[release/1.6] Fix memory leak with kubectl exec >= 1.30.0#10574
Conversation
SHA: 60c4c2b2521fb454ce69dee737e3eb91a25e0535 Signed-off-by: Davanum Srinivas <[email protected]>
Signed-off-by: Davanum Srinivas <[email protected]>
Signed-off-by: Davanum Srinivas <[email protected]>
Signed-off-by: Davanum Srinivas <[email protected]>
Signed-off-by: Davanum Srinivas <[email protected]>
Signed-off-by: Davanum Srinivas <[email protected]>
|
Wondering if these should be an |
Signed-off-by: Davanum Srinivas <[email protected]>
Done! thanks @thaJeztah |
mikebrow
left a comment
There was a problem hiding this comment.
Similar to the review of the borrow/sync PR on 1.7, code review looks good to me appears to not introduce new bad behavior :-) just need to sanity test against older k8s clients
As with 1.7 I also suppose a sanity check v1.23 min through 1.31 is needed, for 1.6, each time we change out the streaming protocol code base..
|
tested the following crictl versions. Note that the 1.30.0 and up has both protocols, so tested both. |
kubectl exec >= 1.30.0

xref: #10568
xref: kubernetes/kubernetes#126608
Fix for panic in the issues logged above are already in latest kubernetes master specifically in staged repositories. Unfortunately both 1.6 and 1.7 are on really old versions of k8s and it's not feasible/practical to move to newer versions of k8s. So we instead copy over the
wsstreampackage which has the fix and change references in our code to the copy here.