backports: for v1.10.6 #11471
Merged
backports: for v1.10.6 #11471
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
frezbo
approved these changes
Jul 30, 2025
Also go1.24.5, Kubernetes default to 1.33.3. Signed-off-by: Andrey Smirnov <andrey.smirnov@siderolabs.com>
This issue is specifically bad with single node controlplanes, and masked by retries in multiple control planes. The core issue is that Talos had a race between creating a `talos` `Service` and `Endpoint`, which might end up with `Endpoint` created without `Service`, and Kubernetes EndpointController cleans up orphaned endpoints. In multiple control planes scenario, and while enabling not on bootstrap, which might be masked by multiple machines re-creating endpoints and the issue is not seen. Fixes siderolabs#11311 Co-authored-by: Mateusz Urbanek <mateusz.urbanek@siderolabs.com> Signed-off-by: Andrey Smirnov <andrey.smirnov@siderolabs.com> (cherry picked from commit 54bd50b)
See siderolabs#11210 This doesn't fix anything, but the logs will be more helpful to understand what exactly is wrong. Signed-off-by: Andrey Smirnov <andrey.smirnov@siderolabs.com> (cherry picked from commit a966321)
See siderolabs/pkgs#1277 Signed-off-by: Andrey Smirnov <andrey.smirnov@siderolabs.com> (cherry picked from commit 506212a)
This is to make them more easily identifiable for clean up if required. Signed-off-by: Andrew Longwill <andrew.longwill@siderolabs.com> (cherry picked from commit bfc57fb)
This showed up in docker runs (not sure why only docker), but the issue is the following: * a service is running which has some volume requirements * `VolumeMountRequests` are created, and `VolumeMountStatus` were established * the service put finalizers on `VolumeMountStatus` * now the service is going to be restarted - so at first it's going to be shut down * on shutdown, the service will remove `VolumeMountRequest`, and remove finalizers on `VolumeMountStatus` * now it's job of other controllers to tear down and remove mounts * as the service starts back up after restart, it will re-create `VolumeMountRequest`, and will try to wait and put finalizers on `VolumeMountStatus` * here comes the race condition: it can be that the service sees tearing down `VolumeMountStatus` which is left from the shutdown time, so it puts a finalizer on it, and it blocks the proper teardown of the previous "generation" of the mount request/status, leading to a deadlock So the fix is to wait for the new status to be created which is not tearing down. Signed-off-by: Andrey Smirnov <andrey.smirnov@siderolabs.com> (cherry picked from commit 06ef710)
Import the fix from siderolabs/go-blockdevice#135 Signed-off-by: Andrey Smirnov <andrey.smirnov@siderolabs.com> (cherry picked from commit d62e255)
|
/m |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
PRs backported: