Limitations for NetBackup Master in a VMware 5.x Guest

[CKWallace]

We have several customers with NetBackup Master Servers in a VMware guest.  Most of these environments are less than 20 FETB and have catalogs less than 100GB.  We have a site at 40FETB with a catalog of over 300GB.  I fully understand limitation for connection of tape drives and the like, but am wondering of scale is an issue for bigger sites.  I should be able to get "whatever is required" for CPU and RAM.  (NetBackup 7.5, VMware 5.x, Cisco UCS Chassis, Windows 2008 R2 and NetApp SAN)

Thanks in advance for your input. - Clark Wallace, LOGICALIS

[RamNagalla]

 

Support for NetBackup 7.x in virtual environments 

http://www.symantec.com/business/support/index?page=content&id=TECH127089

whitepaper-netbackup-architecture-overview

https://www-secure.symantec.com/connect/articles/whitepaper-netbackup-architecture-overview

may be above does would help you ..

[CKWallace]

Nagalla - thanks for the response.  I access to all of these documents.  I was hoping for some real life feedback on when does the NetBackup Master server get too big to run on a VM.

Thanks - Clark

[Marianne · Accepted solution]

Maybe you can use these docs as a guideline:

Designing your backup system
http://www.symantec.com/docs/HOWTO56078

How to estimate the number of master servers needed
http://www.symantec.com/docs/HOWTO56073

Extract:

When making an estimate, consider the following:

• Type of processor

  For a NetBackup master server, Symantec recommends using multiple discrete processors instead of a single multi-core processor. The individual cores in a multi-core processor lack the resources to support some of the CPU-intensive processes that run on a master server. Two physical dual-core processors (for a total of four processors) are better than a single quad-core processor.

http://www.symantec.com/business/support/index?page=content&id=HOWTO56073#id-SF0S0175814 

As you can see, a master needs multiple physical cpu's. 
A sign of a master server that is battling to cope is normally seen during starting phase of jobs - i.e. the time it takes to allocate resources. Delays of 5 - 15 min (or longer) between identifying and allocating resources can be considered as 'taking strain'.

Hope this helps somewhat...

[CKWallace]

Thanks for the input.  We actually decided to stay physical, primarilly so we have the NetBackup master self contained and not imbedded in the infrastructure it is protecting.

-Clark Wallace