| rdogerrit | Joel Capitao proposed openstack/tobiko-distgit antelope-rdo: Dummy commit to rebuild https://review.rdoproject.org/r/c/openstack/tobiko-distgit/+/55658 | 07:48 |
|---|---|---|
| rdogerrit | Merged openstack/tobiko-distgit antelope-rdo: Dummy commit to rebuild https://review.rdoproject.org/r/c/openstack/tobiko-distgit/+/55658 | 08:03 |
| jcapitao[m] | karolinku: there is something wrong in Antelope Trunk builder | 09:03 |
| jcapitao[m] | it does not trigger build with latest distgit commit | 09:03 |
| jcapitao[m] | I'm checking why | 09:03 |
| karolinku[m] | is cron enabled? :) | 09:04 |
| karolinku[m] | (check if device is connected to power supply) | 09:05 |
| jcapitao[m] | yep | 09:05 |
| jcapitao[m] | I checked in database | 09:05 |
| jcapitao[m] | it's stucked to commit before the dummy one | 09:05 |
| karolinku[m] | I think we already had that issue, restart was needed | 09:06 |
| jcapitao[m] | I have a local reproducer | 09:06 |
| apevec | can you strace the process? | 09:06 |
| jcapitao[m] | so there is something wrong in DLRN code or Tobiko | 09:07 |
| apevec | ah you can actually reproduce at will? | 09:07 |
| jcapitao[m] | yes, locally I have the same behavior | 09:07 |
| karolinku[m] | what is local error? | 09:14 |
| karolinku[m] | or it just hangs during dlrn_run? | 09:14 |
| jcapitao[m] | No commits to build | 09:15 |
| jcapitao[m] | with dlrn --head-only --verbose-build --package-name openstack-tobiko | 09:15 |
| karolinku[m] | maybe commit message is not enough to trigger rebuild? | 09:17 |
| karolinku[m] | in https://review.rdoproject.org/r/c/openstack/tobiko-distgit/+/55658 | 09:17 |
| jcapitao[m] | DLRN trigger build for each new combination of source commit and distgit commit | 09:18 |
| jcapitao[m] | c.f https://github.com/softwarefactory-project/DLRN/blob/master/dlrn/drivers/rdoinfo.py#L198-L206 | 09:19 |
| karolinku[m] | for Y and Z you did with spec modification | 09:20 |
| jcapitao[m] | the dummy commit you mean ? | 09:23 |
| karolinku[m] | yes | 09:23 |
| jcapitao[m] | for Antelope as well | 09:24 |
| jcapitao[m] | that's the issue | 09:24 |
| jcapitao[m] | it triggered build for Y and Z, but not A | 09:24 |
| karolinku[m] | the weird think in gerrit it that https://review.rdoproject.org/r/c/openstack/tobiko-distgit/+/55633/2 shows that only cm was modified, which is not true | 09:26 |
| karolinku[m] | is the rebase in PS 2 just canceled .spec modification? | 09:27 |
| jcapitao[m] | DLRN is using GH distgit repo https://github.com/rdo-packages/tobiko-distgit/tree/antelope-rdo | 09:28 |
| karolinku[m] | same story https://github.com/rdo-packages/tobiko-distgit/commit/0d91f7b5c53be5da0e422d54c60a13164e5d361c, it's empty | 09:29 |
| jcapitao[m] | interesting | 09:36 |
| jcapitao[m] | but the content is still there | 09:36 |
| jcapitao[m] | and there is a new commit hash | 09:36 |
| jcapitao[m] | but it might be related | 09:36 |
| jcapitao[m] | let's test with new patch | 09:37 |
| karolinku[m] | agree, content + new hash should be enough, but I still have feeling that we once discuss it. amoralej would remember! | 09:38 |
| amoralej | about what? | 09:38 |
| karolinku[m] | dummy commits with just CM moidifed | 09:39 |
| amoralej | no, it does not work | 09:39 |
| karolinku[m] | :D | 09:39 |
| amoralej | there is something in DLRN which ignore those empty commits | 09:39 |
| jcapitao[m] | that's interesting | 09:40 |
| jcapitao[m] | so it does not check only commit hash | 09:40 |
| jcapitao[m] | I mean distro hash + source hash | 09:40 |
| rdogerrit | Joel Capitao proposed openstack/tobiko-distgit antelope-rdo: Dummy commit to rebuild https://review.rdoproject.org/r/c/openstack/tobiko-distgit/+/55659 | 09:43 |
| amoralej | lemme see | 09:53 |
| amoralej | there is a git command inside dlrn which is what makes it to ignore them | 09:53 |
| jcapitao[m] | I'm checking the DLRN code | 09:54 |
| jcapitao[m] | I think it's in https://github.com/softwarefactory-project/DLRN/blob/master/dlrn/repositories.py | 09:54 |
| amoralej | or in https://github.com/softwarefactory-project/DLRN/blob/master/dlrn/drivers/rdoinfo.py ? | 09:57 |
| jcapitao[m] | I'm tracing the process | 09:58 |
| rdogerrit | Merged openstack/tobiko-distgit antelope-rdo: Dummy commit to rebuild https://review.rdoproject.org/r/c/openstack/tobiko-distgit/+/55659 | 09:58 |
| jcapitao[m] | amoralej: in rdoinfo drive, it exits at https://github.com/softwarefactory-project/DLRN/blob/master/dlrn/drivers/rdoinfo.py#L135 | 09:58 |
| jcapitao[m] | refreshrepo raises an exception | 09:59 |
| amoralej | but note that the list of commits is taken from https://github.com/softwarefactory-project/DLRN/blob/master/dlrn/drivers/rdoinfo.py#L174-L191 iiuc | 10:01 |
| amoralej | it may be ignoring empty ones not in the refresh but when creating the list of commits to process | 10:01 |
| amoralej | ah, i remember, i think | 10:03 |
| amoralej | there was a point '.' | 10:03 |
| amoralej | that was the key | 10:03 |
| amoralej | somewhere | 10:03 |
| amoralej | if you do "git log ." it does not show empty commits | 10:04 |
| amoralej | i think that was the point | 10:04 |
| jcapitao[m] | lemme check | 10:05 |
| karolinku[m] | what a trick o_0 | 10:05 |
| jcapitao[m] | good point | 10:05 |
| jcapitao[m] | that's the issue | 10:05 |
| jcapitao[m] | I see it show a commit but not the latest one | 10:05 |
| jcapitao[m] | actually it shows the one which is in DB | 10:06 |
| amoralej | yep | 10:06 |
| jcapitao[m] | https://github.com/softwarefactory-project/DLRN/blob/7d5cbe815db23911c432d131297bdf8c41a9cb63/dlrn/repositories.py#L149 | 10:06 |
| amoralej | yep | 10:07 |
| jcapitao[m] | thank you ! | 10:07 |
| jcapitao[m] | lemme propose a change for that | 10:07 |
| amoralej | the goal is just to be able to create empty dummy commits ? | 10:07 |
| jcapitao[m] | I'll replace . by HEAD | 10:08 |
| jcapitao[m] | which is the right pointer | 10:08 |
| jcapitao[m] | as we checkout to the right branch before | 10:08 |
| jcapitao[m] | we should git log HEAD | 10:08 |
| jcapitao[m] | amoralej: I'd say not only, but to be sure we build from commit we checkout to | 10:09 |
| jcapitao[m] | the problem here is that we checkout to the right ref (branch in that case), but we don't build from latest commit | 10:10 |
| amoralej | what happens with merge commits ? | 10:10 |
| amoralej | how are we managing those ? | 10:10 |
| amoralej | are merge commits skipped with that? | 10:11 |
| amoralej | that may be the reason to have that . | 10:11 |
| jcapitao[m] | lemme check | 10:11 |
| jcapitao[m] | amoralej: we are build the "Merge" commit hash | 10:13 |
| jcapitao[m] | latest nova https://trunk.rdoproject.org/centos9-master/component/compute/21/e6/21e6a54a81a84f02a2dfe7854824e4adfa019f49_c39e5659/commit.yaml | 10:13 |
| jcapitao[m] | commit_hash: 21e6a54a81a84f02a2dfe7854824e4adfa019f49 | 10:13 |
| jcapitao[m] | which is a merge commit https://github.com/openstack/nova/commit/21e6a54a81a84f02a2dfe7854824e4adfa019f49 | 10:13 |
| amoralej | so we merged both the real commit and the merge commit ? | 10:15 |
| amoralej | we also built 8ab51e4422f1fffd272f153e4f2ed4aba469b178 ? | 10:15 |
| amoralej | so apparently merge commits "see" the parent commit content | 10:15 |
| jcapitao[m] | that's eactly what I'm checking in DB | 10:15 |
| amoralej | then i don't see any other collateral damage | 10:16 |
| jcapitao[m] | yes we are building both real and merge commit | 10:18 |
| jcapitao[m] | https://trunk.rdoproject.org/centos9-master/component/compute/8a/0f/8a0fa58d8c2ecaa141900a25a763f55fc07ecc98_c39e5659/ | 10:18 |
| jcapitao[m] | which is the real commit | 10:18 |
| jcapitao[m] | I'm preparing a patch | 10:22 |
| apevec | hmm, that feels like wasting resources but not sure how to avoid it | 10:42 |
| jcapitao[m] | yeah, I'm not sure too | 10:51 |
| jcapitao[m] | because it depends on how the remote repo configures the merging | 10:52 |
| rdogerrit | Matthieu Huin proposed config master: ci-config: disable default pipeline https://review.rdoproject.org/r/c/config/+/55660 | 10:53 |
| rdogerrit | Merged config master: ci-config: disable default pipeline https://review.rdoproject.org/r/c/config/+/55660 | 11:06 |
| rdogerrit | Bogdan Dobrelya proposed rdo-jobs master: Add a multi-cell non-voting job https://review.rdoproject.org/r/c/rdo-jobs/+/53192 | 11:57 |
| rdogerrit | Yatin Karel proposed rdo-jobs master: [multinode adoption] Pass zuul-params.yml to test operator https://review.rdoproject.org/r/c/rdo-jobs/+/55661 | 12:01 |
| rdogerrit | Fiorella Yanac proposed rdo-jobs master: Add controllers as networkers nodes https://review.rdoproject.org/r/c/rdo-jobs/+/55655 | 13:05 |
| rdogerrit | Merged rdo-infra/ci-config master: Disable project pipelines https://review.rdoproject.org/r/c/rdo-infra/ci-config/+/55627 | 13:34 |
| rdogerrit | Joel Capitao proposed config master: Remove hguemar URI from reviewers https://review.rdoproject.org/r/c/config/+/55646 | 13:48 |
| rdogerrit | Joel Capitao proposed config master: Remove hguemar URI from reviewers https://review.rdoproject.org/r/c/config/+/55646 | 13:52 |
| rdogerrit | Joel Capitao proposed config master: Remove hguemar URI from reviewers https://review.rdoproject.org/r/c/config/+/55646 | 14:23 |
| rdogerrit | Joel Capitao proposed config master: Remove hguemar URI from reviewers https://review.rdoproject.org/r/c/config/+/55646 | 14:28 |
| rdogerrit | Joel Capitao proposed config master: Remove hguemar URI from reviewers https://review.rdoproject.org/r/c/config/+/55646 | 14:30 |
| rdogerrit | Merged rdo-jobs master: [multinode adoption] Pass zuul-params.yml to test operator https://review.rdoproject.org/r/c/rdo-jobs/+/55661 | 16:27 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!