14:37:49 <dgilmore> #startmeeting RELENG (2016-10-24)
14:37:49 <zodbot> Meeting started Mon Oct 24 14:37:49 2016 UTC.  The chair is dgilmore. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:37:49 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
14:37:49 <zodbot> The meeting name has been set to 'releng_(2016-10-24)'
14:37:49 <dgilmore> #meetingname releng
14:37:49 <zodbot> The meeting name has been set to 'releng'
14:37:49 <dgilmore> #chair dgilmore nirik tyll sharkcz bochecha masta pbrobinson pingou maxamillion mboddu
14:37:49 <zodbot> Current chairs: bochecha dgilmore masta maxamillion mboddu nirik pbrobinson pingou sharkcz tyll
14:37:52 <dgilmore> #topic init process
14:37:53 <maxamillion> .hello
14:37:55 <zodbot> maxamillion: (hello <an alias, 1 argument>) -- Alias for "hellomynameis $1".
14:37:58 <maxamillion> .hello maxamillion
14:37:59 <zodbot> maxamillion: maxamillion 'Adam Miller' <maxamillion@gmail.com>
14:38:09 <nirik> .hello kevin
14:38:10 <zodbot> nirik: kevin 'Kevin Fenzi' <kevin@scrye.com>
14:39:59 <mboddu> .hello mohanboddu
14:40:00 <zodbot> mboddu: mohanboddu 'Mohan Boddu' <mboddu@bhujji.com>
14:40:02 <sharkcz> .hello sharkcz
14:40:03 <zodbot> sharkcz: sharkcz 'Dan Horák' <dan@danny.cz>
14:40:27 <sharkcz> I need to leave in a while ...
14:42:22 <dgilmore> sharkcz: okay
14:42:29 <dgilmore> lets do alt arches first
14:42:34 <dgilmore> #topic Alternative Architectures updates
14:43:52 <dgilmore> sharkcz: how is everything?
14:44:02 <dgilmore> pbrobinson: are you here?
14:44:03 <sharkcz> for ppc and s390 all looks good, I'm solving the python-matplotlib right now as a preparation for the ppc koji merger
14:44:11 <dgilmore> okay
14:45:22 <pbrobinson> dgilmore: yep
14:45:26 <pbrobinson> sorry, was making tea
14:45:42 <pbrobinson> yes, all looking good as sharkcz reported
14:45:42 <dgilmore> pbrobinson: :)
14:45:50 <nirik> cool
14:46:11 <pbrobinson> dgilmore: I had cake too.... very exciting! (my life is so boring!)
14:46:13 <dgilmore> pbrobinson: did you see the atomic host commit i sent in on Friday?
14:46:23 <dgilmore> pbrobinson: I could go cake and coffee
14:46:28 <pbrobinson> dgilmore: maybe? I don't remember it
14:46:35 <dgilmore> there is a bakery opening up the street soon, it may be very bad
14:46:51 <pbrobinson> what did I screw up? It's on list to look further today
14:47:00 <dgilmore> https://pagure.io/fedora-atomic/c/37389c874240db4055c8076600a89a5fa3eba4a7?branch=master
14:47:21 <dgilmore> #info changes were made for atomic host in rawhide to support all arches other than s390x and i686
14:47:29 <pbrobinson> dgilmore: master? can we have f25 too?
14:48:07 <dgilmore> pbrobinson: we can, I checked with walters after doing it that the rpm-ostree in f25 had the necessary support
14:48:36 <dgilmore> I believe that the packag lists are correct
14:48:40 <nirik> dgilmore: I saw a failure related to that...
14:48:47 <dgilmore> nirik: ?
14:48:49 <nirik> there's no 'grub' package on x86_64 anymore...
14:48:53 <pbrobinson> looks to be the same version for f24+
14:48:55 <nirik> it's grub2
14:49:01 <dgilmore> nirik: gahh yeah
14:49:07 <dgilmore> typo
14:49:24 <nirik> easy to fix. ;)
14:49:50 <dgilmore> the very last build of rpm-ostree added the ability to specify packages-<basearch>
14:49:56 <dgilmore> #info the very last build of rpm-ostree added the ability to specify packages-<basearch>
14:50:26 <bowlofeggs> .hello
14:50:26 <zodbot> bowlofeggs: (hello <an alias, 1 argument>) -- Alias for "hellomynameis $1".
14:50:28 <bowlofeggs> (late)
14:50:34 <bowlofeggs> .hello bowlofeggs
14:50:37 <zodbot> bowlofeggs: bowlofeggs 'Randy Barlow' <randy@electronsweatshop.com>
14:51:13 <dgilmore> #info we should enable atomic host in rawhide for aarch64 and when we merge ppc64 and ppc64le those as well
14:51:31 <dgilmore> pbrobinson: so are we in a place to merge ppc64 and ppc64le this week?
14:51:48 <dgilmore> we should announce on devel list and the ppc list
14:51:53 <pbrobinson> dgilmore: yep, I think so, was going to sync with you later on that for timing etc
14:52:45 <dgilmore> #info we will announce to devel and ppc lists the migration of ppc64 and ppc64le from ppc.koji.fp.o to koji.fp.o this weekend
14:53:27 <dgilmore> #info s390x to be evaluated for f27
14:53:36 * dgilmore would really like to get s390x done also
14:53:51 <dgilmore> but that is more of a political thing than technical
14:54:02 <pbrobinson> dgilmore: yes, completely
14:54:04 <sharkcz> ack
14:55:02 <dgilmore> anything else to cover on Alternative arches?
14:55:26 <bowlofeggs> yeah, has pbrobinson gotten fedora running on his brain's arch yet?
14:55:46 <bowlofeggs> and is that little or big endian?
14:56:05 <dgilmore> pbrobinson: do you think we should add a iot secotion? or just evaluate down the road?
14:56:14 <pbrobinson> bowlofeggs: not sure linux could cope with that.... I think it we evict itself ;-)
14:56:16 <dgilmore> bowlofeggs: it is endian agnostic
14:56:22 <bowlofeggs> hahaha
14:56:35 <pbrobinson> dgilmore: evaluate later on in F-26
14:56:44 <dgilmore> pbrobinson: cool
14:56:57 * dgilmore wants to ensure its all a first class citizen
14:56:59 <pbrobinson> dgilmore: I'm thinking we'll run it similar to two week atomic, likely as a monthly release
14:57:06 <dgilmore> okay
14:57:59 <dgilmore> lets move on
14:58:07 <dgilmore> #topic pagure migration
14:58:14 <dgilmore> nirik: you said you had an update?
14:58:15 <nirik> so, I have some news here...
14:58:27 <nirik> yeah. So we have already got pagure projects with issues in them...
14:58:46 <nirik> they added a option to the importer to work around this, by adding an offset.
14:59:12 <nirik> So, if you have a pagure project with 10 issues in it, you set an offset to +10 and import and all the imported tickets will be 10 higher
14:59:21 <nirik> so, all the numbers will change, but the data will be there.
14:59:36 <nirik> is that going to work?
15:00:15 <nirik> It might be nice if we could make the hosted ones not change, but I think thats a lot more work...
15:01:39 <nirik> thoughts?
15:01:57 <dgilmore> nirik: going to have to
15:02:12 <dgilmore> its that or we ask people to refile open issues
15:02:20 <dgilmore> and lose history completely
15:02:25 * nirik nods.
15:02:44 <nirik> if thats going to be ok, I can look at moving stuff in the next few days hopefully.
15:02:57 <dgilmore> I do not see any other option
15:03:32 <nirik> ok. can do then...
15:03:42 <dgilmore> cheers
15:08:44 <dgilmore> anyone have any questions? or input?
15:09:08 <maxamillion> eh, I'm not the biggest fan of losing numerical context but it kind of is what it is
15:09:14 <dgilmore> #info we will import the fedorahosted issues with new issue numbers due to existing issues in pagure
15:09:35 <dgilmore> maxamillion: I guess the other option owuld be dump the pagure issues and add them after
15:09:41 <nirik> yeah, people can search for old stuff at least
15:09:41 <dgilmore> but then they change number
15:09:58 <maxamillion> dgilmore: yeah :/
15:10:02 <maxamillion> I'm not sure what's better
15:11:31 <dgilmore> less than ideal, meet less than ideal
15:12:32 <bowlofeggs> haha
15:12:34 <dgilmore> #topic f26 changes
15:12:34 <maxamillion> +1
15:12:39 <dgilmore> https://fedoraproject.org/wiki/ReleaseEngineering/PriorityPipeline#F26_Planned_Tools_Changes
15:12:41 <bowlofeggs> well you could just only change the ones that there are less of
15:12:43 <bowlofeggs> ?
15:13:01 <bowlofeggs> i.e., if there are less "new ones" you could change those in favor of letting the old ones keep their numbers
15:13:02 <dgilmore> so that is what we have in the priority piepline
15:13:09 <dgilmore> bowlofeggs: not sure thats possible
15:13:22 <dgilmore> but that may cause different confusion
15:13:32 <dgilmore> though the effected ones would be ancient
15:13:33 <nirik> I could ask if we wanted, but yeah, might be more difficult code wise
15:13:36 <bowlofeggs> yeah it'll be confusing either way
15:13:50 * nirik thinks we just bite the bullet. ;)
15:13:59 <dgilmore> nirik: same
15:14:13 <maxamillion> +1
15:14:16 <maxamillion> probably best
15:15:11 <dgilmore> does anyone have any questions on the changes?
15:15:30 <dgilmore> #info there will be a flag day for packager changes
15:16:33 <maxamillion> not I
15:16:36 <dgilmore> we will make the change to sha256sum for lookaside cache, force the migration to kerberos for koji auth, change the ssl certs for koji.fp.o
15:16:37 <nirik> we should start deciding when the day should be... to give lots of notice and also to make sure we don't push it out more. ;)
15:16:50 <dgilmore> there was one other change I can not remeber right now that needed it also
15:16:58 <dgilmore> nirik: right
15:17:06 <dgilmore> which is why I am bringing it up now
15:17:13 <nirik> oh, the rawhide sign/tag thing (but that doesn't need flag day I don't thiink)
15:17:23 <dgilmore> oh yeah
15:17:37 <dgilmore> the fedpkg update so we can gate rawhide builds on being signed
15:17:44 <dgilmore> nirik: it does
15:17:48 <nirik> so after f25 release... but before holidays
15:18:01 <nirik> oh? ok
15:18:19 <dgilmore> everyone has to have a new enough fedpkg to support the f26-candidate tag as the destination
15:18:26 <nirik> How about december 6th?
15:18:41 <dgilmore> I would say the week after.
15:18:46 <nirik> thats 3 weeks after f25 (unless we slip)
15:18:52 <dgilmore> I am going to be on PTO for the two weeks before that
15:18:57 <nirik> sure, wfm
15:19:06 <nirik> dec 13th?
15:19:28 <nirik> dunno if it has to be on a tuesday...
15:19:58 <dgilmore> nirik: I would nearly say the 12th
15:20:06 <dgilmore> we can make the changes over the weekend
15:20:16 <dgilmore> then when people come to work monday its changed
15:20:20 <nirik> ok
15:21:02 <dgilmore> most of it all except for the ssl auth changes can be done pretty transparently
15:21:43 <dgilmore> #info weekend of December 10th and 11th all the changes for packager flag day will be finalised
15:22:00 <nirik> so we need to push updates right before that? or everything can be in soon / pushed and thats just the server end?
15:22:26 <dgilmore> #info on December 12th all packagers will need to be using new enough versions of rpkg, koji, fedpkg, fedora-packager
15:22:43 <dgilmore> nirik: everything but the server changes can go out now
15:22:52 <dgilmore> its going to be some server changes
15:23:14 <dgilmore> we will need to identify the versions of tooling needed by end users
15:23:19 <dgilmore> and advertise them
15:23:46 <nirik> yep.
15:23:48 <dgilmore> nirik: fedpkg for a long time has supported sha256sum
15:23:54 * nirik nods.
15:23:57 <dgilmore> we just need to change a flag to default to it
15:24:21 <dgilmore> there may be some updates that need to go out around that time
15:24:32 <dgilmore> but I think if they go out a bit early it will not hurt
15:24:35 <nirik> I'm not fully sure about the kerb switch... I guess we invalidate/revoke certs then and it should hopefully fall back to getting a ticket. we will need to go over that at some point
15:24:56 <dgilmore> nirik: well we should push out new koji configs that default to kerb
15:25:09 <dgilmore> then users will not be able to use ssl certs
15:25:15 <nirik> yeah, then that will need to be done right around the flag day
15:25:23 <dgilmore> at the flag day ssl cert auth will be switched off on the hub
15:25:26 <nirik> or I guess a bit before since it won't hurt
15:26:32 <dgilmore> #info change from ssl to kerberos will be the most distruptive end user change
15:27:15 <dgilmore> #info any changes we want to get in for Fedora 26 we should have listed now
15:27:54 <dgilmore> #topic Open Floor
15:28:03 <dgilmore> any one have anything to bring up?
15:28:15 <nirik> Not sure if it should be listed, but we should figure out how we want to allocate the new buildhw's... as buildhw's or as buildvmhosts.
15:28:41 <nirik> I had 2 quick FYI items beyond that...
15:28:41 <dgilmore> I thought we already said we would do it all as buildvmś
15:28:57 <dgilmore> and switch some to buildhw is the image building was too slow
15:28:58 <nirik> I wasn't sure if we tested that everything works...
15:29:35 <nirik> ok, can look into that. Might need to shuffle around storage...
15:29:37 <dgilmore> not 100% sure
15:30:29 <nirik> The other things: 1) I updated buildvm/buildhw/arm01/arm02 (and soon arm04 if this glibc build would ever finish) to latest updates including the kernel with the fix for that recent CVE.
15:31:00 <nirik> and also I will be gone wed->sunday. If you need me you can send email, I'll probibly check it every once in a while, but likely won't be on IRC any.
15:31:23 <maxamillion> nirik: hope you're off doing something fun! :D
15:31:38 <dgilmore> nirik: time off is good
15:31:41 <nirik> yep. yearly halloween gathering with college friends. ;)
15:31:53 <maxamillion> ohhh right, you mentioned that
15:31:58 <maxamillion> my memory is mush :(
15:32:05 <maxamillion> very cool though :)
15:32:24 <bowlofeggs> new bodhi beta in stg if you want to test. if i don't hear of any blocker bugs by thursday that bad boy's going to prod ☺
15:32:51 <maxamillion> bowlofeggs: SS Shipit
15:32:52 <dgilmore> #info builders updated and rebooted (except arm04 nodes)
15:33:04 <dgilmore> nirik: thanks
15:33:24 <dgilmore> #info bodhi beta in stg, going to Prod Thursday
15:34:15 <dgilmore> anything else or wrap up?
15:34:22 <bowlofeggs> i'll be on PTO next week
15:34:28 <bowlofeggs> just fyi
15:34:38 <dgilmore> #info bowlofeggs on PTO next week
15:34:41 <dgilmore> cheers
15:34:58 <bowlofeggs> actually, if you don't want me to put a bodhi into prod 2 days before PTO that's up to you ☺
15:35:06 <dgilmore> #info nirik on PTO Wednesday, Thursdasy and Friday
15:35:09 <bowlofeggs> i guess i could ask infra if they're cool with that
15:35:17 <dgilmore> bowlofeggs: should be okay
15:35:26 <bowlofeggs> this one doesn't have nearly as much change as 2.2 had
15:35:48 <bowlofeggs> the biggest change is the autosigning feature
15:35:59 <bowlofeggs> and puiterwijk wrote it so he could probably debug if needed
15:37:36 <dgilmore> bowlofeggs: works for me
15:37:41 <dgilmore> .push duty
15:37:44 <dgilmore> .pushduty
15:37:45 <zodbot> dgilmore: The following people are on push duty: Mohan
15:37:48 <zodbot> dgilmore: - https://apps.fedoraproject.org/calendar/release-engineering/
15:38:11 <mboddu> bowlofeggs: I think thats works for me since I am doing pushes this week
15:38:13 <dgilmore> bowlofeggs: just work with mboddu on it
15:40:03 <dgilmore> okay lets wrap up
15:40:06 <dgilmore> #endmeeting