-
Notifications
You must be signed in to change notification settings - Fork 373
version: Update containerd version to 1.3.0 #2100
version: Update containerd version to 1.3.0 #2100
Conversation
/test |
Codecov Report
@@ Coverage Diff @@
## master #2100 +/- ##
=========================================
Coverage ? 52.67%
=========================================
Files ? 110
Lines ? 16450
Branches ? 0
=========================================
Hits ? 8665
Misses ? 6743
Partials ? 1042 |
Yikes, CI really didn’t like this PR. |
32c9e9f
to
e40e0f5
Compare
/test |
@egernst Yeah, lets try this one more time. |
e40e0f5
to
2992882
Compare
/test |
Could you add the changes to containerd in the commit? All the CI's seem to be failing attempting to build CNI plugins. |
2992882
to
48c6ae8
Compare
/test |
Since |
48c6ae8
to
25af68b
Compare
/test |
@amshinde if this PR is related with kata-containers/tests#2003 then you need to put |
@GabyCT I did put it on the commit. |
/test |
1 similar comment
/test |
@grahamwhaley I will be looking at this again to fix the CI failures. |
/retest |
The cni config used for tests from the cri repo has ipv6 routes. Skip this and use a cni config used for crio as well. Depends-on: github.com/kata-containers/runtime#2100 Fixes kata-containers#2002 Signed-off-by: Archana Shinde <[email protected]>
These include features like privileged containers without host devices and support for per runtime annotations. Depends-on: github.com/kata-containers/tests#2029 Fixes kata-containers#2099 Signed-off-by: Archana Shinde <[email protected]>
25af68b
to
62f64bb
Compare
/test |
@chavafg I suspect this PR has been failing as the VM instances do not seem to support ipv6 in spite of trying to enable it with this PR: kata-containers/tests#2003 |
@amshinde I am not sure, but reading some internet resources it seems that we need to make further configuration to enable ipv6. Will check how to do it and how to enable it using our jenkins plugin. |
@chavafg For now, I have avoided using ipv6 with kata-containers/tests#2029 by not using the cni config file that comes with containerd which included ipv6 routes.
@jcvenegas Do you have any input on this? I know you had added the initial commit to switch to testing with the cni config provided by containerd. Was this added to address a certain failure that I am not aware of. |
@chavafg I am seeing that containerd process exits while doing a |
@amshinde I opened kata-containers/tests#2035 to collect them in the CI |
@chavafg Thanks, I have merged it. Will the entire CI need to be rerun again? |
/test |
@lifupan containerd-cri CI seems to be failing. I dont see much from the containerd logs other than
Do you have any idea? I was able to run containerd 1.3.0 with kata shimv1 and v2 locally. |
Hi @amshinde I think PR kata-containers/tests#2043 can fix this issue. |
/test |
The cni config used for tests from the cri repo has ipv6 routes. Skip this and use a cni config used for crio as well. Depends-on: github.com/kata-containers/runtime#2100 Fixes #2002 Signed-off-by: Archana Shinde <[email protected]>
@lifupan Thanks, looks like that did help. I have added another commit to update containerd config here : kata-containers/tests@61725d5 But looks like one of storage tests is failing now. |
/test |
I have tried reproducing the failures locally. Even tried running the kubernetes CI on an Azure instance. @chavafg Let me know if you can reproduce the containerd CI issue, if you have nay cycles. |
Hi @amshinde, I could reproduce the issue and tried to debug... but it seems that containerd tests are not even running kata. I see pod up, but no kata process. Also the journals of kata-runtime and proxy are empty. |
seems like changes here https://github.com/kata-containers/tests/pull/2029/files#diff-69bf746790abbecab6afee63f20ad2f7L29-R36 are not working correctly, could you make them work locally? |
/test |
1 similar comment
/test |
These include features like privileged containers without host devices
and support for per runtime annotations.
Signed-off-by: Archana Shinde [email protected]