Solved

Operation kImageDelete not allowed, catalog item is owned by cluster

  • 10 November 2018
  • 9 replies
  • 2164 views

Userlevel 1
Badge +1
So I had to replace my prism central with a new one, unfortunately it had already taken ownership of the images. With my new prism central at I can not delete any of the old images and I get this error. Operation kImageDelete not allowed, catalog item is owned by cluster 4452164c-24b6-479d-b16b-a8635fbe2741, which I believe is my old prism central UUID. I can't delete them from from the clusters locally because they are hooked to prims central. This is pretty frustrating as I am trying to use PC as my main entry point for managing and deploying vm.
icon

Best answer by Sneha Patekar 10 December 2019, 23:44

View original

This topic has been closed for comments

9 replies

Yes, same problem for me. Is there a way via the CLI to delete those old Images? Tried with acli image.delete but getting the same error.
Same here...
Badge +1
Same issue 😞
Same issue. Has anyone had any luck getting this resolved?
Badge +4
Same problem as the others. Has any one any ideas please ?
For my issue, I ended up contacting support and found that I had images tied to a Cluster that was no longer a part of Prism Central. The only way I could get the images out of Prism Central was for support to provide me with a python script to forcibly remove the bunk images.
Same problem here.
Joe_E could you please share the python script that support provided here?
Userlevel 1
Badge +7

I know this thread is a bit dated, but I wonder if anyone has discovered how to address this problem. I too have a bunch or orphaned image files that are owned by a deleted PC instance. I can’t remove them via Prism element, and acli reports the error: NotSupported: Operation kImageDelete not allowed, catalog item is owned by cluster 000b4257-317… Any ideas what the elusive, support-provided Python script might be doing under the hood? Is there a way to update the ownership metadata on the image files so they will re-associate with the original CE cluster, perhaps? 

Userlevel 2
Badge +1

@DavidTeague @Claus Jan @DennisFFM @tanel @Joe_E @storm @Steven2018 @jeffearl 

Thank you for sharing your questions here, I apologize you had to wait it so long.

To address your issue, the issue in specific was earlier resolved using a workaround which resulted in many other issues. 

But this issue is now resolved in AOS 5.11 and 5.10.5.

 

The following cases are addressed by these changes:

  • Allow images to be deleted if it owned by an older PC from a new PC.
  • If image is an orphan image (which means just present on PC) it can be deleted.
  • If image is on a PE and is not owned by that PE - The image can either be deleted by unregistering the PE or migrating image to new PC and deleting from PC.

Recommendation to customers having this problem is to upgrade to 5.10.5 or later and then attempt to clean up the images again.

As an alternative when an upgrade is not possible or the issue remains post upgrade is to use a new script has been developed by engineering. This script will be part of AOS bundle as of 5.10.9 and 5.11.2. For previous versions it can be downloaded manually. This script will not make any catalog RPC calls and rather just remove metadata cleanly from IDF and it is the last resort to have a way to clean-up the entire image.

I am concerned that the script is not openly shared with the customers but will have to open a Nutanix case with Support who will further help with the issue. 

I hope this was helpful!!

 

Thank you,

Sneha P