16:30:45 <dgilmore> #startmeeting RELENG (2016-01-11)
16:30:45 <zodbot> Meeting started Mon Jan 18 16:30:45 2016 UTC.  The chair is dgilmore. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:30:45 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:30:45 <zodbot> The meeting name has been set to 'releng_(2016-01-11)'
16:30:45 <dgilmore> #meetingname releng
16:30:45 <zodbot> The meeting name has been set to 'releng'
16:30:45 <dgilmore> #chair dgilmore nirik tyll sharkcz bochecha masta pbrobinson pingou maxamillion
16:30:45 <zodbot> Current chairs: bochecha dgilmore masta maxamillion nirik pbrobinson pingou sharkcz tyll
16:30:48 <dgilmore> #topic init process
16:30:52 <dgilmore> who is here?
16:30:53 <nirik> morning
16:30:59 * sharkcz is here
16:31:37 * masta is here o/
16:32:20 <maxamillion> .hello maxamillion
16:32:22 <zodbot> maxamillion: maxamillion 'Adam Miller' <maxamillion@gmail.com>
16:33:20 <dgilmore> lets get started
16:33:33 <dgilmore> there is no tickets with the meeting keyword
16:34:00 <dgilmore> #topic compose process
16:34:12 <dgilmore> quickly wanted to go over the changes for f24
16:34:36 <dgilmore> sharkcz: I have yet to test pungi on secondary arches
16:34:57 <dgilmore> I need to file some tickets over layered configs
16:35:35 <dgilmore> but you should be able to switch in the arches and run the nightly script in pungi-fedora git repo
16:35:55 <dgilmore> sharkcz: it needs the runroot plugin be installed and enabled
16:36:17 <sharkcz> dgilmore: on builders?
16:36:25 <dgilmore> sharkcz: builders and hub
16:36:34 <dgilmore> there is a hub side and builder side
16:36:46 <dgilmore> s390 may be a bit tricky
16:37:10 <sharkcz> yep, especially because the "remote" setup
16:37:46 <dgilmore> yep
16:38:05 <dgilmore> the builders in runroot need to have /mnt/koji mounted rw
16:38:59 <sharkcz> IIRC pbrobinson asked RH IT for firewall changes quite some time ago
16:39:21 <sharkcz> so in theory it could work
16:39:31 <dgilmore> okay
16:39:37 <dgilmore> the builders can not do builds
16:40:50 <sharkcz> you mean the builders with runroot doing the composes?
16:40:55 <dgilmore> i am hoping that rawhide will be changed by friday
16:40:58 <dgilmore> sharkcz: yes
16:41:24 <dgilmore> so we need to get some secondary arch testing this week
16:41:48 <sharkcz> ok, will add it to my plans
16:42:50 <dgilmore> ppc and aarch64 should be simpler
16:43:02 <sharkcz> yep
16:43:18 <nirik> do we need to let adamw know for openqa? ie, are path's or image names going to change any?
16:43:18 <dgilmore> but will need testing also
16:43:27 <dgilmore> nirik: likely yes
16:43:37 <dgilmore> nirik: all the paths and image names will change
16:43:43 <nirik> the base repos should be the same for rawhide users tho right?
16:43:44 <adamw> then yes, i need to know :P
16:43:55 <dgilmore> nirik: metalinks yes
16:44:01 <nirik> ie. mm should do the right thing... yeah.
16:44:03 <adamw> if someone could just send me a list of what all the changes will be and when they're going to happen, that'd be great.
16:44:04 <dgilmore> but the actual paths will change
16:44:25 <dgilmore> adamw: not sure when, but rawhide will look like a TC
16:44:26 <adamw> or if the new paths are the ones that exist already and we're just going to drop the 'legacy' ones, i can work with that too.
16:44:58 <adamw> dgilmore: well, i'll need the paths to the compose locations too of course
16:45:16 <dgilmore> adamw: so rawhide will have Workstation, Cloud, Server, Everything
16:46:01 <adamw> dgilmore: what i'm saying is, will the composes still be in the same folder layout under https://kojipkgs.fedoraproject.org/mash/
16:46:01 <dgilmore> adamw: https://kojipkgs.fedoraproject.org/compose/rawhide/Fedora-24-20160113.n.1/compose/ that what it will look like so far
16:46:10 <dgilmore> adamw: they will not be
16:46:24 <adamw> so, no. OK. I'll set up a branch for fedfind and stick it in there
16:46:30 <dgilmore> adamw: they will be under /compose/rawhide/
16:46:39 <adamw> fwiw I still hold that Rawhide nightly composes do not naturally have a release number. :P
16:47:01 <nirik> dgilmore: does this do spam-o-matic/deps check? and what will be sent to devel about the compose?
16:48:06 <dgilmore> nirik: spam-o-matic will need wired in, there is a tool for a compose report I need to run it
16:48:33 <maxamillion> what's spam-o-matic?
16:48:44 <adamw> the bot that produces the rawhide reports, i think
16:48:57 <nirik> a script that mails broken deps owners about the broken deps
16:49:10 <nirik> it tries to figure out which packages are responsable
16:49:13 <maxamillion> ahhh ok
16:49:28 <nirik> it's basically a repoclosure + mail bomber
16:49:37 <dgilmore> adamw: https://kojipkgs.fedoraproject.org/compose/rawhide/Fedora-24-20160113.n.1/compose/metadata/composeinfo.json
16:50:14 <adamw> dgilmore: thanks
16:51:21 <dgilmore> need to package up https://pagure.io/compose-utils
16:52:11 <nirik> cool
16:52:57 <dgilmore> so expect lot sof composes this week
16:53:06 <dgilmore> as things get narrowed down
16:53:31 * nirik nods
16:53:46 <dgilmore> #topic Secondary Architectures updates
16:53:47 <dgilmore> #topic Secondary Architectures update - ppc
16:53:52 <dgilmore> sharkcz: how is ppc?
16:54:09 <sharkcz> no issues afaik
16:54:47 <sharkcz> update to f23 on builders should follow the nw/net reorg in PHX
16:54:47 <dgilmore> how are they looking in preperation for the mass rebuild
16:54:52 <dgilmore> okay
16:55:29 <dgilmore> #info ppc builder move soon
16:55:33 <dgilmore> #topic Secondary Architectures update - s390
16:55:39 <dgilmore> sharkcz: how is s390?
16:56:07 <sharkcz> build-wise no issues, builder were migrated to f23 cca week ago
16:56:16 <dgilmore> okay
16:56:24 <nirik> we don't yet have the new net/vlan for ppc, but hopefully soon. ;) Also, was going to use that same net/vlan for aarch64/moonshot, FYI
16:56:26 <dgilmore> what version of koji are you using?
16:56:52 <sharkcz> from f23 updates, let me check
16:57:07 <dgilmore> sharkcz: the version in the infra repo should be used
16:57:18 <dgilmore> it does have a schema update that is needed
16:57:56 <sharkcz> hub is covered now more by nirik (or ansible)
16:58:05 <dgilmore> okay
16:58:35 <dgilmore> all the configs for the runroot plugin are in ansible
16:58:44 * nirik nods.
16:58:44 <sharkcz> but yep, will switch the builders to the infra repo
16:58:58 <dgilmore> https://infrastructure.fedoraproject.org/repo/builder-rpms/23/s390x/
16:58:59 <nirik> can we reach the builders from phx2/
16:59:08 <dgilmore> there is a s390x repo for the builders
16:59:11 <nirik> ah, you mean rpms... right.
16:59:36 <sharkcz> but we are again low on disk space on the hub :-(
16:59:42 <dgilmore> nirik: the builders will have to pull the configs from ansible
17:00:00 <sharkcz> and I think I've cleaned up everything ...
17:00:09 <nirik> sharkcz: yeah, was meaning to look at that... noticed it was low last week.
17:00:29 <dgilmore> nirik: one thing that should help some is my plans to move all the old secondary arch releases to archive
17:01:02 <nirik> that would be nice.
17:01:04 <sharkcz> yep, that should help
17:01:11 <nirik> both ppc and s390 are really low.
17:01:14 <dgilmore> we will be able to remove some content that only exists in /mnt/koji/tree
17:01:15 <nirik> let me +1TB?
17:01:44 <nirik> dgilmore: did you get a chance to clean up primary mashes?
17:01:46 <sharkcz> please do if you can
17:01:58 <dgilmore> nirik: not yet
17:02:21 <nirik> primary is at 87%
17:02:59 <dgilmore> kicked of a rm
17:03:42 <dgilmore> #topic Secondary Architectures update - arm
17:03:54 <dgilmore> not sure what aarch64 is up to
17:04:15 <nirik> ok, added 1TB to arm, ppc, s390
17:04:30 <nirik> I can give a quick update on the moonshot. ;)
17:04:39 <dgilmore> nirik: please
17:05:15 <nirik> We had a lot of fun last week because it was longer than our normal racks could hold, but we managed to get a bigger rack, move all the mustangs into it and move it in place and then racked the moonshot
17:05:35 <dgilmore> nirik: joy
17:05:43 <nirik> we are able to get to the iLo console on it. It still needs networking setup from RHIT (hopefully sometime this week)
17:05:56 <nirik> and then we can start hopefully using it.
17:06:03 <dgilmore> that would be good
17:06:06 <nirik> it would be great if we could have it working by mass rebuild
17:06:12 <masta> does it have a 10g connection?
17:06:29 <nirik> nope, we don't have any 10g stuff there to connect it to.
17:06:34 <nirik> it's got multiple 1gs
17:06:43 <dgilmore> nirik: I want to enable 32 bit arm docker base images. creation of does not yet work on aarch64
17:07:02 <dgilmore> nirik: would be good
17:07:34 <nirik> we have I think 4 1g uplinks hooked up... but we could do more if needed.
17:07:47 <dgilmore> I think some work will still be needed to make running armv7 vms work seamlessly
17:07:51 <nirik> it has 30 cartrages
17:08:19 <dgilmore> nice
17:08:31 <dgilmore> any idea the ram, storage etc on them?
17:08:32 <nirik> (and can hold 15more if we want to expand)
17:08:36 <masta> indeed, nice!
17:08:43 <nirik> not sure, I can try and look.
17:12:11 <nirik> it just says "ProLiant m400 Server Cartridge"
17:12:16 <dgilmore> fun
17:12:21 <dgilmore> oh well
17:13:21 <nirik> this is just serial iLO... the real mgmt should have lots more info I would expect.
17:13:32 <dgilmore> nirik: how evil would it be to run kojid from here as a short term solution to enabling docker base image creation
17:13:41 <dgilmore> nirik: likely
17:14:07 <nirik> pretty evil. ;) But I guess it could be a stop gap
17:14:35 <dgilmore> in testing its not working on aarch64
17:16:06 <dgilmore> anyway
17:16:11 <dgilmore> #topic open floor
17:16:17 <dgilmore> any questions?
17:16:31 <maxamillion> I have one
17:16:49 <maxamillion> I was looking over https://fedorahosted.org/fedora-infrastructure/ticket/4857 and noticed that it was closed because the infra side is done, is there still anything needed from the releng side?
17:20:08 <dgilmore> there is all the release engineering side to do
17:20:17 <dgilmore> we have a issue for it somewhere
17:20:32 * lmacken has to run, but would just like to point out an important ticket that didn't get tagged for the meeting
17:20:39 <lmacken> There's a critical Atomic Host bug that was fixed a month ago, and the installer needs to be rebuilt against the update. https://fedorahosted.org/rel-eng/ticket/6332
17:21:07 <lmacken> Right now we're shipping images that boot to the recovery console after installation
17:21:39 <dgilmore> lmacken: please wiat until we are done with maxamillion's issue
17:21:47 <lmacken> dgilmore: I can't, I said I have to run.
17:22:05 * lmacken just putting it on your radar
17:22:07 <dgilmore> lmacken: then add it to next weeks meeting if you will not be here to discuss
17:22:23 <dgilmore> its on my radar, it needs a solution engineered
17:22:27 <maxamillion> dgilmore: if you could find the ticket, I would appreciate it .... I'd like to follow up on it if I can ... if there's naything I can do to help move that work along I'd be happy to help there
17:22:28 <lmacken> it doesn't need to get discussed, it needs to get done before the next weeks meeting
17:23:01 <dgilmore> maxamillion: it needs someone to sit down and write code
17:23:16 <dgilmore> which kinda sucks for a one off not reusable thing
17:23:43 <dgilmore> maxamillion: its likely a weeks worth of work to write the code and do end to end testing
17:23:47 <maxamillion> dgilmore: well, it could be reusable if something else like this ever pops back up
17:23:52 <dgilmore> maxamillion: nope
17:24:13 <dgilmore> maxamillion: it could maybe be forked and used as a basis
17:24:37 <maxamillion> dgilmore: alright, can I pick your brain on that later today for exact needs/requirements? I don't mind writing the code
17:24:45 <dgilmore> there is likely too many hard coded specifics
17:24:53 <dgilmore> maxamillion: sure
17:24:59 <dgilmore> maxamillion: did someone ping you>
17:25:01 <dgilmore> ?
17:25:32 <dgilmore> I just ask because I was emailed about it last week by workstation peopole and I told them we will not be looking at it until after f24 is done
17:25:55 <maxamillion> dgilmore: yeah, someone was asking me if I knew anything about it and I didn't really so I started digging into the tickets and here we are
17:26:18 <maxamillion> dgilmore: oh, I didn't know the workstation folks reached out to you
17:26:58 <dgilmore> maxamillion: sorry you got put in the middle like that
17:27:26 <maxamillion> dgilmore: no worries
17:27:43 <dgilmore> maxamillion: they should have juyst filed a ticket asking
17:27:49 <dgilmore> not gone email people directly
17:28:07 <dgilmore> Firefox can't find the server at taiga.cloud.fedoraproject.org.
17:28:58 <maxamillion> dgilmore: I think it moved to http://taiga.fedorainfracloud.org
17:30:03 <dgilmore> maxamillion: hrrm that is not in my browser history at all
17:33:26 <maxamillion> dgilmore: no idea
17:35:40 <dgilmore> maxamillion: having trouble finding it
17:35:56 <dgilmore> maxamillion: but I am pretty sure acarter filed the issue in taiga
17:37:14 <maxamillion> dgilmore: alright, we can look into it off-meeting
17:39:32 <dgilmore> maxamillion: yep
17:40:01 <dgilmore> as to lmackens ticket I have no idea what we should do. people were informed before f23 GA how things would work
17:40:13 <dgilmore> does anyone have anything else?
17:42:05 <dgilmore> #endmeeting