One of our customer having some issue with their Netbackup master recently. Their environment have two Netbackup Master servers which is on Veritas Cluster Server. Our customer facing an error when the Master backup is failover to passive node. It seem that the Netbackup services is not brought up automatically even when the failover were successfull. We have to issue command "bpup -v -f" to start all of the services manually.Please advice. Thank you

Hi Marianne,Thank you for the advice. We have log a case to Veritas for this issue under the VCS (SFHWA) product. It seems like the VCS helper service need to be re-enter the password on both nodes and everything is fine now. We manage to perform manual failove and perform backup on each node. Regards

You'll probably find that there is a single service that might be failing, or that it might be taking longer to startup than VCS expects. Set the NetBackup Server resource to non-critical. The restart or failover the service group and then monitor the Event viewer application/system events to see if there is a complaint about any service that doesn't start. If there are none, its probably just a timeout issue and you can increase the online timeout of the NetBackup Resource. You can also control which resource VCS monitors if you look in the registry but try the above first.

There seems to be no resource for NetBackup itself. Resouce dependencies work like this (I will try my best to sketch a text version of a service group here): NetBackup (NBU services) / \ / \ Lanman Mount (virtual name) (Drive letter) / \ IP DiskGroup / NIC Resources are onlined from the bottom up. After Lanman and Mount resources have onlined successfully, the NetBackup resource must be onlined. The NetBackup resource seems to be missing from the cluster config. I see these resources going offline and online (you can map them to my picture above): NIC is a persistent resource (always online) - no need to online or offline.NetBackup_Server-IPNetBackup_Server-LanmanNetBackup_Server-MountV-HNetBackup_Server-VMDg-HYou can see that there is no NetBackup resource. Something like NetBackup_server-NBU .This command output will confirm configured resources: hastatus -sum Who performed the NetBackup Clustered installation and configuration? Was it not checked and tested before handing over to the customer? I found this TN that could explain the absence of a NetBackup resource: https://www.veritas.com/support/en_US/article.000014833 (The TN is quite old, but worth checking with with Veritas Support to see if TN is still relevant.) In all honesty, if this is the case, then someone has not done his/her job during initial testing.

Are you referring to automatic failover due to a fault of a critical resource?Or a switch to the other node (manually initiated)? Can you confirm that the NetBackup resource was online at the time of failover?Was the NetBackup resource perhaps in faulted state? If so, did you manually clear the fault? Faults on non-persistent resources are not automatically cleared.The reason for that is that 'something' happened for the resource to fault. You need to troubleshoot the reason and fix it, then clear the fault. A manual switch to another node will online a service group to the same state as it was on the source node. Are your backup resources using VCS commands or VCS GUI to start/stop NBU?In a clustered environment, VCS commands/utilities should be used, not NBU commands. Well, I am only guessing here.To really know what happened we need to see the Engine_A log. Please extract all the lines from slightly before and after the failover and paste in a text file (e.g. engine.txt) and upload here. Please do not upload the entire file as that is normally too big.

Hi all,We have went through the node of VCS and can see a service name "Veritas Cluster Server Helper" is stopped. What does it do? could this related?Thank you

You need to check Event Viewer logs to see why the service was stopped. Herewith explanation of VCS helper service: https://www.veritas.com/support/en_US/article.TECH31331 Please look at the Engine_A log or extract and post here. The reason for VCS-related 'events' is normally clearly logged in this file.

Netbackup Master server on VCS

13 Replies

RiaanBadenhorst
Level 6
7 years ago
You'll probably find that there is a single service that might be failing, or that it might be taking longer to startup than VCS expects. Set the NetBackup Server resource to non-critical. The restart or failover the service group and then monitor the Event viewer application/system events to see if there is a complaint about any service that doesn't start. If there are none, its probably just a timeout issue and you can increase the online timeout of the NetBackup Resource.

You can also control which resource VCS monitors if you look in the registry but try the above first.
Marianne
Level 6
7 years ago
Are you referring to automatic failover due to a fault of a critical resource?
Or a switch to the other node (manually initiated)?

Can you confirm that the NetBackup resource was online at the time of failover?
Was the NetBackup resource perhaps in faulted state? If so, did you manually clear the fault?

Faults on non-persistent resources are not automatically cleared.
The reason for that is that 'something' happened for the resource to fault. You need to troubleshoot the reason and fix it, then clear the fault.

A manual switch to another node will online a service group to the same state as it was on the source node.

Are your backup resources using VCS commands or VCS GUI to start/stop NBU?
In a clustered environment, VCS commands/utilities should be used, not NBU commands.

Well, I am only guessing here.
To really know what happened we need to see the Engine_A log.

Please extract all the lines from slightly before and after the failover and paste in a text file (e.g. engine.txt) and upload here.
Please do not upload the entire file as that is normally too big.
- tarmizi
  Level 5
  7 years ago
  Hi Marianne,
  Responding to your query, it is an automatic failover from node0 to node1 then out of curiosity, we perform another failover (manually trigger) from node1 back to node0 but the Netbackup service is not up and its why we ran "bpup -v -f" command. And we found an error in Event Viewer (as per shown below) in node0 that might what triggered the automatic failover on VCS.
  Thank you
  - Marianne
    Level 6
    7 years ago
    The Event Viewer error is a Microsoft error.
    
    Please provide Engine log as per previous requests.
tarmizi
Level 5
7 years ago
Hi all,
We have went through the node of VCS and can see a service name "Veritas Cluster Server Helper" is stopped. What does it do? could this related?
Thank you
- Marianne
  Level 6
  7 years ago
  You need to check Event Viewer logs to see why the service was stopped.
  
  Herewith explanation of VCS helper service:
  
  https://www.veritas.com/support/en_US/article.TECH31331
  
  Please look at the Engine_A log or extract and post here.
  The reason for VCS-related 'events' is normally clearly logged in this file.

Forum Discussion

Netbackup Master server on VCS

13 Replies

Related Content

SMTP nofication windows 2K master/media server.

Netbackup Install and upgrade for large server fleet

Cannot make Virtual Netbackup Appliance 3.2 Master Server

Re: Import KMS-files to a new NetBackup Master Server with already imported old catalog (NetBackup 8.2)

NetBackup Client Token for second master server

Recent Discussions

MS-SharePoint policy restore error (2804) .

How to restore a backup

How to configure RBAC

command: bperror

10 years old netbackup appliance database service down, ssl certification out date