2020-02-21 10:21:31 UTC - Jon Bennett: @Jon Bennett has joined the channel ---- 2020-02-21 17:01:27 UTC - rani: @rani has joined the channel ---- 2020-02-22 00:07:20 UTC - Roman Popenov: @Roman Popenov has joined the channel ---- 2020-02-22 00:08:57 UTC - Roman Popenov: Hello! I am running into issues while running `build.sh` script in docker folder (<https://github.com/apache/pulsar/blob/master/docker/build.sh>). ---- 2020-02-22 00:10:48 UTC - Roman Popenov: ---- 2020-02-22 00:11:39 UTC - Roman Popenov: It appears that <https://github.com/apache/pulsar/blob/master/pulsar-client-cpp/docker/build-wheel-file-within-docker.sh#L25> is what causes the failure ---- 2020-02-22 00:12:27 UTC - Roman Popenov: I think this might be related to <https://stackoverflow.com/questions/30091681/why-does-docker-prompt-permission-denied-when-backing-up-the-data-volume> I am not exactly sure what would be the appropriate way to solve this issue (without just giving too many permissions or along those lines) ---- 2020-02-22 00:13:38 UTC - Roman Popenov: My ultimate goal is to build images using Java 11 ---- 2020-02-22 00:13:46 UTC - Roman Popenov: Has anyone successfully done this? ---- 2020-02-22 00:15:07 UTC - Devin G. Bost: I haven’t tried running the build images using Java 11, but I haven’t had those issues with running `build.sh`. I wonder if it’s related to how your environment is configured. ---- 2020-02-22 00:16:07 UTC - Roman Popenov: Yeah, I was wondering the same :thinking_face: ---- 2020-02-22 00:16:33 UTC - Devin G. Bost: I haven’t seen that “permission denied” exception, or if I did, it was so long ago that I don’t remember it. ---- 2020-02-22 00:17:24 UTC - Roman Popenov: I did try with adding to the docker run command in `build-wheels.sh` <https://github.com/apache/pulsar/blob/master/pulsar-client-cpp/docker/build-wheels.sh#L78> to `--privileged` ---- 2020-02-22 00:18:54 UTC - Roman Popenov: That didn’t help :confused: ---- 2020-02-22 00:19:34 UTC - Devin G. Bost: Have you tried building the original image before you made any changes? ---- 2020-02-22 00:19:39 UTC - Devin G. Bost: That might be a useful sanity check. ---- 2020-02-22 00:21:18 UTC - Roman Popenov: Yes, I have ---- 2020-02-22 00:21:21 UTC - Roman Popenov: It is the same issue ---- 2020-02-22 00:21:29 UTC - Roman Popenov: I am trying to build 2.5.0 branch ---- 2020-02-22 00:21:38 UTC - Roman Popenov: I haven’t tried master ---- 2020-02-22 00:22:28 UTC - Devin G. Bost: I would try master. I’d also try an earlier release branch, like 2.4.0 or 2.4.2. If you continue getting the issue, it’s probably related to your docker installation and settings. ---- 2020-02-22 00:23:16 UTC - Roman Popenov: I should build my docker container in docker I guess :joy: ---- 2020-02-22 00:23:34 UTC - Devin G. Bost: I had that idea cross my mind, but I don’t know if it will cause side-effects or not. It’s worth a shot. ---- 2020-02-22 00:24:04 UTC - Roman Popenov: I would like to avoid it :stuck_out_tongue_closed_eyes: +1 : Devin G. Bost ---- 2020-02-22 00:24:33 UTC - Roman Popenov: > it’s probably related to your docker installation and settings. Is there anything special that I should do? ---- 2020-02-22 00:25:09 UTC - Devin G. Bost: I’d probably just dig through stack overflow answers. I’m not a docker expert, so I’m not sure. ---- 2020-02-22 00:26:32 UTC - Sijie Guo: if you are building docker image, use `mvn clean install -DskipTests -Pdocker` +1 : Devin G. Bost ---- 2020-02-22 00:27:26 UTC - Roman Popenov: So it’s not package? ---- 2020-02-22 00:27:41 UTC - Sijie Guo: package works as well ---- 2020-02-22 00:28:05 UTC - Roman Popenov: I was simply running *`docker/build.sh`* for now ---- 2020-02-22 00:28:06 UTC - Sijie Guo: `mvn package` or `mvn install` works. ---- 2020-02-22 00:28:29 UTC - Roman Popenov: I will try to change the command in the script to `mvn clean install -DskipTests -Pdocker` ---- 2020-02-22 00:28:35 UTC - Sijie Guo: that’s fine ---- 2020-02-22 00:28:43 UTC - Sijie Guo: docker/builder.sh works ---- 2020-02-22 00:28:54 UTC - Sijie Guo: but it will run all the tests ---- 2020-02-22 00:29:42 UTC - Roman Popenov: I am hitting <https://apache-pulsar.slack.com/archives/C5ZSVEN4E/p1582330248324500> ---- 2020-02-22 00:30:52 UTC - Sijie Guo: @Roman Popenov did you build the project before? ---- 2020-02-22 00:31:37 UTC - Sijie Guo: check those files’ permissions ---- 2020-02-22 00:32:14 UTC - Sijie Guo: building docker images will build the python clients in a docker. ---- 2020-02-22 00:32:46 UTC - Sijie Guo: if you have manually build python clients or cpp clients without docker, that means files are created before with different permissions. ---- 2020-02-22 00:33:06 UTC - Roman Popenov: oh, yeah, I think I have done that long long time ago ---- 2020-02-22 00:33:09 UTC - Sijie Guo: run `mvn clean` first to clean files from previous build, and run `docker/build.sh` crossed_fingers : Roman Popenov ---- 2020-02-22 00:33:15 UTC - Roman Popenov: I’ve manually build the cpp client once ---- 2020-02-22 00:33:30 UTC - Roman Popenov: Will do ---- 2020-02-22 00:35:01 UTC - Sijie Guo: kk ---- 2020-02-22 00:37:56 UTC - Roman Popenov: Same error ---- 2020-02-22 00:40:53 UTC - Sijie Guo: `mvn clean` probaby doesn’t remove those files. ---- 2020-02-22 00:41:04 UTC - Sijie Guo: `git clean -fdx` to clean those files. ---- 2020-02-22 00:41:09 UTC - Sijie Guo: then run `docker/build.sh` ---- 2020-02-22 00:41:43 UTC - Devin G. Bost: Yeah, manually building Python and CPP can really make a mess of things… ---- 2020-02-22 00:44:08 UTC - Roman Popenov: Let me clean the docker cache and co while I am at it ---- 2020-02-22 00:44:20 UTC - Devin G. Bost: Sorry to change the subject here, but @Sijie Guo, when you get a chance, I’m getting a really odd result from the Github actions… I tried rebasing, but I’m getting the same issue. I’d message @tuteng, but he’s probably on his Saturday break. What’s happening is that all of the actions in my PR are passing in under 25 seconds. It’s thinking that my code changes are documentation-only changes. We can chat in a thread. ---- 2020-02-22 00:44:46 UTC - Devin G. Bost: ---- 2020-02-22 00:44:55 UTC - Devin G. Bost: I also messaged @Ali Ahmed, but he hasn’t gotten back to me. ---- 2020-02-22 00:45:49 UTC - Ali Ahmed: doc only changes are skipped in ci ---- 2020-02-22 00:45:56 UTC - Devin G. Bost: It was a code change though. ---- 2020-02-22 00:45:59 UTC - Devin G. Bost: That’s the problem. ---- 2020-02-22 00:46:07 UTC - Devin G. Bost: I made several code changes, but it won’t build. ---- 2020-02-22 00:46:27 UTC - Ali Ahmed: whats the link ---- 2020-02-22 00:46:54 UTC - Devin G. Bost: <https://github.com/apache/pulsar/pull/6202> ---- 2020-02-22 00:48:08 UTC - Devin G. Bost: The only change I made to the workflow was changing `ci-unit.yaml` to build the surefire artifacts `always()` instead of on `failure()`, but that didn’t work either. ---- 2020-02-22 00:48:35 UTC - Devin G. Bost: (I was trying to get more info about a test that periodically hangs.) ---- 2020-02-22 00:48:47 UTC - Roman Popenov: Oh, yes, that fixed it :heart: ---- 2020-02-22 00:49:15 UTC - Devin G. Bost: So, I’m currently blocked in fixing the flaky tests that I can’t reproduce locally. ---- 2020-02-22 00:50:02 UTC - Ali Ahmed: force a change to the branch ---- 2020-02-22 00:50:02 UTC - Devin G. Bost: I also worry that if it’s something strange that happened to the git history, merging into master will propagate the problem. ---- 2020-02-22 00:50:06 UTC - Devin G. Bost: Just did. ---- 2020-02-22 00:50:21 UTC - Devin G. Bost: ---- 2020-02-22 00:50:23 UTC - Devin G. Bost: I rebased from master. ---- 2020-02-22 00:51:06 UTC - Ali Ahmed: try creating a new pr ---- 2020-02-22 00:51:18 UTC - Devin G. Bost: Good idea. ---- 2020-02-22 00:54:12 UTC - Devin G. Bost: Hmm it’s not letting me do it from that same PR. Should I try from a different branch (with different code)? ---- 2020-02-22 00:54:53 UTC - Devin G. Bost: I could branch my branch. I’ll try that. ---- 2020-02-22 00:56:21 UTC - Roman Popenov: Alright, that have solved the issue tada : Devin G. Bost ---- 2020-02-22 00:59:52 UTC - Devin G. Bost: Reproduced in new PR: <https://github.com/apache/pulsar/pull/6395> ---- 2020-02-22 01:00:04 UTC - Devin G. Bost: Look at that…. 41 successful checks??? ---- 2020-02-22 01:01:16 UTC - Devin G. Bost: The only one that’s actually building is the license check. ---- 2020-02-22 01:02:44 UTC - Devin G. Bost: I’m changing the surefire trigger back to `failure()` in the first PR as a sanity check. ---- 2020-02-22 01:03:15 UTC - Devin G. Bost: Didn’t help. ---- 2020-02-22 01:17:20 UTC - Devin G. Bost: Anyone know a Git expert? ---- 2020-02-22 01:28:23 UTC - Devin G. Bost: I created a branch from master, and I’m manually copying over the files from my first PR to the new branch, hoping it will create a clean git history. ---- 2020-02-22 01:30:07 UTC - Ali Ahmed: use cherry pick on clean master push to a new branch and create a new pr ---- 2020-02-22 01:30:36 UTC - Devin G. Bost: Thanks. I’ll try that. ---- 2020-02-22 02:02:39 UTC - tuteng: It seems that there are some problems here <https://github.com/apache/pulsar-test-infra/blob/master/diff-only/entrypoint.sh#L10> +1 : Devin G. Bost ---- 2020-02-22 02:03:15 UTC - tuteng: github actions move to `checkout@v2` ---- 2020-02-22 02:03:42 UTC - Devin G. Bost: If that’s the issue, you’re a life-saver because I was sure it was something I did, and I was working on a Git bug report haha. :sweat_smile: ---- 2020-02-22 02:04:15 UTC - Devin G. Bost: > github actions move to `checkout@v2` What do you mean by this? ---- 2020-02-22 02:05:24 UTC - Devin G. Bost: It looks like that commit was 17 days ago though. ---- 2020-02-22 02:05:44 UTC - Devin G. Bost: I’ve been seeing issues only in the last 2 days or so. ---- 2020-02-22 02:07:40 UTC - tuteng: <https://github.com/apache/pulsar/pull/6386/files> This pr move actions/checkout@v1 to actions/checkout@v2 ---- 2020-02-22 02:08:08 UTC - Devin G. Bost: Ah, good catch! ---- 2020-02-22 02:16:14 UTC - Devin G. Bost: What does that change do? ---- 2020-02-22 02:19:07 UTC - tuteng: @yijie will fix it +1 : Devin G. Bost ---- 2020-02-22 02:19:44 UTC - Devin G. Bost: You guys are awesome! ---- 2020-02-22 02:21:14 UTC - Devin G. Bost: Good thing other PRs haven’t been merged since that change or their bugs might have gone unnoticed. ---- 2020-02-22 05:25:47 UTC - Sijie Guo: FYI - the Github Action issue is now fixed. Many thanks to @yijie for fixing the problem. ---- 2020-02-22 06:01:44 UTC - Devin G. Bost: Awesome! Thanks for the update! ----
