Solved

Probem with restore AKS


Userlevel 1
Badge +5

Hello All, 

From few days we are struggling with restore AKS from our environment. Commvault  backup working without any issue. We tried to restore AKS by using Linux Node and Windows Node. Every time we received the same issue. Error is: 

Error Code: [91:145]
Description: Unable to write data to the Virtual Disk [/opt/commvault/iDataAgent/jobResults/CV_JobResults/iDataAgent/VirtualServerAgent/Restore/201136/solace-dev`HelmChart`solace-dev`d91d2911-dfc4-4c32-a6ed-2e1bd553bac6/VMConfig.cvf]. Please ensure that the proxy is able to communicate with the virtual machine host and that there is sufficient disk space available. Details: [Success]
Source: 10.29.164.144, Process: vsrst

Could someone can help or show the way how investigate that type of issue?

Regards, 

Michal 

icon

Best answer by Denney 12 August 2022, 17:03

View original

14 replies

Userlevel 7
Badge +23

Hi @Michal128 , thanks for the post!

I’ve seen a few cases with this error.

Most were resolved by restoring to a different data store and at least one was caused by AV scanning the directories involved.

Here’s the link for the Windows recommendations:

https://documentation.commvault.com/11.24/essential/8665_recommended_antivirus_exclusions_for_windows.html

Can you take a look at these (and/or try to restore to another data store) and update us here?

Thanks!

 

Userlevel 1
Badge +5

Hello Mike, 

Thanks for suggestion, but exclusions on AV is added to the server which store data and which one prepared the restore process.  Currently we are using Linux VSA proxy server (Access Node) without any AV. 

About storage account. I don’t see option to restore Application in-place with set storage account for restore. 

In think the most important option for restore is in-place restore. 

Regards, 

Michal  

Userlevel 7
Badge +16

Are you running the latest maintenance release? I know they have made a lot of improvements over the last few months. 

Userlevel 7
Badge +23

Appreciate the reply!

Check what @Onno van den Berg is suggesting, though if that is not the solution, I’d advise creating a support case to dive deeper.

If you do, share the case number here so I can track accordingly!

Userlevel 1
Badge +5

Hello, 

Opened incident  case number is: 220629-286

I am working on FR25, I know that is not the most new one, but the same one we have on production server. 

Currently restore files from PVC is working, but we can;t still restore entire application. 

Regards, 

Michal 

Badge

Hello, I’m also getting this error. I’m running 11.24.56 and k8s v1.21. Local k8s cluster in my case, as opposed to AKS.

I’m able to restore to a different namespace (same k8s cluster) but attempts to restore to the same namespace (whether overwriting the existing application or doing an out-of-place restore to a new application name) fails with the message:

Unable to write data to the Virtual Disk [application-name]. Please ensure that the proxy is able to communicate with the virtual machine host and that there is sufficient disk space available. Details: [Success]
Source: dremio_ap, Process: vsrst

I’m running Commvault as an all-in-one on a Win2019 machine. k8s cluster is bare-metal and not very utilized.

Userlevel 7
Badge +18

Hi @Denney - the previous issue here resulted in a patch that is currently being created and validated. Without investigation we can’t be sure if its the same issue, but the symptoms seem to match.

The internal patch (form ID) is 2483 for FR28.

From here I’d suggest a support case and reference the above updates for FR28 and ask to validate if it could be the same issue.

 

 

Userlevel 7
Badge +16

Not sure if it is actually true, but I heard FR28/2022e contains an updated/enhanced foundation for K8 workloads. Can someone confirm this? Would that for example mean that for customers who have a large K8 install basis to move to Fr28 once it reaches GA? 

Userlevel 7
Badge +23

@Mathew Ericson , going right to the source.  Can you confirm @Onno van den Berg ‘s question above?

Badge

Thank you all for the attention!

Customer support helped me open 220809-603 to help track.

At the moment, my email ID isn’t associated with a partner account so I’m not able to view the ticket just yet. But I’m happy to post the log archive once I’m able.

Userlevel 4
Badge +5

@Mathew Ericson , going right to the source.  Can you confirm @Onno van den Berg ‘s question above?

 

Sure @Mike Struening and @Onno van den Berg - yes 2022e has a new protection mode for applications that protects all resources in the namespace and at the cluster level (cluster-scoped or non-namespaced objects).

But the error above looks more related to existing functionality and a patch has been identified by Damien. Best course of action here is a support case - and of course upgrade to latest :)

Life moves pretty fast in the Kubernetes world - running the latest possible Commvault release is recommended to get the latest supportability and fixes.

Userlevel 7
Badge +16

Thanks for the quick confirmation. I totally agree with your statement however a lot of customer stick to specific LTS versions hitting issues who are fixed in newer feature releases who contain massive changes who cannot be backported. Now 2022e is going to be a LTS release so customers moving to that version will be good for K8 but the next release might deliver new under the hood improvements. Most of the times the under the hood improvements are not being mentioned both on the documentation page and or during the feature release webcast. I think marketing and advocating these improvement is smart todo and the community might be the right medium for it. 

Badge

Hello all,
I wanted to reply with an update/resolution for my particular situation.

  1. I reinstalled Commvault on Aug 10th and may have picked up a new MR (my prior install was July 27th) as I’m now able to perform in-place restores.
  2. Commvault Support helped determine that the out-of-place restores to same namespace is presently unsupported for applications deployed via HELM chart.

Thank you for the assistance!

Userlevel 7
Badge +23

Agreed, thanks @Onno van den Berg and @Mathew Ericson for the help!

Reply