cancel
Showing results for 
Search instead for 
Did you mean: 

Netbackup vmware policy best practice

shahriar_sadm
Level 6
Hi dear all, my vmware backup environment includes a vm backup host and about 20 esx host with about 700 virtual machines. Is it necessary to create one policy for each vm? If I use intelligent policy example for each host, maybe backup for 40 virtual machine will run together. Netbackup recommend 4 concurrent vm backup.
Generally whats the best practice and your suggestion for manage this number of vm and host.
Thanks
2 ACCEPTED SOLUTIONS

Accepted Solutions

Michal_Mikulik1
Moderator
Moderator
Partner    VIP    Accredited Certified

Hello,

NetBackup recommends 4 concurrent backups - but per one datastore. This is set by Host Properties\Master\Resource Limit tab and supersedes other limits if they are higher.

So for example, when in your case 40 VMs on 3 different datastores are picked up by an intelligent policy, there will be maximum of 12 VM backups running concurrently, if not restricted more by limits on other levels (storage unit, policy, etc.).

Regards

 

View solution in original post

Hello,

following a couple of links useful for me when I had to tuning vmware backups.
IA B29: Best Practices for Backup and Recovery of VMware
NetBackup 7.6 Blueprints - VMware

Tuning depends also on transport you use and storage you have. 
You may create different VIP policies for different storage so you may limit jobs per policy.

Hope it helps,

regards

 

View solution in original post

8 REPLIES 8

Michal_Mikulik1
Moderator
Moderator
Partner    VIP    Accredited Certified

Hello,

NetBackup recommends 4 concurrent backups - but per one datastore. This is set by Host Properties\Master\Resource Limit tab and supersedes other limits if they are higher.

So for example, when in your case 40 VMs on 3 different datastores are picked up by an intelligent policy, there will be maximum of 12 VM backups running concurrently, if not restricted more by limits on other levels (storage unit, policy, etc.).

Regards

 

Hello,

following a couple of links useful for me when I had to tuning vmware backups.
IA B29: Best Practices for Backup and Recovery of VMware
NetBackup 7.6 Blueprints - VMware

Tuning depends also on transport you use and storage you have. 
You may create different VIP policies for different storage so you may limit jobs per policy.

Hope it helps,

regards

 

Thanks dears for reply,

I still have a problem, imagine I create one policy per each esxserver, for every day one esx server and vms backup. what happend if one vm move between esx servers before next backup. esx1backup sundays and esx2 backup wedensdays. if vm1 will backup in sundays on esx1 and before wedensday move to esx2, vm1 has duplicate backups and if imagine reverse mybe no backup in a week! 

Thanks 

Marianne
Moderator
Moderator
Partner    VIP    Accredited Certified

You should ensure sufficient backup infrastructure to backup ALL your ESX/VMs every day.

They don't need to be full every day - Fulls over a weekend and Incr on weekdays.

shahriar_sadm,

it depends on your environment, but generally it is not a good idea to have vmware policies fragmented per esx host.
Please check this: NetBackup for VMware best practices

If I understand correctly, you configured policies per ESX, so that when a vm moved from a esx to another it is backed up in another policy. A full is made when a vm is newly backuped up in a vmware policy.
In order to  avoid switching of vms among policies, it is better using queries in VIP and Setting global limits on the use of VMware resources

Hope it helps

 

Hi Andrea,

Corrently I have a policy for each vm! about 700 policies :| its very hard to tracking all changes in vmware environment like create and delete VMs...

I am confused, if I create a policy for each esx server also I have to use VIP. and if I create a policy for each datastore what happend on storage migration, you said create a policy for each esx server is not good idea, what is the better practice? using VIP quesry to select VMs on each datastore? or somting like specific tags? whats your suggestion?

Thanks

 

shahriar_sadm, 

my suggestion is to create some (~10 or 20.? it depends on your vmware configuration) vip policies based on vmware cluster (using field Cluster AnyOf "..." in query) or based on vmware datastore clusters. You may use Test Query... button in each policy to have a preview  It is not a problem to have dozens of Vms in same policy, if your storage is good. After policy configuration, please set some Resource Limits (in master server properties): a good start point could be 2-3 per esx, 2-3 per datatostore, 8-10 per DatastoreFolder, and limits jobs per policy to 6-8.
Remember that main rule is test and check, you have to check storage latency etc. Limits also depend on your storage and vmware configuration so check it before with some tests.

Hope it helps
 

Genericus
Moderator
Moderator
   VIP   

We have about 1000 VM hosts. We split them into groups and back them up in groups.

I have a DailyA.PROD & DailyB.PROD each with about 150 clients - running fulls every day (BOOST to a data domain)

I have some WeeklyA/B/C.PROD polices each with about 100 clients 

I have daily.DEV and Weekly.DEV policies, and Daily.QA and Weekly.QA, the same with UAT.

I use the limits mentioned earlier in the mastersettings to limit the number per host, along with storage limits.

 

We are looking into using VM tagging to define the groups and run policies that use the tags to choose hosts to backup.

 

 

 

NetBackup 9.1.0.1 on Solaris 11, writing to Data Domain 9800 7.7.4.0
duplicating via SLP to LTO5 & LTO8 in SL8500 via ACSLS