Solved

Kubernetes - How do our restores actually function? Plus some bonus backup questions

  • 6 January 2021
  • 2 replies
  • 264 views

Userlevel 1
Badge +1

As the Kubernetes documentation is very light, I am not entirely certain as to how they function.

 

1.Full application restore: Does this cache/swap on disk on the VSA proxy before committing to the K8S environment? If so, are the disk space requirements detailed? Is it 100% of the application size?

2.Out-of-place full application restore: Do we restore the config files and PVC?  Do we edit the config to have the application connect to a new PVC?

3.Will we be able to specify a PVC for restores? Or it just auto-creates based upon the new application name?

4.For file level restores to a PVC, you are prompted for volume and path. Path is not covered in our documentation. What is the default location and what is the format for this? Will a browse button become available for path?

Backup questions:

1. Will we ever be able to read data in from the worker nodes directly?  Or always worker to master to VSA?

2.Customers may be using AKS or other services where the master node is outside of their Commvault environment.  Will we be able to proxy the communications - ie. have the VSA reach out via an HTTPS proxy to get to the master node/application server?

 

 

icon

Best answer by DJoshi 6 January 2021, 20:32

View original

2 replies

Userlevel 1
Badge +2

Reached out to k8s team for answers -

1.Full application restore: Does this cache/swap on disk on the VSA proxy before committing to the K8S environment? If so, are the disk space requirements detailed? Is it 100% of the application size?

Answer: Job results cache is used only for config files. No cache/swap needed for application data restore.  

2.Out-of-place full application restore: Do we restore the config files and PVC?  Do we edit the config to have the application connect to a new PVC?

Answer: Yes and yes.

3.Will we be able to specify a PVC for restores? Or it just auto-creates based upon the new application name?

Answer: We autocreate the new PVC.

4.For file level restores to a PVC, you are prompted for volume and path. Path is not covered in our documentation. What is the default location and what is the format for this? Will a browse button become available for path?

Answer: Default location is the root of the volume. We will not show a browse button for the volume but a path can be entered manually and the data will be restored to that path.

Backup questions:

1. Will we ever be able to read data in from the worker nodes directly?  Or always worker to master to VSA?

Answer: Currently no. But this may change in the future.

2.Customers may be using AKS or other services where the master node is outside of their Commvault environment.  Will we be able to proxy the communications - ie. have the VSA reach out via an HTTPS proxy to get to the master node/application server?

Answer: Yes as long as the access node can access the API node over the proxy connection we are fine.

 

Userlevel 7
Badge +23

Hey @Will Hawkins - how did you go with setting this up? all good?

Reply