Skip to content

driver: deploy csi-addons sidecar in its own DaemonSet #269

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 2 commits into from
Jul 21, 2025

Conversation

black-dragon74
Copy link
Member

Describe what this PR does

This patch introduces a new DaemonSet for csi-addons sidecar that uses pod network instead of the host networking.
Provide some context for the reviewer

Copy link
Collaborator

@Madhu-1 Madhu-1 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@black-dragon74 can you adjust the operator testing to enable csi-addons and verify extra daemonset for rbd?

This patch introduces a new DaemonSet for csi-addons sidecar that
uses pod network instead of the host networking.

Signed-off-by: Niraj Yadav <niryadav@redhat.com>
@black-dragon74 black-dragon74 requested a review from Madhu-1 July 14, 2025 09:06
Signed-off-by: Niraj Yadav <niryadav@redhat.com>
Copy link
Collaborator

@Madhu-1 Madhu-1 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

@Madhu-1
Copy link
Collaborator

Madhu-1 commented Jul 15, 2025

@nb-ohad PTAL

@black-dragon74
Copy link
Member Author

In my testing, sometimes a deployment deadlock is reached due to strict pod anti-affinity rules (pods are unable to be scheduled on the same node as they are already running).

This is leading to new replicaset always being in pending state, unless we go ahead and delete the old pods manually.

To remedy this we should set maxSurge to 0. Discussing it here before including in this PR.

@Madhu-1
Copy link
Collaborator

Madhu-1 commented Jul 17, 2025

In my testing, sometimes a deployment deadlock is reached due to strict pod anti-affinity rules (pods are unable to be scheduled on the same node as they are already running).

This is not related to this PR right?

This is leading to new replicaset always being in pending state, unless we go ahead and delete the old pods manually.

Are you testing on single node or multiplde node. Single node is not a big problem as its only for testing.

To remedy this we should set maxSurge to 0. Discussing it here before including in this PR.

Is this related to this PR? we are using default what kubernetes provides and This is already configurable in the Driver CR unless am missing anything

@black-dragon74
Copy link
Member Author

This is not related to this PR right?

Correct. But we could have the updated strategy as a part of this PR itself. Since it is already in the review process.

@Madhu-1
Copy link
Collaborator

Madhu-1 commented Jul 17, 2025

This is not related to this PR right?

Correct. But we could have the updated strategy as a part of this PR itself. Since it is already in the review process.

This is a different problem, it should have its own issue and one PR not as part of this one.

@black-dragon74
Copy link
Member Author

This is a different problem, it should have its own issue and one PR not as part of this one.

Tracked at #273

@Madhu-1 Madhu-1 merged commit 1ab5ad6 into ceph:main Jul 21, 2025
15 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants