Monday, 2016-10-03

*** mordred has quit IRC00:48
*** mordred has joined #openstack-requirements00:52
*** coolsvap has joined #openstack-requirements03:41
*** mugsie has quit IRC05:50
*** mugsie has joined #openstack-requirements05:51
coolsvaptonyb: around?05:55
openstackgerritDirk Mueller proposed openstack/requirements: Updated from generate-constraints  https://review.openstack.org/38070306:38
*** jpena|off is now known as jpena07:20
openstackgerritOpenStack Proposal Bot proposed openstack/requirements: Updated from generate-constraints  https://review.openstack.org/38070307:25
*** zigo has quit IRC07:39
*** zigo has joined #openstack-requirements07:42
*** zigo is now known as Guest2475607:42
*** Guest24756 is now known as zigo08:54
openstackgerritIhar Hrachyshka proposed openstack/requirements: Downgrade pycparser to 2.13  https://review.openstack.org/38101109:46
*** bauwser is now known as bauzas09:54
openstackgerrityolanda.robla proposed openstack/requirements: Downgrade pycparser to 2.13  https://review.openstack.org/38101409:57
openstackgerritIhar Hrachyshka proposed openstack/requirements: Downgrade pycparser to 2.13  https://review.openstack.org/38101110:09
*** ihrachys has joined #openstack-requirements10:14
ihrachyshowdy10:14
bauzasihrachys: aloha10:14
ihrachysI see we need something like https://review.openstack.org/#/c/380958/ too for swift10:15
ihrachysbut it does not work10:15
bauzasihrachys:  so, I'm paying attention to your change, that seems to be workarounding the situation at least for the nova unittests10:15
bauzasihrachys: before moving forward, do you know if there is any open bug against that?10:15
ihrachysbauzas: yeah, but we won't be able to land it without swift fixed :)10:15
ihrachysI don't know of a bug10:16
bauzasokay, filing one10:16
bauzasihrachys: oh, that's because they're not using upper-requirements.txt ? I see10:16
ihrachyshm, I look at swift unit test cross check logs, and I see that constraints do not reflect patch up for review. see:10:17
ihrachysRequirement already satisfied (use --upgrade to upgrade): pycparser===2.14 in ./.tox/py27/lib/python2.7/site-packages (from -c /home/jenkins/workspace/gate-swift-python27-ubuntu-xenial/upper-constraints.txt (line 262))10:17
ihrachyshttp://logs.openstack.org/58/380958/2/check/gate-swift-python27-ubuntu-xenial/4817576/tox/py27-2.log.txt10:17
ihrachysthat's for the patch that I pushed10:17
bauzasswift is co-gating with nova ?10:17
ihrachysso the constraints file used is probably from git, not the one up for review10:17
ihrachysbauzas: swift tests are used to validate upper-constraints.txt changes in requirements repo10:18
bauzasoh sad10:18
ihrachysbauzas: oh right, I get why swift fails. because it does NOT include my patch.10:18
ihrachysso basically, to pass my patch, I need swift's one; and to pass swift's one, I need mine10:18
ihrachysFUN10:18
ihrachysif that would be a single repo, we would squash and live our lives10:19
ihrachyswe could probably temporarily break the cogating thru project-infra10:19
bauzasihrachys: that's what I'm afraid :/10:19
ihrachysthat would probably make sense because it's not correct to gate constraints against swift now while they don't really co-gate back10:20
ihrachyslemme untangle them10:20
ihrachysbauzas: have a bug number?10:23
*** tdurakov has joined #openstack-requirements10:24
bauzasihrachys: just now https://bugs.launchpad.net/nova/+bug/162983010:25
openstackLaunchpad bug 1629830 in OpenStack Compute (nova) "pycparser-2.14 wheel is wrongly built and raises AssertionError" [Critical,Confirmed]10:25
ihrachysbauzas: tnx, I will add it to project-config patch10:25
bauzasI'll ask other projects to mention them in the bug report10:26
ihrachysbauzas: patch to disable cross check https://review.openstack.org/38103110:28
bauzastonyb: around ?10:28
bauzasihrachys: FWIW, I found the original bug report https://bugs.launchpad.net/cinder/+bug/162972610:38
openstackLaunchpad bug 1629726 in Cinder "cinder-manage db sync fails with recompiled pycparser 2.14" [Undecided,Confirmed]10:38
ihrachysack10:38
ihrachysI use yours for the gerrit topic already. I will change it if I need to respin10:38
bauzasihrachys: np, I duplicated it, so that's finre10:41
bauzasgiven matt wrote an e-r query against it, it was far better to keep it10:41
* bauzas bbiab - lunch10:44
*** jpich has joined #openstack-requirements10:44
openstackgerritIhar Hrachyshka proposed openstack/requirements: Downgrade pycparser to 2.13  https://review.openstack.org/38101110:46
ihrachysok I proposed backports with different change-ids and set depends-on accordingly: https://review.openstack.org/#/q/topic:bug/162983010:48
*** openstackgerrit has quit IRC10:48
*** openstackgerrit has joined #openstack-requirements10:49
openstackgerritIhar Hrachyshka proposed openstack/requirements: Downgrade pycparser to 2.13  https://review.openstack.org/38101111:18
* dims tuning into the pycparser discussion11:39
*** jrollen has joined #openstack-requirements11:40
dimsbauzas : ihrachys : so https://review.openstack.org/#/q/topic:bug/1629830 is the consensus so far? fix upper-constraints?11:41
ihrachysdims: it's either that, or upstream fixing their pypi11:42
ihrachysdims: we could also try to disable binary wheels for the package, but I don't know how to do it properly in gate context.11:42
bauzasdims: yup11:42
dimsAJaeger : agree? ^^ (is there any other toggles we can use)11:42
bauzasihrachys: the problem I see is that we're using u-c, not requirements.txt11:43
bauzasihrachys: we could maybe try to use the git context11:43
ihrachysbauzas: elaborate11:43
ihrachysdims: long term we could have our mirror of pypi where we could apply fixes like that (dropping bad wheels)11:43
ihrachysdims: but I don't think that will happen :)11:44
bauzasihrachys: something like http://stackoverflow.com/questions/16584552/how-to-state-in-requirements-txt-a-direct-github-source11:44
bauzasusing the git commit11:44
ihrachysbauzas: I think we forbid git links in requirements.txt11:44
ihrachysbauzas: there was some check in requirements gate for that11:44
* amrith joins pycparser discussion11:44
amrithg'morning11:44
dimsihrachys : just dropping the bad wheels from our pypi mirror?11:44
ihrachysbauzas: or maybe we just discouraged that, not sure11:44
bauzasihrachys: tbh, I'm not opiniated, I just provided another possbility11:45
ihrachysbauzas: I think there was a claim that those links will be broken in next pbr releases. not sure it happened. prolly not.11:45
bauzasihrachys: but yours is good to me11:45
dimsihrachys : our pypi mirror has the tar.gz, what if we just cleanup the wheels, not sure if bandersnatch will go fetch the wheels11:45
ihrachysdims: dunno. I preferred to stick to less hacky way - ban broken version11:45
bauzasihrachys: I just tried to think about other possibilities but not downgrading the package11:45
bauzasbut I'm fine if we drop it11:46
* dims not approving anything yet...11:47
amrithihrachys, is it merely banning pycparser 2.14?11:48
ihrachysamrith: that's my plan11:48
bauzasdims: FWIW btw. https://bugs.launchpad.net/nova/+bug/1629726/comments/211:48
openstackLaunchpad bug 1629726 in OpenStack DBaaS (Trove) "recompiled pycparser 2.14 breaks Cinder db sync and Nova UTs" [Critical,Confirmed]11:48
ihrachysdims: yeah, we need to wait for CI at least11:48
ihrachysso far it looks good11:48
ihrachysbut grenade hasn't voted yet11:48
bauzasthe main problem was with grenade yeah11:49
bauzasbecause ihrachys needed to Depends-On on the newton bacport11:49
bauzaswhich itself depends on the mitaka one :p11:50
bauzassooooo lovely :)11:50
dimsright11:51
ihrachysok gate-grenade-dsvm-neutron-ubuntu-xenial passed11:51
bauzason master ?11:52
ihrachysthere are some more voting jobs to go, but that's asuring11:52
ihrachysyes, master11:52
bauzas_o/11:52
bauzas\o/ even11:52
*** pallavi has joined #openstack-requirements12:01
dimsbauzas : ihrachys : which pkg currently/transitively pulls pycparser?12:02
ihrachysno idea12:03
amrithdims, I've just put pycparser<2.14 in all my requirements.txt files and am rerunning devstack, will let you kniw of that works. looking for the dependency right now.12:03
ihrachysdims: bauzas: amrith: all patches +1'd by jenkins12:09
amrithmy devstack is also going further with 2.13 hammered into upper-constraints.txt12:11
dimsihrachys : looks like pycparser change does not end up changing requirements files, so we won't have to re-release any artifacts12:13
dimsbauzas : amrith : agree with my evaluation?12:14
bauzasdims: sec, I found the package namez12:14
ihrachysdims: oh yeap. and I just made a silly comment on the thread regarding that one.12:14
bauzascfiii IRC12:14
bauzascffi rather12:14
dimsthat was my main concern12:14
dimsbauzas : that's probably from cryptography12:15
amrithdims, since this can be done with only a u-c change, no need to re-release everything. I agree.12:15
bauzaswell, I provided a stack in the bugreport12:15
bauzaseg. http://logs.openstack.org/25/379525/2/check/gate-nova-python27-db-ubuntu-xenial/026aba5/console.html12:15
bauzasdims: ^12:15
amrithand people who don't respect upper-constraints are, ... well.12:15
bauzasdims: in case of Nova, this is cffi which uses a backend parser being pycparser http://logs.openstack.org/25/379525/2/check/gate-nova-python27-db-ubuntu-xenial/026aba5/console.html#_2016-10-03_09_25_12_29429312:16
dimsk12:17
bauzasbut that can be different from any other stracktrace tho12:17
dimsany other requirements core folks around?12:17
bauzascryptography could be hurt as well, sure, if they share the same transitive dep12:17
dimstonyb ? coolsvap ? odyssey4me ?12:17
amrithhmm12:17
coolsvapdims: hello12:17
amrithwith 2.13 I have the same failure12:17
amrithare you folks sure that 2.13 would solve the problem?12:18
*** jrollen has quit IRC12:18
coolsvap2.13 looks good with requirements gates12:18
amrithsee: https://gist.github.com/anonymous/56538a9eaa9d432291c63c5c1816f2e312:18
dimsbauzas : ihrachys : if you make one of the nova reviews depend on this series, https://review.openstack.org/#/q/project:openstack/requirements++topic:bug/162983012:19
dimsbauzas : ihrachys : does it make it pass?12:19
ihrachysdims: I haven't checked THAT but I saw zuul failing grenade jobs before the latest version of the patches12:19
odyssey4medims I'm not q requirements core. I think perhaps you mean prometheanfire :)12:19
ihrachysdims: for requirements repo12:19
dimsodyssey4me : your vote counts too :)12:19
amrithand confirmed, I'm using pycparser 2.1312:20
amrithamrith@amrith-work:/opt/stack$ pip list  | grep pycparser12:20
amrithpycparser (2.13)12:20
amrithamrith@amrith-work:/opt/stack$12:20
dimsi am looking for company to help me push the button (and carry blame)12:20
dims:)12:20
bauzasdims: the problem is that we're pulling the requirements from an HTTP call12:20
amrithihrachys, dims, not working for me.12:20
odyssey4mepycparser does seem to have hit anyone - the sdist is fine, but the wheel is broken12:20
ihrachysamrith: no idea. what does pip freeze show in the venv?12:20
bauzasdims: I couldn't just depend-on it12:20
odyssey4meso we're working around it with --no-binary12:20
ihrachysok you answered already12:20
amriththis is stack.sh, no venv12:20
amrithI'm seeing failure in devstack12:20
dimsodyssey4me : ack. thanks12:20
ihrachysamrith: works for gate at least12:21
amrithmy venv based tests don't need this and they pass12:21
dimsamrith : ack, will let you poke at it12:21
bauzasdims: but I can take a faulty change and modify my venv so that I'm using the downgraded version12:21
amrithI will depend on your master change and float a trial baloon in trove12:21
bauzasdims: actually, I think tdurakov did that12:21
ihrachysok folks if you think it's bad direction after all I leave you to come up with a better fix :)12:21
ihrachysI am now going into meeting marathon for next hours.12:22
bauzasihrachys: I don't think dims is relucting to approve that, he just wants a proof of success :)12:22
dimsyep12:23
amrithdims, ihrachys see: https://review.openstack.org/#/c/381075/12:23
ihrachysfine with me. you have all the knobs to proceed. I just state that I am off for next hours12:23
bauzasdims: okay, let me then ack that for Nova12:23
bauzasdims: but I'll only test against our py27 job, not grenade12:24
bauzaswhich is borked too12:24
amrithdims, what's the plan?12:24
ihrachysbauzas: how come? another breakage for grenade?12:25
bauzasihrachys: nah nah, just explaining that in Nova, we have both py27, tempest and grenade which are broken12:25
ihrachysoh ok.12:26
ihrachysso you are lucky, you can trigger it with unit tests.12:26
bauzaszactly12:26
bauzasI don't have a grenade environment ready locally12:26
bauzasI could be testing Tempest tho12:26
bauzassince I have a local devstack with it there12:26
bauzasbut given in Nova, Tempest tests and UTs share the same trace, easier for me to just spin a local test with tox12:27
bauzasamrith: not sure your trial ballon will show something right12:29
AJaegerdims, ihrachys : Regarding https://review.openstack.org/#/c/381011 - IMHO we should block it also in global-requirments. That's what we did in the past. And then it works also for jobs that don't use upper-constraints...12:29
ihrachysAJaeger: right after you locate it in global-requirements12:29
amrithbauzas, it will ... see https://review.openstack.org/#/c/380985/12:29
ihrachysAJaeger: timer started :)12:29
amrithmore specifically http://logs.openstack.org/85/380985/1/check/gate-trove-functional-dsvm-mysql/48ae79c/logs/devstacklog.txt.gz#_2016-10-03_09_04_15_20012:30
amrithbauzas, ^^12:30
bauzasAJaeger: well, I tend to agree with you, but I think some projects will need to get an unnecessary sync12:30
bauzasAJaeger: for example, Nova is heavily using u-c, so we won't get something good at syncing something we don't really use12:31
AJaegerihrachys: Ah - that's something you should have mentioned in the commit message since it's not obvious.12:31
ihrachysAJaeger: agreed, but as I said in the email thread, I am bad at foresight. seemed obvious to me when I was producing the patches.12:31
bauzaswe *could* add that in global-reqs because that's a transitive dependency12:32
AJaegerihrachys: just came back online and read IRC first, so haven't seen your email12:32
bauzasbut like I said, it's IMHO too huge12:32
ihrachysbauzas: what's the point of having it there if no project uses it directly?12:33
AJaegerbauzas: if nova would have X in it's *requirements.txt files, we should add a "!=version" for breakage to document it. We give a range of working versions while u-c just explains what was tested.12:34
AJaegerbauzas: but since X is not in global-requirements, no need to add it now...12:34
bauzasAJaeger: yeah I agree, and ihrachys yes I agree with you, I'm just explaining we *could* do that, but that's not necessary12:35
AJaegerbauzas: agreed, it's not necessary ;)12:35
ihrachysI agree we all agree12:35
bauzas:D12:36
AJaegerihrachys: +1 ;)12:36
ihrachysnow let's join hands and hearts and voices and dance12:36
AJaeger;)12:37
* ihrachys sings kumbaya song12:37
* amrith mumbles something about counting chickens12:38
amrithbauzas, the trove gate seems happier with the change than my sandbox12:40
*** jrollen has joined #openstack-requirements12:41
bauzasdims: it fixes the nova UTs too (oh surprising it is :p ) http://paste.openstack.org/show/583989/12:42
bauzasdims: throw your axe !12:43
*** jpena is now known as jpena|lunch12:43
dimsbauzas : coolsvap : prometheanfire : dhellmann : ihrachys : so blessing these 4 - https://review.openstack.org/#/q/project:openstack/requirements++topic:bug/1629830+status:open12:44
dimstonyb : ^12:44
ihrachysamen12:44
coolsvapdims: i already have a +2 on master12:45
coolsvapwe need someone from stable team to ack on stable branch changes12:45
ihrachyscoolsvap: I am from stable ;)12:45
*** jrollen has quit IRC12:45
ihrachysand as per stable rules, you can push with a single +2 then ;)12:45
*** jrollen has joined #openstack-requirements12:45
coolsvapihrachys: i cannot +2 on stable changes12:46
ihrachysI mean, dims or someone with perm12:46
dimscoolsvap : ack thanks12:46
coolsvapi think dims done12:46
coolsvapdims: :)12:46
dimsoh boy, am i going to be in lot of trouble today :)12:46
* dims readies his fingers to point at co-conspirators12:46
ihrachyshaha. I will share the blame.12:46
coolsvapdims: lets see :)12:46
dimsihrachys : please drop a note to -dev. i gotta dash to wake up kids12:47
ihrachysack12:47
*** jroll has quit IRC12:53
*** jrollen is now known as jroll12:55
dhellmanndims : thanks!13:28
*** jpena|lunch is now known as jpena13:42
*** rtheis has joined #openstack-requirements13:57
openstackgerritMerged openstack/requirements: Downgrade pycparser to 2.13  https://review.openstack.org/38101113:57
jrollb 214:00
jrolloops, sorry14:00
prometheanfirehit14:02
prometheanfireyou sunk my battleship14:02
dimsLOL14:05
jroll\o/14:10
*** yatinkarel has joined #openstack-requirements14:17
*** mriedem has joined #openstack-requirements14:19
mriedemihrachys: thanks for the u-c patches14:19
* ihrachys bows14:20
dims++ ihrachys14:22
amrithihrachys, dims, bauzas ... the Trove balloon is good. https://review.openstack.org/#/c/381075/14:23
* amrith does not have glasses so can't read IRC :)14:23
bauzascool14:23
dimsamrith : cool14:24
*** pallavi has quit IRC14:25
openstackgerritDirk Mueller proposed openstack/requirements: Updated from generate-constraints  https://review.openstack.org/38070314:32
openstackgerritDirk Mueller proposed openstack/requirements: Updated from generate-constraints  https://review.openstack.org/38070314:40
amrithall hail dims14:46
dimsamrith : that would be ihrachys, bauzas and others who did the work before i even woke up :)14:47
amriththey too serve who take a nap (that's been my motto for a long time)14:47
ihrachysI accept bank notes14:47
amrithihrachys, dims took those14:48
ihrachysdammit14:48
dimshaha14:48
bauzasdims: nah, I just supported ihrachys :p14:49
bauzasall credits go to him :)14:49
* AJaeger throw some confusion in the mix and thanks everybody that fixed it14:50
dimsall's well that ends well :)14:55
amrithall hail ihrachys ... moveon.org petition has been launched to get bank notes back from dims15:03
AJaegerdims: until the next time ;)15:04
dimsawww15:04
amrithAJaeger, until next time what? dims takes a nap and ihrachys fixes the worlds problems?15:04
AJaegeramrith: the next broken package gets released. This kind of breakage will happen again - perhaps on a smaller level...15:06
mriedemfyi, the wheel on pycparser 2.14 was just removed15:17
mriedemhttps://pypi.python.org/pypi/pycparser15:17
mriedemso we could revert those u-c changes if we wanted to...15:17
mriedemat least on master15:17
bauzasmriedem: that sounds a reasonable thing15:29
bauzasreverting needs to be carefully done tho, given the depends-on15:30
prometheanfiremriedem: submit the review :D (amending the deps-on)15:31
prometheanfiregit revert -e I think15:31
openstackgerritMatt Riedemann proposed openstack/requirements: Revert "Downgrade pycparser to 2.13"  https://review.openstack.org/38119215:32
mriedemor just gerrit15:32
mriedembuttons are easy15:32
prometheanfiremriedem: with the wheel removed can we still install 2.14?15:34
prometheanfireguess it's still there15:34
mriedemyeah it will be the same as before i think15:34
prometheanfireya15:34
bauzasmriedem: do you want me to test again on my venv ?15:35
bauzasat least for the nova UT15:35
prometheanfireI'd think we should be fine, but never know15:36
mriedembauzas: sure15:36
bauzasokay, that should be fine in 5 mins or so15:37
prometheanfiredims: tonyb: what happened to the cross project gating we were doing?15:38
bauzasmriedem: prometheanfire: FWIW, that works (oooh surprising it is)15:42
dimsbauzas : prometheanfire : y since the wheel was nuked, our pypi mirror cleansed itself of the wheel, so now we can add that back15:46
bauzaskewl15:46
prometheanfiredims: yarp, assumed that's what was happening15:48
*** mriedem has quit IRC16:00
*** ihrachys has quit IRC16:19
*** jpich has quit IRC16:28
*** armax has joined #openstack-requirements17:00
*** armax has quit IRC17:05
*** armax has joined #openstack-requirements17:14
openstackgerritDirk Mueller proposed openstack/requirements: Updated from generate-constraints  https://review.openstack.org/38070317:19
*** jpena is now known as jpena|away17:46
*** strigazi is now known as strigazi_AFK17:53
*** AJaeger has left #openstack-requirements18:16
*** rtheis_ has joined #openstack-requirements19:25
*** rtheis has quit IRC19:26
openstackgerritMerged openstack/requirements: Revert "Downgrade pycparser to 2.13"  https://review.openstack.org/38119220:00
*** jpena|away is now known as jpena20:05
*** jpena is now known as jpena|off20:22
openstackgerritSteve Baker proposed openstack/requirements: Raise python-heatclient upper-constraint to 1.5.0  https://review.openstack.org/38130320:46
*** openstackgerrit has quit IRC21:19
*** openstackgerrit has joined #openstack-requirements21:19
*** rtheis_ has quit IRC21:21
* tonyb plays catchup23:43

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!