 

End should look like this:

Repeat this for ‘Linux VMware Image’ plug-in if needed.

EMC Avamar replication not working after renaming grid

So i was about to finish a job with the last item to do was start replication between PROD and DR.  Problem is that the replication job gets stuck in waiting indefinitely (see end of blog for job log).  After a support chat with EMC this is a known issue in version 7 of Avamar (though they might fix it in 7.0.2).  After renaming a grid, the Avamar Client Agent (or avagent) still looks for the original name of the grid for replication.  Below you will see the steps for getting around this until a fix from EMC comes out.

• Run the following line, you should see the old grid name and the new grid name:

root@host:~/#: avmgr getl --path=/MC_SYSTEM
1  Request succeeded
1  OLDGRIDNAME.domain        location: hex number      pswd: hex number
1  NEWGRIDNAME.domain       location: hex number      pswd: hex number

• Next create avagent.cmd in /usr/local/avamar/var/

root@host:~/#: vi /usr/local/avamar/var/avagent.cmd

• Enter the following two lines and save:

--hostname=OLDGRIDNAME.domain
--mcsaddr=CURRENTIPADDRESS (not old one)

• Restart the Avamar Client Agent:

root@host:~/#: service avagent restart

• You should see the following

avagent Info: Stopping Avamar Client Agent (avagent)...

avagent Info: Client Agent stopped.

avagent Info <5008>: Logging to /usr/local/avamar/var/avagent.log

avagent Info <5174>: - Reading /usr/local/avamar/var/avagent.cmd

avagent Info <5417>: daemonized as process id 52585

avagent Info: Client Agent started.

• Once the service has restarted, go into Avamar Administrator, Policies and select the domain your grid is in.  You should see your new grid name and old one.  Make sure neither of them are disabled and replication will start working.

• For those interested in seeing what is in the log for the stuck job:
  

  <logdir ack="no" errdesc="Message intended for Client &apos;04012b6ed263497e84e7b3f0a96300d6fd12989f&apos; received incorrectly at agent on NEWGRIDNAME:1.1.1.1(e433961fb4975d1aa2ce635993242f2"

  21083161d) currently registered to &apos;newprod&apos;" pidnum="1008" targetCid="" browsetimeout="10" maxelems="-1" file="MOD-1352394809289" supportbundle="false" sync="fg" cmd="getworklog" cid="04012b600d6fd129d26349789fee84e7b3f0a963" errno="6207" showd