20:00:06 #startmeeting EPEL (2022-05-04) 20:00:06 Meeting started Wed May 4 20:00:06 2022 UTC. 20:00:06 This meeting is logged and archived in a public location. 20:00:06 The chair is tdawson. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions. 20:00:06 Useful Commands: #action #agreed #halp #info #idea #link #topic. 20:00:06 The meeting name has been set to 'epel_(2022-05-04)' 20:00:06 #meetingname epel 20:00:07 The meeting name has been set to 'epel' 20:00:07 #chair nirik tdawson pgreco carlwgeorge salimma dcavalca 20:00:07 #topic aloha 20:00:07 Current chairs: carlwgeorge dcavalca nirik pgreco salimma tdawson 20:00:27 .hi 20:00:28 carlwgeorge: carlwgeorge 'Carl George' 20:00:35 Hi carlwgeorge 20:00:43 EHLO 20:00:47 morning 20:00:51 .hi 20:00:52 salimma: salimma 'Michel Alexandre Salim' 20:01:06 Hi SSmoogen 20:01:11 Hi nirik 20:01:18 Hi salimma 20:01:19 .hello robert 20:01:20 rsc: robert 'Robert Scheck' 20:01:22 .hi 20:01:23 davide: Sorry, but user 'davide' does not exist 20:01:31 Hi rsc 20:01:40 Hi user that does not exists 20:01:42 hi! 20:01:50 Oh, sorry, Hi davide :) 20:02:01 Hi dherrera 20:02:13 One day I'll figure out how to fix my account so this works properly with the matrix nick :) 20:02:56 .hi 20:02:57 pgreco: pgreco 'Pablo Sebastian Greco' 20:03:01 Hi pgreco 20:04:19 davide: I just follow kparal's blog post to the letter :p 20:04:46 #topic EPEL Issues https://pagure.io/epel/issues 20:04:47 https://pagure.io/epel/issues?tags=meeting&status=Open 20:05:30 The only thing there is the CVE, and we still have several weeks before that ... moving on. 20:05:41 #topic Old Business 20:06:30 The only thing I have for old business is sending email on retirement ... and I believe we are waiting on a doc pull request for that. 20:07:24 Oh, I also have epel7-testing cleanup ... but I haven't had any time to do any of that. 20:07:43 Is there any other old business that I don't have? 20:08:07 This might be one of those fast meetings I've hear about. 20:08:17 tdawson: don't jinx it :) 20:08:23 #topic EPEL-7 20:08:23 CentOS 7 will go EOL on 2024-06-30 20:08:37 Anything for epel7? 20:09:03 I've seen a couple of epel7 repo issues today ... that's a bit odd. 20:09:05 fun aside: should we do retirement parties the way we do release parties? 20:09:30 I know many people that partied when 6 was retiredd 20:09:32 *retired 20:09:32 Like, when epel7 goes away, have a retirement part? 20:09:40 I'm in favor of more parties 20:09:40 the end of the pre-systemd era 20:10:33 That would be interesting. we could gather all the fun/odd things we remember about epel7 20:11:15 there has been a problem with epel7 mirorrs 20:11:43 I think it is something with the quick-mirror script as when running it, it says everything is fine and nothing to do in epel-7 20:11:50 but there are a lot of files which aren't fine 20:12:15 the ticket on it is https://pagure.io/releng/issue/10775 but it may need to move over to fedora-infra 20:12:19 Joking aside, it'd be great to do a retrospective on the previous release on retirement 20:12:38 What went well, lessons learned, etc 20:12:54 I think 7 started an era, and nobody will miss 8 20:12:55 With a strong drink in hand if that's easier of course 20:13:14 9 is already looking better 20:13:35 SSmoogen: Is it only hapening on epel7 repo? 20:13:49 I found problems with epel9 also 20:14:03 * nirik is sucked into a doom with a server. 20:14:10 davide: very good ideas, I've written them in my notes. 20:14:11 I only did /pub/epel so I am not sure if fedora is affected. 20:14:16 one mirrors issues seems beyond the scope of this meeting tho 20:14:42 mirrors.kernel.org and download-ib01 were reported to have this issue 20:15:06 but agreed. one fire at a time 20:15:24 davide: yup 20:15:33 sort of retirement party + eulogy 20:15:41 "it was a good release, but..." 20:16:45 Hopefully I remember to send out emails a couple months before epel7 goes away. But if I forget, someone else remember a few months to send something out for epel7. 20:16:57 I'm going to move on ... 20:17:03 #topic EPEL-8 20:17:04 CentOS Stream 8 goes EOL in 2024-05-31 20:17:30 Just a reminder that RHEL 8.6 comes out "Sometime in May" 20:18:32 Other than the KDE / kernel wierdness, I don't think there are any/many things affected. 20:19:14 Is there anything else to talk about for epel8 ? 20:19:44 not from me 20:20:03 .hello ngompa 20:20:04 Eighth_Doctor: ngompa 'Neal Gompa' 20:20:13 Hi Eighth_Doctor 20:20:16 mentioning KDE summons Neal :) 20:20:23 :) 20:20:43 do we have KDE Plasma 5.24 in EPEL8? 20:20:43 Quick ... I'll change the topic ... maybe he won't see. ;) 20:21:11 tdawson: too late 20:21:14 :) 20:21:32 Eighth_Doctor: No, epel8 is currently at 5.23.3 20:22:07 5.24.4 requires a newer wayland-protocols (which I think I can get around) and a new libinput (which I might be able to work around, maybe) 20:23:07 I'm sorta hoping that 5.24.4 will fix the screen locking issue. 20:23:51 I've already tested, and plasma-5.23.5 doesn't fix the screen locking issue (or the ssdm issue) 20:24:26 Anything else for epel8 ? 20:25:02 not from me again 20:25:07 #topic EPEL-9 20:25:21 anything for epel9 ? 20:25:23 will RHEL 9 be released concurrently with 8.6? or separately 20:25:28 no idea 20:26:01 if I had to bet, I'd say no 20:26:06 I've been told that for technical reasons they can't be done on the same day 20:26:11 2 consecutive weeks 20:26:23 but I expect they won't be far apart 20:26:25 carlwgeorge: oh interesting 20:26:34 for epel9 there was an email for overlap in EL9 packages and EPEL 20:26:39 I wonder what that is, but understand we can't know 20:26:57 * salimma thinking of how "for technical reasons" Alaska Airlines' schedule blew up every beginning of month 20:27:00 so when you put meat into the grinder.. you have to put one at a time 20:27:11 https://lists.fedoraproject.org/archives/list/epel-devel@lists.fedoraproject.org/thread/O5M7WH5XODHBXSVY3VJXJ6TNK44JTZ2O/ 20:27:34 ^ what SSmoogen was referring to 20:27:43 the first time, and not about the grinder 20:28:19 you can only fill one taco shell at a time? 20:29:42 to be fair if I'm the release manager and I have to release both on the same day I'll probably scream too 20:29:44 sorry I derailed myself 20:29:45 So, I did a quick check, and all of those packages (libwmf pybind11 tesseract and tesseract tesseract-tessdata) were requested through the normal way. 20:29:51 I'm looking into the pybind11 one, no idea yet how that branch request happened 20:30:23 pybind11 - https://pagure.io/releng/fedora-scm-requests/issue/42248 20:30:27 so I am guessing they all went through BEFORE the package was in a place where the pdc? thing could say 'nope sorry' 20:31:18 When did the pdc thing get put in place? 20:31:19 with things moving around before the final release of 9.0, there have been all kinds of 'buildroot-only' to 'crb' to 'appstream' and back again 20:31:29 so I am not sure about pdc being the tool 20:31:42 I think the check has been there since it was allowed to make branches... 20:31:48 but it checks against centos stream 20:31:58 there is something that the releng person checks which looks to see if the package is in a json which is generated about what is currently in RHEL 20:32:31 or in this case CentOS Stream 20:32:53 https://composes.stream.centos.org/production/latest-CentOS-Stream/compose/metadata/rpms.json 20:33:01 I will let nirik take over on how that works because my knowledge is dead 20:33:38 I think carlwgeorge wrote it, so I would defer to him. ;) 20:33:48 you know... huh... 20:34:18 theory: all those packages have a Epoch... does that confuse the logic in the script? 20:34:57 So, I just checked on libwmf, that was only built in CentOS STream, just a little bit after the branch was created. So, that one was a race condition that epel won on. 20:34:59 ie, tesseract-tessdata-0:4.1.0-3.el9.src the 0: might be throwing it for a loop? 20:35:17 Michel Alexandre Salim 🎩: I'll bet that they'll be announce released on the same day 20:35:48 but they'll drip the bits out in stages 20:37:18 I don't think an epoch would do it, but I'm still investigating 20:37:48 my theory is the person processing the branch requests has an outdated fedscm-admin tool 20:38:03 thats possible sure. 20:38:08 Oh ... that's possible. 20:38:23 oh wait no, that would fail for pkg_el9.json not existing 20:38:26 I wonder if we could add a check to the top of the tool to check it's version against the latest... 20:38:46 of course that will go away when automation lands. 20:38:49 the code in question https://pagure.io/fedscm-admin/blob/main/f/fedscm_admin/utils.py#_762-778 20:40:43 more investigation is needed, we can move on for this meeting 20:40:48 OK, moving on. 20:41:00 #topic General Issues / Open Floor 20:41:03 so one of the fragile things we dealt with on flag days was stuff moving around in EL-X releases. We don't really check to see if something has a conflict after it was ok unless told about it 20:41:51 I was wondering if something like what miro ran to find the current conflicts could be run daily between epel and centos stream to alert us sooner 20:42:14 I bet it could 20:43:14 i've done it before by hand for epel8, and had a to-do item to repeat the process (which i've forgotten) on epel7 20:43:32 anyway that was my only constructive thing I could add 20:43:42 i can probably implement a script to do this, but i don't know how to make it a toddler as you suggested in the email thread 20:44:06 Anything else for open floor? 20:44:39 carlwgeorge, I would talk to pingou about that 20:44:51 no i do not have anything for open floor 20:45:12 Is everyone ok leaving 10-15 minutes early? 20:45:25 sure. 20:45:42 Anyone beside SSmoogen? 20:45:46 all good 20:45:46 nothing else from me either 20:45:53 I'll take my 15 minutes 20:45:55 Cool 20:46:09 thanks tdawson and everyone 20:46:13 Thank you all for coming. The discussions were short, but good. 20:46:15 thanks all 20:46:18 thanks guys, until next week! 20:46:19 thanks 20:46:20 I'll talk to you all next week, if not soone. 20:46:22 thanks! 20:46:28 #endmeeting