| *** ijolliffe has quit IRC | 00:02 | |
| *** riuzen has joined #starlingx | 01:56 | |
| *** dpenney has quit IRC | 02:30 | |
| *** riuzen has quit IRC | 03:11 | |
| *** bengates has joined #starlingx | 03:48 | |
| *** bengates has quit IRC | 03:53 | |
| *** sgw has quit IRC | 04:59 | |
| *** sgw has joined #starlingx | 05:17 | |
| *** gkadam has joined #starlingx | 05:40 | |
| *** zbsarash1 has quit IRC | 06:54 | |
| *** bengates has joined #starlingx | 07:30 | |
| *** bengates_ has joined #starlingx | 07:31 | |
| *** bengates has quit IRC | 07:35 | |
| *** bengates_ has quit IRC | 07:53 | |
| *** bengates has joined #starlingx | 07:54 | |
| *** ijolliffe has joined #starlingx | 12:33 | |
| *** mthebeau has joined #starlingx | 12:51 | |
| sgw | Morning all | 13:02 |
|---|---|---|
| bwensley | morning | 13:03 |
| *** njascanu has joined #starlingx | 13:53 | |
| sgw | slittle1: https://review.opendev.org/#/q/projects:starlingx/+AND+topic:python3+(status:open+OR+status:merged) | 14:25 |
| *** mthebeau has quit IRC | 14:52 | |
| sgw | slittle1: can you ping dpenney to join | 14:52 |
| slittle1 | I probably have 1-15 min before kids demand lunch | 14:52 |
| slittle1 | pinged Don | 14:54 |
| *** dpenney has joined #starlingx | 14:54 | |
| dpenney | sorry... had to reboot for a windows update, and forgot to log abck into IRC again :) | 14:54 |
| slittle1 | ok, have python3 project up | 14:54 |
| slittle1 | sgw: what did you want to bring to our attention | 14:55 |
| slittle1 | sgw: I think you mentioned the generator change in root ? | 14:57 |
| sgw | so the question is because of the conflicting files brought in by the addition of python3-rpm-generators which conflicts with rpm package, which is why they are removing files in that specfile | 14:57 |
| sgw | So my concern in this case is if we add the lst (https://review.opendev.org/#/c/728323/) change first without the associated change to rpm (https://review.opendev.org/#/c/728326/) will we get an ISO build failure | 14:59 |
| dpenney | a case where the changes need to merge at the same time, as it's effectively moving files from one package to another... do we know the impact of having the rpm change first, dropping the files, without the new pkgs to bring the files back in? | 15:00 |
| sgw | this is an edge case because both python3-rpm-generator (binary) and our rpm (build from source) deliver the same files, there is a change in upstream Centos's rpm specfile to remove the duplicated files, which is what the https://review.opendev.org/#/c/728326/ patch provides | 15:00 |
| sgw | I hope this makes sense | 15:00 |
| slittle1 | ok, I need to poke into a few things first. As you describe it, it does appear that we need to make an exception in this case. | 15:08 |
| sgw | I just want to fully understand how .lst files affect mirror download vs ISO creation. If I understand it yes, we might need an execption which means we need to +W these with the dependency | 15:11 |
| *** mthebeau has joined #starlingx | 15:17 | |
| sgw | slittle1: dpenney: so are we good with the lst going seperate from the specfile changes for RPM? I see Don gave a +2 | 16:21 |
| dpenney | so adding the rpms to LST files won't themselves result in inclusion in the ISO. They'd have to be either in an image.inc file or pulled in via a dependency. so I think in this case, we should be safe? | 16:28 |
| dpenney | because these are new RPMs, not upversioned ones | 16:28 |
| dpenney | and the updates to use them / include them are separate | 16:28 |
| *** bengates has quit IRC | 16:30 | |
| sgw | sorry guys, I am still confused then, if we add python3-rpm-generators to the lst files only how do they get used and thus conflict with the rpm package (this is what we are trying to avoid), clearly this is happening someplace? In this case I think we need to merge https://review.opendev.org/#/c/728326/ FIRST then the lst files | 16:42 |
| sgw | thoughts? | 16:42 |
| sgw | I am trying to avoid additional build failures | 16:42 |
| *** zbsarash1 has joined #starlingx | 16:54 | |
| dpenney | RPMs are pulled into ISO by 1) direct inclusion via image.inc file, or 2) via RPM dependencies "Requires"... installed into mock env with BuildRequires... if a new package is not in an image.inc and not in a BuildRequires or Requires, it should just be "dormant" | 17:11 |
| dpenney | note that it could be a chain of dependencies pulling it in, of course... X Requires Y Requires Z... | 17:13 |
| sgw | Sounds right, but I am still concerned that might not be the case, based on what Ran has been saying and how she structued the commits and dependencies. Again, looking at the commits, I think that we need the RPM specfile change first , note it's not dependent on any other changes Then the python3-rpm-generators change | 17:13 |
| dpenney | I guess the question would be whether something is pulling in the python3-rpm-generators somehow with a dependency. If not, then we've got "rpm" missing these files, if we do the rpm change by itself. If so, then we've got a conflict if we do the LST change by itself | 17:16 |
| dpenney | I don't know what the impact of not having these files would be | 17:17 |
| sgw | I have not seen any dependency on the python3-rpm-generators so unless there is some binary rpm that I missed | 17:18 |
| dpenney | we could merge the LST file update and then kick the mirror job (which runs a few times a day anyway). Once the mirror job is successful, we can merge others | 17:22 |
| sgw | Let's do that then | 17:22 |
| dpenney | I should be able to kick the mirror job once zuul merges | 17:24 |
| sgw | Ok, I will add my +2 +W to the python3 add now | 17:25 |
| sgw | done | 17:25 |
| dpenney | do we want to merge https://review.opendev.org/728324 too? | 17:26 |
| sgw | Yes that one also | 17:26 |
| sgw | +W added | 17:26 |
| dpenney | ok, I'll watch Zuul and kick the mirror after | 17:27 |
| dpenney | https://review.opendev.org/728324 isn't getting picked up... the parent commit ID is confusing... It seems like it's pointing to the one merging now, by title, but the commit ID doesn't match | 17:35 |
| sgw | Oh come on, this is ridiculous! | 17:36 |
| sgw | Sorry frustrating day | 17:36 |
| dpenney | I can't find the review associated with the parent, or state... another abandoned parent? https://review.opendev.org/gitweb?p=starlingx%2Ftools.git;a=commitdiff;h=69ddcf15fb81f49d8cd94ba60d22afc1d10d690b | 17:37 |
| dpenney | is this like the one you had to rebase? | 17:37 |
| sgw | We can force a rebase | 17:39 |
| sgw | let me give it a go | 17:39 |
| sgw | reased | 17:39 |
| sgw | rebased | 17:39 |
| dpenney | using 69ddcf15fb81f49d8cd94ba60d22afc1d10d690b as the search in gerrit brings me to the update that just merged, even though it has a different commit ID | 17:40 |
| dpenney | fun fun | 17:40 |
| sgw | No kidding | 17:40 |
| sgw | not the way I wanted to spend my morning | 17:40 |
| dpenney | :) | 17:40 |
| dpenney | well, the rebase looks like it fixed the parent ID, and zuul is running the check, then the gate, then merged... and THEN I'll kick the mirror job | 17:41 |
| sgw | One has merged waiting for the second one | 17:49 |
| dpenney | merged... kicking the job | 17:57 |
| sgw | Coordinated effort thanks! | 17:57 |
| sgw | hope this is the exception rather than the rule! | 17:58 |
| dpenney | ok, so the mirror repo sync has been failing for a couple of weeks it seems, as http://mirror.centos.org/centos/7.7.1908/readme has been deprecated and moved to the vault | 18:03 |
| sgw | Huh? | 18:04 |
| sgw | we have had failures un-detected? | 18:05 |
| dpenney | that's one problem... syncing of repos to CENGN... it looks like it's setup to send emails to Scott on failure on that particular job | 18:05 |
| sgw | slittle1: ^^^???? | 18:05 |
| sgw | njascanu: Hi there, does the failure you reported about sanity happen with Virtual or just bare metal? | 18:41 |
| njascanu | Hi. Only with baremetal. Virtual look good. | 19:43 |
| *** njascanu has quit IRC | 19:53 | |
| *** larainema has quit IRC | 20:38 | |
| *** sgw has quit IRC | 21:00 | |
| *** sgw has joined #starlingx | 21:59 | |
| sgw | dpenney: please let me know if there is any issues, so we can all so we call allow the rest of the python3 related patches go in | 22:44 |
| dpenney | slittle1 took over updating the cengn mirror with the vault... I haven't heard an update from him | 22:46 |
| sgw | slittle1: status check? | 22:46 |
| dpenney | it looks like the repo sync job was successful | 22:47 |
| sgw | OK then I will +2 +W the rest as apporpiate, please take a look at those also! | 22:48 |
| dpenney | not sure if the mirror download did its job. it seems to succeed too quickly to have done anything | 22:48 |
| dpenney | so I think the second job, the mirror_download_sync, still has some issue | 22:48 |
| dpenney | I'll ping him on Zoom | 22:48 |
| sgw | really, then I guess we need to figure out what's going on first | 22:49 |
| *** bwensley has quit IRC | 22:58 | |
| dpenney | yeah, he needs to look at this before we go further | 23:09 |
| sgw | slittle1: dpenney: please post an update to the -discuss list so we know what's going on and the next steps. | 23:27 |
| *** ijolliffe has quit IRC | 23:56 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!