When use crictl rmp -a -f, Avoid excessive coroutines causing context… #1749
Conversation
k8s-ci-robot
added
kind/bug
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: jokemanfire The full list of commands accepted by this bot can be found here.
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
k8s-ci-robot
added
the
size/XS
… timeout Increase the maximum number of co processes limit Signed-off-by: jokemanfire <[email protected]>
|
@jokemanfire have you tried to increase the default timeout, which is rather short, instead? |
It could , But is it more user-friendly to increase the concurrency limit on goroutines here？Otherwise, I would have to try a better threshold on timeout. Thanks |
I would not say it's more user friendly, because it may be intentional to have more than |
In my opinion, the design should be more inclined towards thread pools or database connection pools (given a limit, after all, containerd processing requests also have their own concurrency). If 1000 pods are deleted and 1000 coroutines or threads are started at the same time, I think this is a problem in the design, after all, crictl is not a tool for testing concurrency limit. This is just my opinion. It may be better to conduct concurrency testing using a cri-test. |
There were some issues when creating approximately 400 containers and 200 pods.When I use crictl rmp -a -f.
The error occur
removing the pod sandbox "5c963da93d21cb6af42513ddbb682491aac7c91fd687f8e3bbf4d51a7dfaffed": rpc error: code = DeadlineExceeded desc = context deadline exceededWhat type of PR is this?
/kind optimize
What this PR does / why we need it:
Increase the maximum number of co processes limit.