Solved

Hedvig umount 'device is busy

  • 10 September 2022
  • 10 replies
  • 148 views

Badge +3

 

 

When following the document on how to stop/start a Hyperscale X appliance node

 

https://documentation.commvault.com/2022e/expert/133467_stopping_and_starting_hyperscale_x_appliance_node.html

 

I get to the step unmount the CDS vdisk, and after getting the proper vdisk name and running the cmd

# umount /ws/hedvig/<vdiskname>

i get the message ‘device is busy’.

What’s the proper way to remediate this and continue?

Thanks

icon

Best answer by GordA 14 September 2022, 00:10

View original

10 replies

Userlevel 7
Badge +16

Can you check if there are actually no running Commvault processes anymore?

Command: commvault list

Userlevel 7
Badge +16

You missed out on tea with the Mad Hatter 🤪

Jokes aside, hope this was a one time issue 👍

Userlevel 7
Badge +16

No worries :)

The CvMountd is what's keeping the disk busy.

Can you retry a commvault stop and check if all the services have stopped after the second attempt?

If not logs need to be checked why they aren't stopping.

Let us know ;)

Userlevel 7
Badge +23

@GordA let me know if you’re holding off on RCA, or if we should mark your reply as the best answer.

Thanks!

Badge +3

Yeah I think this issue is behind us now.

I was able to continue using the same lazy umount

umount -l /ws/hedvig 

to do the reboots of our other nodes and update to 11.28.19.

Thanks!

G.

Badge +3

Hi, thanks for the reply. Still learning CV and HyperScale X.

Output from commvault list after commvault stop shows 

 

Badge +3

Hi, so in talking to support , the thought was that the services are defunct but the ps -ef | grep defunct doesn’t show any defunct services.

Lazy umount was used to continue the process and reboot the node. After the reboot and restart node, etc the Commserv was able to successfully update the node to 11.28.19

I’ll continue on my other nodes which were also exhibiting the same behaviour of update job failures.

 

G

Userlevel 7
Badge +23

@GordA , were they able to determine any sort of RCA on why this happened?

I know sometimes whatever logging would be needed is either long gone, or the debug level was not high enough at the time...

Badge +3

No, we didn’t get that far down the rabbit hole. 

G

Userlevel 7
Badge +23

Appreciate the confirmation, @GordA !

Reply