elek opened a new pull request #1020: URL: https://github.com/apache/hadoop-ozone/pull/1020
## What changes were proposed in this pull request? HDDS-3461 introduced a new CSI smoketest (to be sure that the CSI daemon can be started). It was reverted because a failure on the master and commited after an additional check is added to wait until the CSI socket is created. Unfortunately this check is bad. In some cases it can fail: For example in here: https://github.com/jsoft88/hadoop-ozone/runs/734147343?check_suite_focus=true ``` connection error: desc = "transport: Error while dialing dial unix /tmp/csi.sock: connect: no such file or directory" ``` Thanks to @jsoft88 , who reported this problem. ## What is the link to the Apache JIRA https://issues.apache.org/jira/browse/HDDS-3717 ## How was this patch tested? This is the important part. I tested the earlier version, but due to other errors in my `/tmp` it showed false positive result. Now I tested it with executing robot test: ``` robot hadoop-ozone/dist/src/main/smoketest/csi.robot ``` You can see that without the patch it passes the `Check if CSI server is started` phase even if `/tmp/csi.socket` is not created locally. After apply the patch, you can do the same: ``` robot hadoop-ozone/dist/src/main/smoketest/csi.robot ``` The same check at the beginning should be pending and failed 3 minutes. You can also create a socket during the 3 minutes: `nc -U /tmp/csi.sock -l` In this case the check will be passed. ---------------------------------------------------------------- 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: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
