From Fedora Project Wiki
Attendees
- adamw (110)
- handsome_pirate (71)
- kparal (39)
- tflink (32)
- Viking-Ice (26)
- Martix (19)
- pwhalen (19)
- nirik (9)
- brunowolff (8)
- mjg59 (7)
- zodbot (4)
- mkrizek (2)
- jsmith (2)
- misc (1)
- croberts (1)
Agenda
- ARM as a primary arch
- Fedora 20 Change review
- Open floor
ARM as a primary arch
- ARM has been proposed as a primary arch for F20: any QA concerns?
- handsome_pirate has been trying to get ARM to mirror the QA process
- Architectures/ARM/qa-machines lists remotely-accessible machines available for testing purposes
- There are issues with running autotest on ARM hosts on Fedora 18: we agreed that automated testing on ARM should not blocker primary arch elevation as we are not enforcing the tests yet
- QA does not see any definite roadblocks to ARM becoming a primary arch for F20, but expects that minor changes will be needed to the release criteria to define the ARM 'release blocking images' and assistance will be needed from ARM-focused testers to complete ARM validation testing
Fedora 20 Change review
- Tabled until next week due to lack of time
Open floor
N/A
Action items
N/A
IRC Log
| adamw | #startmeeting Fedora QA meeting | 15:00 |
|---|---|---|
| zodbot | Meeting started Mon Jul 15 15:00:23 2013 UTC. The chair is adamw. Information about MeetBot at http://wiki.debian.org/MeetBot. | 15:00 |
| zodbot | Useful Commands: #action #agreed #halp #info #idea #link #topic. | 15:00 |
| adamw | #meetingname fedora-qa | 15:00 |
| zodbot | The meeting name has been set to 'fedora-qa' | 15:00 |
| adamw | #topic Roll call | 15:00 |
| adamw | any rolls? calling all rolls | 15:00 |
| * Martix is lurking here | 15:00 | |
| adamw | ahoyhoy | 15:01 |
| adamw | any other rolls? | 15:01 |
| * kparal also rolling | 15:01 | |
| * mkrizek is here | 15:01 | |
| Martix | adamw: ahoj | 15:02 |
| * adamw wonders what happened to america | 15:02 | |
| * brunowolff is here | 15:02 | |
| adamw | oh good, it didn't sink | 15:03 |
| kparal | do we have tflink? | 15:04 |
| adamw | doesn't look like it | 15:04 |
| adamw | or viking-ice | 15:04 |
| * jsmith lurks | 15:05 | |
| * adamw checks to see if anyone sent apologies | 15:05 | |
| jsmith | adamw: Does it count if I apologize for being here? | 15:05 |
| * handsome_pirate waves | 15:05 | |
| adamw | doesn't seem like it | 15:05 |
| adamw | man the cannons! | 15:05 |
| croberts | lol | 15:05 |
| Martix | adamw: have you read my msgs to you about ARM desktop testing? | 15:06 |
| adamw | handsome_pirate: have jsmith walk the plank | 15:06 |
| adamw | Martix: no, have I been missing them? | 15:06 |
| handsome_pirate | adamw: Arrr | 15:06 |
| Martix | adamw: on Thursday | 15:06 |
| Martix | adamw: on fedora-qa | 15:06 |
| tflink | whoops, not paying attention :-/ | 15:07 |
| adamw | Martix: oh, sorry | 15:07 |
| adamw | alrighty, we'd better get going | 15:07 |
| adamw | Martix: can you send me an email or something? | 15:07 |
| Martix | ok | 15:07 |
| adamw | oh hi, tflink | 15:07 |
| adamw | #topic ARM as a primary arch | 15:07 |
| adamw | so this has been the big bikeshed of the week | 15:07 |
| handsome_pirate | Heh | 15:08 |
| adamw | for anyone who hasn't read it, there's a megathread on devel@ at https://lists.fedoraproject.org/pipermail/devel/2013-July/184962.html | 15:08 |
| adamw | just wanted to throw a topic on the agenda so we can air out any concerns that aren't already covered | 15:08 |
| kparal | I saw the thread and that's why I didn't read it | 15:08 |
| kparal | just skimmed it a bit | 15:08 |
| Martix | adamw: check you INBOX | 15:08 |
| Martix | *your | 15:08 |
| * tflink also has been skimming more than anything | 15:08 | |
| handsome_pirate | adamw: Mind if I go with this just a wee bit, considering my involvement in ARM? | 15:09 |
| brunowolff | There seems to be a lot of arguing over whether being a primary arch means you need to have comparable deliverables to i686/x86_64. | 15:09 |
| adamw | i think the biggest issue raised so far is the question of what exactly qa takes responsibility for | 15:10 |
| adamw | handsome_pirate: shoot! | 15:10 |
| handsome_pirate | Okay | 15:10 |
| adamw | brunowolff: right, though i'm not sure that concerns us a lot | 15:10 |
| handsome_pirate | I've been trying to get ARM to mirror the QA process | 15:10 |
| adamw | #chair handsome_pirate kparal tflink | 15:10 |
| zodbot | Current chairs: adamw handsome_pirate kparal tflink | 15:10 |
| handsome_pirate | #info handsome_pirate has been trying to get ARM to mirror the QA process | 15:10 |
| tflink | is there really enough time before branch to get everything coordinated? | 15:11 |
| handsome_pirate | They're getting better, but not quite there | 15:11 |
| kparal | handsome_pirate: oh, you're jdulaney. how sneaky :) | 15:11 |
| handsome_pirate | FOr instance, they don't seem to understand how blocker bugs are supposed to work | 15:11 |
| handsome_pirate | kparal: Arrr | 15:11 |
| handsome_pirate | (or, the process, I should say) | 15:12 |
| adamw | well they seemed to stop proposing stuff as primary blockers after we talked to them about it during alpha | 15:12 |
| adamw | or did you mean something else? | 15:12 |
| handsome_pirate | At the end of the F19 cycle they finally started doing testing matrices | 15:12 |
| handsome_pirate | adamw: The intention was to get them to file against the arm blockers | 15:12 |
| handsome_pirate | Very few bugs that were blockers for them were actually filed against the trackers | 15:13 |
| adamw | right | 15:13 |
| adamw | ah, i see | 15:13 |
| handsome_pirate | My evaluation from a QA perspective is that they're close, but not quite ready for F20 | 15:14 |
| adamw | so yeah, i do feel like we're going to need a lot of co-operation from 'arm people', however we describe it exactly (the stuff in the thread about what is technically the responsibility of which group feels a bit trivial) | 15:14 |
| tflink | another question I have is whether we consider automated tests a requirement for primary arch | 15:15 |
| adamw | practically speaking, if we're going to sign off on a bunch of ARM images as 'primary deliverables', we at least need to test them all | 15:15 |
| * tflink doesn't remember if pwhalen was able to get autoqa-on-arm working | 15:15 | |
| adamw | and 'QA' as currently stands doesn't have all that hardware | 15:15 |
| handsome_pirate | tflink: The two tests we have now will work on ARM | 15:15 |
| tflink | handsome_pirate: I thought it was more of an autotest/autoqa issue than the individual tests | 15:15 |
| adamw | tflink: that's an interesting one that hasn't been raised yet | 15:15 |
| Martix | adamw: I can test Gnome/KDE on ARM, see my mail | 15:15 |
| handsome_pirate | adamw: I'll get some h/w for QA | 15:16 |
| adamw | Martix: sure, so can I | 15:16 |
| handsome_pirate | adamw: I assume you and tflink would need some | 15:16 |
| adamw | but I still don't think we have *all* the hardware that ARM images are built for, and even if we do, there's the time question of testing them all | 15:16 |
| adamw | no, not really. i have more than I can practically use, | 15:16 |
| handsome_pirate | tflink: The last I tried, the autotest issue was the autotest on f18 issue | 15:16 |
| tflink | handsome_pirate: which I may run into soon, I should have f18 and f19 clients by the end of the week | 15:17 |
| * nirik notes we have some SOC's setup for qa in phx2 already. | 15:17 | |
| nirik | https://fedoraproject.org/wiki/Architectures/ARM/qa-machines | 15:17 |
| Martix | handsome_pirate: do you have something with currently supported GPU in F19? | 15:17 |
| Martix | handsome_pirate: I mean 3D accel | 15:17 |
| handsome_pirate | nirik: Aye | 15:17 |
| * tflink has some arm hardware but doesn't use it much | 15:18 | |
| handsome_pirate | Martix: Everything I have is headless save for my chromebook | 15:18 |
| adamw | tflink: so the detail here is that autotest does not run on ARM on f18+? | 15:18 |
| pwhalen | tflink, autoqa worked well when using earlier versions | 15:18 |
| tflink | adamw: yeah, need to dig into it | 15:18 |
| pwhalen | I have access to all supported hardware for testing | 15:18 |
| tflink | I have an idea of what the problem is, need to test it | 15:18 |
| kparal | adamw: autoqa needs to be updated to worked with newer autotest versions. but we didn't bother, since we use RHEL6 and old autotest version | 15:19 |
| * tflink has been rebuilding and scripting virthosts | 15:19 | |
| pwhalen | I've tried to follow the release validation testing as best as possible for f19 | 15:19 |
| adamw | #info there are issues with running autotest on ARM hosts on Fedora 18 | 15:19 |
| tflink | pwhalen: do you still have stuff running against f17 on arm? | 15:19 |
| adamw | pwhalen: sure, that's what handsome_pirate was talking about | 15:19 |
| Martix | handsome_pirate: well, I have some hw with Mali (supported in F19) and Adreno (userspace support will be in F20) | 15:19 |
| pwhalen | tflink, no | 15:19 |
| handsome_pirate | Martix: Chromebook is Mali | 15:20 |
| kparal | I don't understand one thing - I don't see any ARM devices around me, anywhere. except for RaspPi, which is not going to be supported anyway. so I might not understand what primary architecture means, but why would we have PA for something that is so thinly available? | 15:20 |
| handsome_pirate | kparal: Only a few specific devices are supported, but they're fairly cheap and easily had | 15:21 |
| adamw | kparal: it's kind of a split issue: the vague idea we might run on zillions of ARM tablets or something, and a very specific market for ARM servers. | 15:21 |
| misc | kparal: because it is expected to be avaliable later, i think | 15:21 |
| handsome_pirate | kparal: I'll bring up about getting you some h/w if you want | 15:22 |
| adamw | plus a hedge against ARM Taking Over The World. | 15:22 |
| kparal | easily had, but I don't see them around me or my colleagues, or my friends, or my friends' friends, ... | 15:22 |
| kparal | anyway, probably a topic for the mailing list | 15:22 |
| * handsome_pirate notes that the biggest issue with arm is proprietary gfx | 15:22 | |
| Martix | kparal: HP Touchpad, Cubieboard... | 15:23 |
| adamw | kparal: still - yeah, that's not a QA topic in particular | 15:23 |
| brunowolff | OLPC machines also use arm, though they need a special kernel now. But I have done updates and installs of non-kernel arm packages from Fedora. | 15:23 |
| brunowolff | A number of Fedora people got OLPC machines last summer. | 15:24 |
| adamw | so in terms of specific concerns: handsome_pirate is worried about following procedure in terms of using blocker bugs, tflink is worried about autoqa | 15:24 |
| adamw | do we have anything else? | 15:24 |
| brunowolff | Documentation of the deliverables that we need to test. | 15:24 |
| * Viking-Ice joins in | 15:24 | |
| tflink | time for adjustment | 15:24 |
| pwhalen | we've been using blocker bugs, though work could likely be done, absolutely open to all suggestions | 15:25 |
| kparal | adamw: the biggest concern is whether we need to do the release validation for every single subarch (or how it is called) of ARM | 15:25 |
| tflink | we have what, 3-4 weeks before F20 branch | 15:25 |
| kparal | as you already stated | 15:25 |
| handsome_pirate | tflink: Mayve a release cycle where QA treats arm as primary but it doesn't actually block? | 15:25 |
| * Viking-Ice playing arm catchup | 15:25 | |
| brunowolff | You could have us only block on some of the stuff that we would block on for i686/x86_64. | 15:26 |
| adamw | kparal: that's my concern, yeah, which ties in with the 'following procedure' thing, as in practice we will need pwhalen and other ARMers to ensure we can do that | 15:26 |
| handsome_pirate | kparal: Actually, the arm team is trying to make it so that there are only a limited number of images to test | 15:26 |
| handsome_pirate | ie, make them work on as wide a variety of h/w as possible | 15:26 |
| adamw | handsome_pirate: that didn't look to be in the works for f20, though | 15:26 |
| kparal | handsome_pirate: how many? | 15:26 |
| handsome_pirate | kparal: right now, there are three images | 15:27 |
| handsome_pirate | kparal: One supported piece of h/w requires a vfat boot partition | 15:27 |
| kparal | because it's not 'ARM as a primary arch', but 'ARM as an additional three primary archs', which sounds much scarier :) | 15:27 |
| brunowolff | Is the arm team going to expect the Gnome 3 desktop to be a deliverable? | 15:27 |
| handsome_pirate | kparal: Otherwise, lpae and a standard image | 15:28 |
| handsome_pirate | ie, lpae kernel | 15:28 |
| Viking-Ice | with a late chime in as I see it basically the main problem with us regarding arm is for someone to actually own arm devices and being able to test on them and apply relevant critera where applicable | 15:28 |
| adamw | Viking-Ice: we do have several people with some ARM devices, but we're still trying to determine exactly how much coverage is expected | 15:28 |
| handsome_pirate | adamw: Most of the testing is the same | 15:29 |
| Viking-Ice | adamw, simple really if the plan is to release an Gnome arm image the relevant Gnome release criteria applies to that image, if the plan is to release only KDE image only the KDE criteria applies etc | 15:29 |
| handsome_pirate | well, a lot of it is | 15:29 |
| pwhalen | Viking-Ice, I do my best to do that now and will going forward. I tested all images for f19 | 15:29 |
| handsome_pirate | ANd, x86 tests will be sufficient | 15:29 |
| pwhalen | we're trying to reduce this to a single image at some point, currently we have two image types due to partitioning for omap requiring vfat. | 15:32 |
| adamw | pwhalen: so for f20, what exactly are the expected ARM deliverables assuming it's a primary arch? | 15:32 |
| handsome_pirate | pwhalen: We have an lpae image as well, yes? | 15:33 |
| Viking-Ice | yeah basically what we need to know ( QA ) is which "spins/de" arm architecture will be released on and we need to test that against the criteria and block based on matching criteria | 15:33 |
| pwhalen | adamw, that to my knowledge is not yet decided. dgilmore? | 15:33 |
| kparal | maybe some arm images might not be considered release blocking, if they are served to just a small audience? | 15:33 |
| pwhalen | handsome_pirate, there's our official offerings https://fedoraproject.org/en/get-fedora-options#2nd_arches | 15:34 |
| Viking-Ice | kparal, define small audience ;) | 15:34 |
| kparal | Viking-Ice: well, an example is GNOME vs XFCE | 15:35 |
| kparal | I don't know how many arm devices are covered by that or that image | 15:35 |
| adamw | okay, so there's basically two 'arches' for ARM for f19 | 15:35 |
| kparal | just throwing ideas | 15:35 |
| handsome_pirate | pwhalen: My mistake | 15:35 |
| pwhalen | handsome_pirate, lpae kernel is installed on the hfp image | 15:35 |
| handsome_pirate | pwhalen: Ah | 15:35 |
| Viking-Ice | kparal, there are two "release" blocking images there on the link that pwhalen posted KDE and Minimal | 15:36 |
| Viking-Ice | Gnome is not listed there so | 15:36 |
| handsome_pirate | kparal: Most Fedora Arm folks I know of tend to run xfce, or at least that used to be the case | 15:36 |
| adamw | so if things were the same for f20 i'd expect the kde and 'minimal' images of both 'arches' to be release blocking: four images | 15:36 |
| kparal | so, it's a single arch, and two install images, right? but the exactly same packages? | 15:36 |
| adamw | Viking-Ice: they didn't do a GNOME image on the basis nothing would actually have hw accel support on it | 15:36 |
| handsome_pirate | kparal: AYe | 15:36 |
| kparal | handsome_pirate: ok, I get it finally | 15:36 |
| pwhalen | kparal, right | 15:37 |
| nirik | there's a single arch. | 15:37 |
| adamw | nirik: i was using the term incorrectly on purpose, hence the scare quotes | 15:37 |
| mjg59 | adamw: Not quite. No hw support and the sw fallback was broken. | 15:37 |
| nirik | one tree/collection of packages. ;) | 15:37 |
| Viking-Ice | adamw, we only match what the do against our criteria and based on that link only two ( in current release criteria setup ) are blocking the minimal and KDE | 15:37 |
| adamw | i was hoping everyone could follow that =) | 15:37 |
| Viking-Ice | s/the/they | 15:37 |
| Martix | pwhalen: any idea why Gnome is missing on the list? | 15:37 |
| kparal | another question, is anaconda expected to work for ARM in F20? | 15:37 |
| adamw | Martix: see above. | 15:37 |
| pwhalen | omap devices require vfat, wheras every other device works with ext | 15:37 |
| adamw | pwhalen: is there a chance of the vfat-vs-non-vfat thing being reconciled for f20? | 15:37 |
| Martix | adamw: I can run Gnome Shell on Adreno GPU | 15:37 |
| adamw | or are we stuck with that? | 15:38 |
| handsome_pirate | Martix: llvmpipe is borked because llvm is borked (/me is working on that) | 15:38 |
| Viking-Ice | in a perfect pony world all images would be run through the critera and blocked upon but hey we have an default... | 15:38 |
| pwhalen | adamw, I believe so, I think it was very close, dgilmore was working on it | 15:39 |
| Martix | handsome_pirate: work on llvm-pipe is great, but there is some HW loke Chromebook which has open source 3D GPU driver | 15:39 |
| handsome_pirate | Martix: Chromebook sort of has one | 15:40 |
| adamw | okay. so given that picture of the deliverables, the work doesn't seem crazily difficult | 15:40 |
| kparal | is anaconda expected to work for ARM in F20? that would affect the QA release validation time a lot | 15:40 |
| handsome_pirate | Martix: Chromebook is complicated, and not likely to be officially supported due to the bloody firmware | 15:40 |
| handsome_pirate | kparal: No | 15:40 |
| adamw | pwhalen: do you confirm re anaconda? | 15:40 |
| kparal | good :) | 15:40 |
| handsome_pirate | No Anaconda | 15:41 |
| handsome_pirate | adamw: I talke about that during job interview with clumens | 15:41 |
| Viking-Ice | no anaconda means we need to adjust the release criteria to match whatever installation method they use | 15:41 |
| kparal | without anaconda support, it might be actually so much work | 15:41 |
| adamw | Viking-Ice: it doesn't, really | 15:41 |
| kparal | (as I expected) | 15:41 |
| handsome_pirate | Installation method: dd to sd card | 15:42 |
| adamw | Viking-Ice: the criteria are couched in terms of 'the installer must blah' and 'the installer must foo' | 15:42 |
| handsome_pirate | If device has internal storage, then dd image to that | 15:42 |
| Viking-Ice | adamw, ah excellent | 15:42 |
| adamw | they don't actually explicitly say 'all images must use the installer' or anything | 15:42 |
| mjg59 | Anaconda was a required part of the promotion criteria | 15:42 |
| adamw | i've covered that on the devel@ list thread; i'm not too worried about that angle | 15:42 |
| adamw | mjg59: OFF TOPIC FOR QA MEETING | 15:42 |
| nirik | anaconda works as far as I can see. | 15:42 |
| adamw | god, this has already gone on for 40 minutes. | 15:42 |
| nirik | at least the ones I have installed have used anaconda text mode just fine. | 15:42 |
| pwhalen | we use anaconda for headless server installs, and preoduce images for dev boards | 15:42 |
| mjg59 | adamw: I'm aware. I'm just pointing out that if there's no plans for Anaconda to work in F20, worrying about primary arch QA for F20 seems premature | 15:43 |
| kparal | so it might be arm+anaconda or nothing, right? | 15:43 |
| adamw | well, what pwhalen just said doesn't sound anything like 'anaconda doesn't work'. | 15:43 |
| kparal | and nirik | 15:44 |
| handsome_pirate | mjg59: My talking with the Anaconda team led me to believe they have no plans for anaconda on arm | 15:44 |
| nirik | anaconda works. kickstart works. | 15:44 |
| Viking-Ice | is there something that says PA have to be installable by Anaconda | 15:44 |
| * nirik thinks this is in the weeds. | 15:44 | |
| handsome_pirate | Anaconda does work, it's just not used very much | 15:44 |
| kparal | let's assume anaconda support will be added, because we have it confirmed that it works at least partially | 15:44 |
| * nirik uses it all the time. well, kickstarts usually with text mode to watch | 15:45 | |
| kparal | which means more installation matrices! | 15:45 |
| tflink | ooh, that sounds like fun :-/ | 15:45 |
| Viking-Ice | limiting PA to anaconda dont make much sense since someday someone might want to write and use another installer for their "spin" | 15:45 |
| adamw | Viking-Ice: there's nothing QAish that says that, but mjg59 is saying someone else said that. i assume FESCo. | 15:45 |
| mjg59 | handsome_pirate: Oh, right. I don't know that anyone on the Anaconda team is working on it - it'd be the ARM SIG doing it | 15:45 |
| handsome_pirate | adamw: mjg59 said that | 15:45 |
| handsome_pirate | adamw: He wrote the criteria | 15:45 |
| mjg59 | https://fedoraproject.org/wiki/Secondary_Architecture_Promotion_Requirements | 15:46 |
| * Martix brought some snack to watch the flame :-) | 15:46 | |
| mjg59 | Drafted rather than wrote, but they were signed off on by fesco | 15:46 |
| handsome_pirate | Anyway | 15:46 |
| adamw | well, the question of whether fesco wants to hold to that requirement is not ours to answer | 15:47 |
| Viking-Ice | I would see the reasoning for requiring the use of Anaconda | 15:47 |
| mjg59 | Anyway, sorry for diverting it. nirik says he has personal experience of it working, so it's not relevant | 15:47 |
| adamw | i don't think we can really block on ARM ever becoming a primary arch because it gives us another giant pile of matrix to work through, however much we want to :) | 15:47 |
| Viking-Ice | s/would see/would like to see | 15:47 |
| adamw | that's just going to wind up being one of those things we need extra people for | 15:47 |
| handsome_pirate | adamw: Hence my trying to get the arm folks to follow the qa processs | 15:48 |
| handsome_pirate | That way, they can do the work | 15:48 |
| adamw | Viking-Ice: can you take that to fesco or something? it's a fesco requirement | 15:48 |
| pwhalen | adamw, I will be on the hook for filling said matrices | 15:48 |
| adamw | handsome_pirate: right. | 15:48 |
| Viking-Ice | actually it might spark new willingness to test thus interest in participating in the QA community | 15:48 |
| Viking-Ice | ( arm that is ) | 15:48 |
| adamw | so let me try and boil this down | 15:49 |
| adamw | we didn't really get anywhere on the autoqa thing; tflink, do you have a path forward there? | 15:49 |
| tflink | adamw: a little bit, it'll depend on the results of my experiments | 15:49 |
| kparal | mkrizek had some patches, they might be obsolete already | 15:50 |
| adamw | okay | 15:50 |
| tflink | kparal: IIRC, they didn't work | 15:50 |
| adamw | are you actually *worried* about autoqa not working on arm, practically speaking? | 15:50 |
| adamw | or are you just concerned someone else might consider it a box that needs checking? | 15:50 |
| mkrizek | they did, but I dont think they do anymore | 15:50 |
| adamw | i'm not sure anything autoqa currently achieves is vital enough that we need to make it an obstacle to primary arch status | 15:50 |
| pwhalen | tflink, let me know if there is anything I can do, I got a fair amount of it working, however had some bugs with the latest version (would need to check which) | 15:50 |
| kparal | we can have upgradepath working for arm updates. depcheck might be a problem. since depcheck is very broken, I don't see it as a major issue | 15:51 |
| tflink | adamw: those are two different questions - I'm not sure it's going to work, no | 15:51 |
| adamw | though obviously it'd be nice | 15:51 |
| handsome_pirate | adamw: Autoqa (or its replacement) on arm is on my to do list | 15:51 |
| tflink | kparal: I thought depcheck was still mostly working | 15:51 |
| Viking-Ice | I dont consider autoqa not working for arm do be somekind of blocker for it from a QA perspective | 15:51 |
| Viking-Ice | s/do be/to be/ | 15:51 |
| kparal | tflink: yes, but it is arch-specific, so we might find some issues with it. upgradepath can be run on x86 box to check arm updates | 15:52 |
| tflink | kparal: but that's different from it being completely broken, no? | 15:52 |
| tflink | more of a not-quite-known? | 15:52 |
| kparal | tflink: ok, sometimes broken :) | 15:52 |
| kparal | bad wording from my side | 15:52 |
| adamw | so overall it seems like we don't feel like we need to make a big fuss about the autoqa angle | 15:53 |
| tflink | adamw: I don't think automation-on-arm is a blocker for F20 as we're not gating updates yet | 15:53 |
| adamw | but we should probably bake it into taskbot | 15:53 |
| tflink | if we start gating updates - even in rawhide, I think it'll be an issue | 15:53 |
| kparal | adamw: I wouldn't consider autoqa to be a large concern, no | 15:53 |
| * handsome_pirate can confirm that as of the last time he tried, taskbot worked on arm | 15:53 | |
| adamw | #info we agreed that automated testing on ARM should not blocker primary arch elevation as we are not enforcing the tests yet | 15:53 |
| adamw | okay. | 15:53 |
| * tflink wonders if we should be adding arm clients to taskbot from the get-go, but that's off topic for here | 15:54 | |
| handsome_pirate | Mind, that was about a month ago, I don't know how much tflink has worked on it since then | 15:54 |
| adamw | the other major angle we covered is the intersection of 'what deliverables will there be for ARM as a primary arch' and 'what resources do we have for testing' | 15:54 |
| tflink | handsome_pirate: not nearly as much as I would like :'( | 15:54 |
| adamw | it seems like the deliverables load is not too bad, but we feel like we'll definitely need co-operation from pwhalen and ARM-focused volunteers to ensure we can cover all the testing | 15:55 |
| handsome_pirate | tflink: If you get some work done, ping me and I'll fire it up on arm | 15:55 |
| adamw | is that a fair summary? | 15:55 |
| handsome_pirate | adamw: I'd say so | 15:55 |
| tflink | adamw: yeah, I think so. that and making the processes work together | 15:55 |
| pwhalen | adamw, I can look into our planned deliverables. and provide testing coverage | 15:55 |
| kparal | adamw: let's add that it is unclear how many different devices we need to test? | 15:56 |
| Martix | adamw: it's ok | 15:56 |
| adamw | kparal: just like x86 =) | 15:56 |
| Viking-Ice | speaking of upgrades I dont think the upgrade criteria takes effect until $NEXT_RELEASE for new PA as in we do not "support" upgrading from f18/19 on to f20 in the case of arm but only "support" upgrading from F20 to F21 and onwards ( the release architecture became PA and onwards ) | 15:57 |
| handsome_pirate | kparal: Actually, arm has fewer devices than x86 | 15:57 |
| handsome_pirate | In that regaurd, arm is easier to test | 15:57 |
| adamw | #info as described by handsome_pirate and pwhalen, the ARM deliverables set does not look unmanageable, but we feel like we'll definitely need co-operation from pwhalen and ARM-focused volunteers to ensure we can cover all the testing | 15:57 |
| adamw | Viking-Ice: that seems reasonable | 15:57 |
| adamw | so overall, let me punt a #agreed | 15:57 |
| handsome_pirate | ack | 15:58 |
| Viking-Ice | ack | 15:58 |
| Martix | ack | 15:58 |
| pwhalen | ack | 15:58 |
| adamw | er, that wasn't it =) | 15:58 |
| tflink | what exactly are we ack'ing | 15:58 |
| tflink | ? | 15:58 |
| Viking-Ice | awesomeness and bacon | 15:59 |
| kparal | ack | 15:59 |
| kparal | you should have said it earlier | 15:59 |
| adamw | proposed #agreed QA does not see any definite roadblocks to ARM becoming a primary arch for F20, but expects that minor changes will be needed to the release criteria to define the ARM 'release blocking images' and assistance will be needed from ARM-focused testers to complete ARM validation testing | 15:59 |
| Viking-Ice | ack | 15:59 |
| tflink | ack | 15:59 |
| kparal | ack | 16:00 |
| * handsome_pirate assumed that was what he was acking | 16:00 | |
| adamw | wow, you're clairvoyant too?! | 16:00 |
| handsome_pirate | arrr | 16:00 |
| brunowolff | ack | 16:00 |
| adamw | coolbeans | 16:00 |
| pwhalen | ack | 16:00 |
| adamw | #agreed QA does not see any definite roadblocks to ARM becoming a primary arch for F20, but expects that minor changes will be needed to the release criteria to define the ARM 'release blocking images' and assistance will be needed from ARM-focused testers to complete ARM validation testing | 16:00 |
| adamw | and we're right on time, with no time for the Change review topic :( | 16:01 |
| adamw | before i close out, does anyone have any really major concerns with any of the F20 Changes that can't wait till next week? | 16:01 |
| handsome_pirate | pwhalen: Wednesday, you want to take this? | 16:01 |
| handsome_pirate | adamw: I have something for the end | 16:02 |
| tflink | not really, I need to start reading those threads more closely | 16:02 |
| kparal | adamw: the AppInstaller might be worth discussion, but it can wait I think | 16:02 |
| * tflink wonders if the get-rid-of-syslog is going to turn interesting | 16:02 | |
| adamw | okay, let's file those for next week | 16:02 |
| adamw | a very quick: | 16:02 |
| adamw | #topic open floor | 16:02 |
| adamw | handsome_pirate: what did you have? | 16:02 |
| * handsome_pirate needs a room mate for FLOCK | 16:02 | |
| handsome_pirate | My room mate for the past several fudcons is not going | 16:03 |
| handsome_pirate | EOF | 16:03 |
| adamw | heh, okay | 16:03 |
| adamw | #info handsome_pirate looking for a roommate for FLOCK, applications on postcards | 16:03 |
| adamw | anything else? | 16:03 |
| adamw | yikes, if i don't clear up mine/tflink's room type problem you might wind up rooming with one of us... | 16:04 |
| tflink | adamw: I thought that was cleared up | 16:04 |
| Viking-Ice | tflink, thanks to FESCO/FPC yes it's going to be a more pain in the ass for us then it would be. I tried work in advance for the inevitable for us in QA with this https://fedoraproject.org/wiki/Features/systemd-journal but notting said no... | 16:04 |
| handsome_pirate | adamw: SInce I don't have a room booked, maybe you could not clear it up? :) | 16:05 |
| adamw | tflink: i'll ping you | 16:05 |
| adamw | alrighty, thanks for coming everyone! | 16:06 |
| adamw | Change review next week | 16:06 |
| adamw | #endmeeting | 16:06 |
Generated by irclog2html.py 2.12.1 by Marius Gedminas - find it at mg.pov.lt!