onichols-pivotal edited a comment on pull request #5230: URL: https://github.com/apache/geode/pull/5230#issuecomment-641658570
@jdeppe-pivotal Since this commit, Upgrade test has consistently failed in ClientServerMiscBCDUnitTest. testPingWrongServer, both in the develop pipeline and the PR pipeline (even your own PR #5228 is now hitting this). Examples: 1. [UpgradeTestOpenJDK8 #254](https://concourse.apachegeode-ci.info/teams/main/pipelines/apache-develop-main/jobs/UpgradeTestOpenJDK8/builds/254) 2. [UpgradeTestOpenJDK11 #250](https://concourse.apachegeode-ci.info/teams/main/pipelines/apache-develop-main/jobs/UpgradeTestOpenJDK11/builds/250) 3. [UpgradeTestOpenJDK11 #251](https://concourse.apachegeode-ci.info/teams/main/pipelines/apache-develop-main/jobs/UpgradeTestOpenJDK11/builds/251) 4. [UpgradeTestOpenJDK11 #8214](https://concourse.apachegeode-ci.info/teams/main/pipelines/apache-develop-pr/jobs/UpgradeTestOpenJDK11/builds/8214) 5. [UpgradeTestOpenJDK11 #8216](https://concourse.apachegeode-ci.info/teams/main/pipelines/apache-develop-pr/jobs/UpgradeTestOpenJDK11/builds/8216) It did pass in: 1. [UpgradeTestOpenJDK8 #255](https://concourse.apachegeode-ci.info/teams/main/pipelines/apache-develop-main/jobs/UpgradeTestOpenJDK8/builds/255) 2. [this revert PR](https://github.com/apache/geode/pull/5230) This rules out the possibility of an environmental issue, so this PR has already served some useful purpose even with a -1 blocking it from being merged. The ones that did fail have various [n] suffix for the failing test, which along with the 1 out of 6 that passed, suggests some degree of indeterminism. If the failure had nothing to do with this commit, perhaps the Upgrade tests are sensitive to test ordering? https://issues.apache.org/jira/browse/GEODE-8236 has been filed for this issue, although it looks like it has been seen before in https://issues.apache.org/jira/browse/GEODE-8176 On a recent [dev list discussion](https://lists.apache.org/x/thread.html/rcccca408873c92df789ddf7f2f871f9251498d4b902b9019da7200eb@%3Cdev.geode.apache.org%3E) many in the community favored revert-first-then-figure-it-out, since the alternative of just leaving the pipeline and all PRs broken isn't great either. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
