16:00:18 #startmeeting RELENG (2020-02-19) 16:00:18 Meeting started Wed Feb 19 16:00:18 2020 UTC. 16:00:18 This meeting is logged and archived in a public location. 16:00:18 The chair is mboddu. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:18 Useful Commands: #action #agreed #halp #info #idea #link #topic. 16:00:18 The meeting name has been set to 'releng_(2020-02-19)' 16:00:19 #meetingname releng 16:00:19 #chair nirik sharkcz pbrobinson pingou mboddu dustymabe ksinny jednorozec 16:00:19 #topic init process 16:00:19 The meeting name has been set to 'releng' 16:00:19 Current chairs: dustymabe jednorozec ksinny mboddu nirik pbrobinson pingou sharkcz 16:00:28 Thanks nirik for reminding it 16:00:32 .hello2 16:00:33 dustymabe: dustymabe 'Dusty Mabe' 16:00:33 hey 16:00:35 * mboddu is still in Tue mode 16:00:44 it's still monday to me. ;( 16:01:10 dustymabe: I have a question for you, but lets take it to releng 16:01:16 I mean #fedora-releng 16:01:22 So, lets get started 16:02:06 mboddu: ok 16:03:17 #topic #9266 prune EOL content from OSTree repos 16:03:23 #link https://pagure.io/releng/issue/9266 16:03:46 dustymabe: So, finally pruning stuff :) 16:04:31 mboddu: correct :) 16:04:48 mboddu: we can do this in a few different ways 16:05:05 we can either prune everything AFTER we do the migration 16:05:09 or we can do it before 16:05:10 What does "--no-prune" will give us from the first script in that ticket? 16:05:23 Better do it after migration, right? 16:05:40 mboddu: `--no-prune` is like `--dry-run` 16:05:48 Ah okay 16:05:56 mboddu: yeah maybe after migration is better 16:06:06 looks like we have 2T volume so it should all fit 16:07:13 so maybe we should talk migration strategy 16:08:31 Okay 16:08:43 what do you folks think about how we should pull off the migration? 16:08:54 dustymabe: 2T in cloudfront or somewhere else? 16:09:59 mboddu: it's a 2T netapp volume 16:10:05 I think like any other data move: 16:10:12 see https://pagure.io/releng/issue/8811#comment-602395 16:10:48 sync data, schedule time for no changes, do final sync, switch web/pointers to the new location 16:11:43 nirik: so we'd want to do it at a time where no pungi composes are happening, correct? 16:11:49 yeah 16:12:11 nirik: If we are only touching 27 and 28, we can do it at anytime 16:12:25 mboddu: we're migrating the entire repo 16:12:45 since pungi still makes silverblue, we have to care about pungi 16:12:46 Okay, then dont mind me, you guys are right 16:13:18 * mboddu thought we are only migrating eol'd content 16:13:53 nirik: so should we enumerate what needs to happen 16:13:59 1. make backup of content 16:14:11 ^^ I guess this can just be "leave the content in place" 16:14:12 backup? 16:14:18 ah yeah 16:14:39 ok. so let's do it different 16:14:49 1. make sure no writers to ostree repos are running 16:15:01 ^^ includes pungi and new-updates sync 16:15:19 2. copy content from production repos into new netapp volume 16:15:44 3. maybe mv prod repos to another directory (minimizes confusion) 16:16:19 4. set up /mnt/koji/ostree /mnt/koji/compose/ostree mounts on composer to point to new netapp volume 16:16:37 5. do we need to do anything to our webservers/cloudfront? 16:16:55 I'd add a '0. make initial copy of content' which we can do anytime and will make step 2 much quicker 16:17:15 as long as the new volumes are in the same path for webserver we should be fine for cloudfront 16:19:11 nirik: i.e. an initial rsync and then a second rsync will take less time? 16:19:20 yeah. 16:19:29 ok sounds good to me 16:19:34 anything else missing? 16:19:34 since it will only need to copy the changes the second time 16:19:52 when did we want to schedule it? 16:20:03 tomorrow? 16:20:03 On that, next week is freeze time 16:20:32 we can do the initial rsync today 16:20:40 I suppose tomorrow might work... 16:22:03 #info We will work on migration first and then pruning, nirik will start working on migration tomorrow and then mboddu will work on pruning later 16:22:38 mboddu: i should be able to handle pruning myself once we get things over into openshift 16:22:44 I can't start anything today until this afternoon. I am in meetings all morning 16:23:11 but I can look this afternoon and we can try and do the switch tomorrow after branched/rawhide finish? 16:23:30 nirik: +1 i'll grab you in 3 or 4 hours 16:23:37 if that's ok 16:24:11 more like 4-5... but feel free to ping me and I'll let you know 16:24:25 #undo 16:24:25 Removing item from minutes: INFO by mboddu at 16:22:03 : We will work on migration first and then pruning, nirik will start working on migration tomorrow and then mboddu will work on pruning later 16:24:40 #info We will work on migration first and then pruning, nirik will start working on migration tomorrow and then dustymabe will work on pruning later 16:24:47 mboddu will help as needed :) 16:26:54 +1 16:27:00 ok i gotta go run my other meeting 16:27:03 thanks mboddu and nirik 16:27:55 Thanks dustymabe 16:27:58 Okay, lets go back to old tickets 16:28:05 #topic #7400 Rethink how we handle networking config with createImage (oz / ImageFactory) 16:28:12 #link https://pagure.io/releng/issue/7400 16:30:26 * mboddu trying to find koji ticket 16:30:34 was there one? 16:31:20 I dont think so 16:31:50 #info mboddu will create a upstream koji ticket and will check with mikem 16:31:54 so we could file one I guess.. I don't have time to commit to do this one... 16:32:33 sounds good 16:32:46 nirik: Understandable 16:34:27 #topic #7555 scratch permission to build livecd for users 16:34:34 #link https://pagure.io/releng/issue/7555 16:35:16 so, not sure whats left here... 16:35:25 are our compose images still scratch? 16:35:31 nirik: Yes 16:35:49 But can we give the users 'livemedia' permissions? 16:36:19 hum, I wonder how that interacts with GC 16:37:24 fwiw (i haven't read through the entire ticket), but that change proposal was never submitted 16:38:34 so, it looks like our cloud images at least are normal... but we never move them to updates tags so they get GCed 16:38:36 Sorry, they are not running using scratch 16:39:31 and even our GA ones get GCed. ;( 16:40:02 But we import them into our compose, so..., thats still fine 16:40:23 Wed Oct 23 17:08:41 2019: Fedora-Xfce-31-1.9 tagged into f31-updates-candidate by releng 16:40:23 Sat Oct 26 12:15:33 2019: Fedora-Xfce-31-1.9 untagged from f31-updates-candidate by oscar 16:40:23 Fri Jan 10 12:24:35 2020: Fedora-Xfce-31-1.9 tagged into trashcan by oscar 16:40:23 Sat Feb 8 01:39:33 2020: Fedora-Xfce-31-1.9 untagged from trashcan by oscar 16:40:55 so, we might want to start tagging some of them to keep em... but I guess we still have them on ftp / mirror 16:41:07 Exactly 16:41:30 anyhow, I'd be open to granting some people perms to spin media... 16:41:33 bcotton: Yeah, I haven't seen the change as well 16:41:48 this might go hand in hand with the idea of having a acutal list of maintainers. 16:41:53 * kalev is interested to spin Workstation test composes. 16:42:22 * Southern_Gentlem interested 16:42:32 nirik: Its good to give them to spins/labs maintainers, but what about people who want add new media? 16:42:47 s/who want/who want to/ 16:42:51 sure, we could just also do it on a case by case basis? 16:44:16 I can't recall... can you use a local ks for a scratch livemedia build? 16:44:56 I am not sure either, but I remember giving a ks file from my fork, so that should help 16:45:59 yeah, looks like no, but I think it will accept any pagure.io git for them 16:46:26 So, how should we handle this? 16:46:30 so, whats next steps here? 16:46:31 ha. 16:47:02 Lets give it couple of folks who need it and see how they use it and get feedback and then announce it out for people who want it? 16:47:09 how about we grant Southern_Gentlem and kalev perm and ask them to try it out and report back to us if it's useful/working? 16:47:21 +! 16:47:24 nice. great minds think alike. ;) 16:47:27 Lol, exactly the same thing twice :) 16:47:27 +1 too 16:47:43 Southern_Gentlem and kalev : Are you okay with this? 16:47:47 go with kalev first because i will need to work with someone to learn what i am doing 16:48:26 mboddu: sure, works for me 16:48:53 Southern_Gentlem: Okay, but its simple 16:49:11 .fasinfo kalev 16:49:12 mboddu: User: kalev, Name: Kalev Lember, email: klember@redhat.com, Creation: 2009-02-25, IRC Nick: kalev, Timezone: Europe/Stockholm, Locale: en, GPG key ID: 8F62AF74, Status: active 16:49:13 I am wondering if the limitations will make it not so useful, but we can see. 16:49:14 mboddu: Approved Groups: cla_fedora cla_done fedorabugs provenpackager cla_fpca gitspin-kickstarts qa +packager hosted-content @gnome-sig 16:49:25 mboddu, i have been doing respins for years on my own equipment 16:49:35 you do have to pass it a install tree... so it has to use a existing one of those... 16:49:48 nirik: install tree and the repo as well 16:50:03 Southern_Gentlem: Right, it should definitely help you a lot 16:50:38 kalev: You should have the permissions now 16:50:44 nice, thanks! 16:50:45 repo for ks yeah 16:51:24 koji spin-livemedia --help 16:51:37 make sure to use --scratch :) 16:51:50 kalev: Right 16:51:55 ok, let me give it a try :) 16:53:32 mboddu: 7min left... anything else? 16:53:52 o/ 16:55:24 #topic Open Floor 16:55:38 * bcotton has a thing 16:55:38 freeze and bodhi enable is next tuesday? 16:55:45 nirik: So, I found the issue with stuck 32 compose, I will work with dustymabe on that 16:55:49 nirik: correct 16:55:50 nirik: yes 16:56:20 I have something for that. can we change the time as we discussed a bunch last cycle? pretty please? 16:58:09 nirik: ? 16:58:15 Which one are you talking about? 16:58:43 the freeze starting at 00:00 tuesday... this is confusing to everyone 16:59:06 ah yes, can you just move it one minute earlier? no real changes, just makes it much more clear when it actually starts 16:59:37 first test scratch build running: https://koji.fedoraproject.org/koji/taskinfo?taskID=41658761 16:59:41 I was hoping to move to to something like tuesday at 14UTC (or 15, whatever we normally release on) 17:00:06 nirik: Ah, right, 14UTC seems okay 17:00:09 bcotton: ^ 17:00:20 or later. or whatever. just not 00:00 17:00:43 agreed, just as long as it's not 00:00 :) 17:00:44 WFM, so long as everyone has a shared understanding 17:00:55 i can update the schedule text to say 1400 UTC 17:01:18 mboddu: can you be around at 14utc to enable bodhi? 17:01:26 it would be more consistant across the board, people know we release around 1400 17:01:48 nirik: Yes, I would be around 17:02:02 if everyone is ok with that, perhaps a email this week with a reminder/note about it to devel-announce would be good. 17:02:12 and we should do the same for final, etc. 17:02:36 nirik++ 17:02:42 so for final that would mean a updates push at 14 and then freezing. 17:03:24 #link https://pagure.io/fedora-project-schedule/issue/175 17:03:25 Right 17:03:33 Thanks bcotton 17:03:45 So, lets close this for today 17:03:49 * mboddu got to run 17:03:54 Thanks everyone for joining 17:03:55 i'll try to get the schedule updated by EOB tomorrow so that you can point to it when sending the remidner 17:04:12 bcotton: Okay, please ping me once its done 17:04:12 bcotton++ 17:04:12 nirik: Karma for bcotton changed to 12 (for the current release cycle): https://badges.fedoraproject.org/tags/cookie/any 17:04:19 mboddu: ack 17:04:20 thanks mboddu 17:04:25 bcotton++ nirik++ kalev++ Southern_Gentlem++ 17:04:26 mboddu: Karma for bcotton changed to 13 (for the current release cycle): https://badges.fedoraproject.org/tags/cookie/any 17:04:29 mboddu: Karma for jbwillia changed to 2 (for the current release cycle): https://badges.fedoraproject.org/tags/cookie/any 17:04:39 #endmeeting