fix: randomize lxc container names #651
Draft
Conversation
This file contains 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
mikemccracken
requested review from
rchincha,
smoser and
hallyn
as code owners
|
Draft because it might need work if it breaks CI. the test suite is not reliable on my system right now so I am just using the github runners |
In situations where concurrent stacker runs are happening on a system, and they are building containers with the same name, and they are being done inside a mount namespace, and the path name given as the roots dir is the same, but the actual mounted volume is different, then both stackers will be able to acquire the file lock at $rootdir/.lock, and will go ahead and start containers named $name, which will then race to set up the lxc control socket, which is named after the container name and the rootfs path, which are both the same here. The fix is to add some randomness to the lxc container name, which ensures that the socket won't clash. This should not affect other uses of the image name, which will still use the un-randomized name. Signed-off-by: Michael McCracken <[email protected]>
mikemccracken
force-pushed
the
2024.11.01/main/uniquify-container-names
branch
from
669c12a to
74db868
Compare
|
The CI failures are real, there is a case now where a build creates a new layer when it shouldn't. empty-layer.bats checks for this and that's what's failing. However it isn't obvious to me what exactly is going wrong, I spent a half day looking at it before Thanksgiving and didn't conclude. Will revisit later. Maybe the Right Way is to have lxc include the mount ns id in its socket naming code? |
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Outdated suggestions cannot be applied.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.
What type of PR is this?
bug
Which issue does this PR fix:
In situations where concurrent stacker runs are happening on a system, and they are building containers with the same name, and they are being done inside a mount namespace,
and the path name given as the roots dir is the same, but the actual mounted volume is different,
then both stackers will be able to acquire the file lock at $rootdir/.lock, and will go ahead and start containers named $name, which will then race to set up the lxc control socket, which is named after the container name and the rootfs path, which are both the same here.
What does this PR do / Why do we need it:
The fix is to add some randomness to the lxc container name, which ensures that the socket won't clash. This should not affect other uses of the image name, which will still use the un-randomized name.
If an issue # is not available please add repro steps and logs showing the issue:
#645
Testing done on this change:
manual only, CI pending
By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.