Feature Request: Support for revisionHistoryLimit

[reefland]

Is this a bug report or feature request?

• Feature Request

Enable support for revisionHistoryLimit to be specified in the CephCluster CR, which can then be used in the helm chart.

• To allow reducing the Kubernetes default of 10 to something else like 3 as every plugin, ceph-tools, crash collector, mds, rgw, mgr, operator, every OSD will maintain 10 copies.
• Loading the rook-ceph namespace and working with it from the CLI or a GUI like ArgoCD is cumbersome, waiting for all the copies of each objects to load and render.

This was discussed #12693 where @travisn started a list of changes to be considered.

[travisn]

Since this is not dependent on helm installs, I removed Helm from the title since all installs would benefit from this feature request.

[ashutosh887]

Let me work on this @subhamkrai

[subhamkrai]

@ashutosh887 I see you have commented on a few other issues to work on. How about you pick any one issue first and move to the next issue? Thanks for looking at the issues :)

[ashutosh887]

@subhamkrai
Did you assign me on any of them?

[subhamkrai]

@subhamkrai Did you assign me on any of them?

@ashutosh887 you can self-assign this issue now https://github.com/rook/rook/blob/master/Documentation/Contributing/development-flow.md#self-assign-issue

[ashutosh887]

Thanks a lot @subhamkrai
I didn't knew that's enabled here

/assign

[github-actions]

Thanks for taking this issue! Let us know if you have any questions!

[subhamkrai]

Thanks a lot @subhamkrai I didn't knew that's enabled here

sure

[travisn]

Unassigned due to inactivity

[BlaineEXE]

I think it would be best to provide this configurable on each CRD that creates ReplicaSets as well. CephCluster, CephObjectStore, CephFilesystem, CephNFS are the big ones I can think of.

I wonder if the other controllers should default to the CephCluster value if unset as well.

@travisn thoughts?

[obnoxxx]

changed the label from Bug to Feature since that's what the title and description say.

[travisn]

I think it would be best to provide this configurable on each CRD that creates ReplicaSets as well. CephCluster, CephObjectStore, CephFilesystem, CephNFS are the big ones I can think of.

I wonder if the other controllers should default to the CephCluster value if unset as well.

@travisn thoughts?

If there is a reason to customize each daemon, sounds good to add it to each CR, but perhaps we just start with the CephCluster CR setting to keep it simple for all daemons to inherit.

[Sheetalpamecha]

/assign

[github-actions]

Thanks for taking this issue! Let us know if you have any questions!

[obnoxxx]

@Sheetalpamecha I'm also interested in taking this up but didn't want to just take it away from you. Would you be OK if I also started working on a patch for this? If you have already started serious work on it, I'll refrain from looking further 😄

[Sheetalpamecha]

Hey @obnoxxx I'm almost done with the patch. :)

[travisn]

@Sheetalpamecha Still planning on this?

[github-actions]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed in a week if no further activity occurs. Thank you for your contributions.

[parth-gr]

./assign

[github-actions]

Thanks for taking this issue! Let us know if you have any questions!

[BlaineEXE]

I wonder if this should be an operator-scope setting, like ROOK_REVISION_HISTORY_LIMIT. I don't imagine that many users would want to change this setting on a per-cephcluster basis -- probably cluster-wide instead. And an operator setting might make it more clear that this behavior will be used for all Rook pods and not just the CephCluster resource.