Solved

Problem with one node of a four cluster FileServer

  • 9 September 2021
  • 8 replies
  • 627 views

Userlevel 4
Badge +15

We have a file server cluster with four nodes. Each node has a certain number of resources (drives) which can also move from one node to another. 
The second node had to be taken out of the cluster a few weeks ago and is now operational again. If resources move to this node, they can no longer be backed up, because the Commvault plugin does not recognize this. 
It only works again when I select the resource in Commvault and then perform a manual synchronization. 
Has anyone ever had this problem ? 
I also have a few screenshots attached to visualize the problem a bit.

Regards

Thomas

icon

Best answer by thomas.S 15 September 2021, 16:22

View original

8 replies

Userlevel 6
Badge +13

That doesn’t seem to be out of the ordinary. Do you have to sync that node up every time the backup runs? If it’s only when the cluster is brought back that should be normal if changes were made that would\could have affected the registry that tells the plugin service how to work. 

Are you able to manually bring the plugin service online when that node is introduced again? 

Userlevel 4
Badge +15

Hi @Aplynx
 

Does this mean that I only need to do the manual syn the first time a node is added back to the cluster and when these resources automatically change, this is also automatically noticed by the plugin ? 
The remaining three nodes have no problems here when a resource moves to another node. 
On the node I can not start the plugin. As soon as I force all nodes to sync via Commvault, the resources that were moved to the new node come online with the Commvault plugin. 

 

Userlevel 6
Badge +13

It sounds like the cluster is missing the registry information to start the plugin which is what the sync process is bringing back. It’s odd that you have to sync the plugin, but the physical WinFS itself is intact. It’s usually not a bad idea to sync it either way when it’s brought back online just to make sure that any changes are seen by that machine. You should only need to do it the one time, unless it’s removed again.

Userlevel 4
Badge +15

We added the node back to the cluster and packed resources onto it. Afterwards we got the services running for the resources in Commvault with a sync. 
Now we will wait to see if when resources migrate this is automatically detected by Commvault. 

Userlevel 7
Badge +23

Appreciate the status update, @thomas.S !  Looking forward to confirmation.

Userlevel 7
Badge +23

Hi @thomas.S , following up to see if this was resolved based on @Aplynx ‘s advice.

Thanks!

Userlevel 4
Badge +15

Sorry for the late feedback. 
Since we have not seen any more problems so far, I assume that the problem can also be considered fixed.

Userlevel 7
Badge +23

All good!  Glad to get late good news versus early bad news!

Reply