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
openshift node becomes dysfunctional because of devicemapper issues with docker #14601
Comments
Likely core issue: moby/moby#33603 |
@rhvgoyal FYI -- I know you are active on the upstream issue |
Related devicemapper issue: https://bugzilla.redhat.com/show_bug.cgi?id=1461370#c6 Pod creation failed with error: |
Two workarounds so far (both maybe needed):
And I get to push the issue away in future.
|
I updated original issue. We suspect that a recent commit in docker caused this cookie/semaphore leak issue. @nhorman is looking into writing a patch for it now. |
Check if following PR fixes the issue. |
Hi, I'm trying to help someone who is having similar issue with Docker 1.12.6; those references to Docker issue and PR all seem to reference recent changes, but you wouldn't be using such a recent Docker with Kubernetes, would you? |
sudo printf 'kernel.sem = 250\t32000\t32\t8192\n' > /etc/sysctl.d/99-kernelsem.conf this works for me , since it will keep permantently the semaphore limit as 8192 . one can issue the command "ipcs -su" to know how manay semaphores are in use , but how can I know what process is using these semaphore , can anone help me , thank you . |
We have backported fix for this in projectatomic/docker as well. So please take latest docker build from your source and it might have the fix. projectatomic/docker#256 |
We are using docker-1.12.6-28.git1398f24.el7.centos.x86_64.rpm and we are seeing the following issues on an upgrade of docker version. Jun 29 18:49:13 openshift-master-01 systemd[1]: Starting Docker Application Container Engine... It looks like this is fixed in the upcoming docker-1.12.6-32.git88a4867.el7 version
An alternative appears to be downgrade/remain at docker-1.12.6-16.el7 according to the Bugzilla report. Edit: |
i'm afraid the issue is not fixed in the docker version mentioned above
and
as i still got hit by it |
It is actually likely PR #33376 that you are after: |
@nhorman do you know if that's back-ported to an existing 1.12.6 version? |
@xThomo not off the top of my head, no, but it should be pretty easy to go and see, its a very small patch |
cheers @nhorman ! |
Issues go stale after 90d of inactivity. Mark the issue as fresh by commenting If this issue is safe to close now please do so with /lifecycle stale |
Stale issues rot after 30d of inactivity. Mark the issue as fresh by commenting If this issue is safe to close now please do so with /lifecycle rotten |
I don't see this issue anymore, plus the workaround steps unblock me anyway. Closing this issue. Please reopen if needed. |
openshift node becomes dysfunctional because of devicemapper issues. Happens on a running cluster, where the node was previously functioning well.
See issue: moby/moby#23089
Version
openshift v3.6.94
kubernetes v1.6.1+5115d708d7
etcd 3.1.0
Steps To Reproduce
Current Result
Node not ready
Origin node gives this error in the log:
Expected Result
Node should be ready and pods should run
Additional Information
Docker daemon restart gives this error:
The text was updated successfully, but these errors were encountered: