[ https://issues.apache.org/jira/browse/FLINK-30757?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17687321#comment-17687321 ]
Shipeng Xie edited comment on FLINK-30757 at 2/11/23 12:16 AM: --------------------------------------------------------------- The e2e [error|https://github.com/apache/flink-kubernetes-operator/actions/runs/4147827390/jobs/7175460340#step:9:1340] in busybox init container is {_}wget: error getting response: Connection reset by peer{_}. It looks like similar error is already reported as an [issue|https://github.com/docker-library/busybox/issues/162]. I pushed another commit to downgrade from 1.36.0 to 1.35.0 but the workflow run still needs approval. was (Author: JIRAUSER296422): The e2e [error|https://github.com/apache/flink-kubernetes-operator/actions/runs/4147827390/jobs/7175460340#step:9:1340] in busybox init container is `wget: error getting response: Connection reset by peer`. It looks like similar error is already reported as an [issue|https://github.com/docker-library/busybox/issues/162]. I pushed another commit to downgrade from 1.36.0 to 1.35.0 but the workflow run still needs approval. > Ugrade busybox version to a pinned version for operator > ------------------------------------------------------- > > Key: FLINK-30757 > URL: https://issues.apache.org/jira/browse/FLINK-30757 > Project: Flink > Issue Type: Improvement > Components: Kubernetes Operator > Reporter: Gabor Somogyi > Assignee: Shipeng Xie > Priority: Minor > Labels: pull-request-available, starter > > It has been seen that the operator e2e tests were flaky when used the latest > version of the busybox so we've pinned it to a relatively old version. > https://github.com/apache/flink-kubernetes-operator/commit/8e41ed5ec1adda9ae06bc0b6203abf42939fbf2b > It would be good to do 2 things > * Upgrade the busybox version to the latest in a pinned way > * Add debug logs of the busybox pod in case of failure (to see why failed) -- This message was sent by Atlassian Jira (v8.20.10#820010)