Thursday, 2018-01-04

clarkbEmilienM: ok left some comments00:04
EmilienMclarkb: thx, I'll take a look00:07
clarkbcorvus: ok ze04 is running an executor again I think the init script is working00:13
*** xarses_ has joined #openstack-infra00:16
mnaserfyi00:18
mnaserhttps://git.centos.org/commitdiff/rpms!kernel.git/6202b1a49906b21f0cdce08b4ff853a65c65875000:18
*** sdague has quit IRC00:18
mriedemdmsimard: another thing with the logs stuff is that you now have to download the config zips http://logs.openstack.org/67/529867/1/check/tempest-full/23d2919/controller/logs/etc/nova/00:19
mriedemassuming because devstack-gate's post run code isn't expecting the new directory structure00:20
clarkbmnaser: good to know, are those rpms published yet?00:20
mnaserclarkb: commit was only 90 minutes ago00:20
mnaseri pinged #centos-devel to ask, but im not sure how the internals of the centos builds work00:20
clarkbmnaser: sure but often times all that is staged for once embargo is lifted right? in any case I think our biggest risk is not actually on our centos machines00:21
clarkbits the trusty hypervisors in infra cloud00:21
mnaserwell it looks like the embargo was broken00:21
clarkbya it was00:21
mnaserand a poc is out unfortunately00:21
mnaserim not sure about ubuntu kernels and what their timelines are :(00:22
clarkbwe shouldn't run arbitrary code anywhere but single use VMs and transitively on our hypervisors for infra cloud00:22
clarkbmnaser: ya no usn for it yet that I see00:22
mnaseri mean the centos stuff wasnt even announced00:23
mnaseri was just digging around00:23
clarkbthat and our laptops00:23
clarkbor wherever you run javascripts and the like00:23
clarkb(though I don't think there is a poc via that vector (yet)00:24
*** daidv has quit IRC00:24
mnaserhttp://kernel.ubuntu.com/git/ubuntu/ubuntu-trusty.git/log/00:25
mnasernothing there unfortunately00:25
openstackgerritEmilien Macchi proposed openstack-infra/system-config master: Add Puppetlabs mirror for Ubuntu Xenial  https://review.openstack.org/53103000:26
openstackgerritMerged openstack-infra/storyboard-webclient master: Remove bower and add yarn support  https://review.openstack.org/52843600:27
mnaserhttps://www.redhat.com/archives/rhsa-announce/2018-January/thread.html00:30
mnaseri guess canonical should announce it anytime now00:30
*** claudiub has quit IRC00:41
*** caphrim007 has quit IRC00:44
openstackgerritKendall Nelson proposed openstack-infra/storyboard master: Remove Update tox Step  https://review.openstack.org/53042700:46
*** xarses_ has quit IRC00:47
*** gothicmindfood has quit IRC00:47
*** ijw has quit IRC00:50
armaxhi folks, how long does it take for change https://review.openstack.org/#/c/530915/ to become effective? I rechecked on https://review.openstack.org/#/c/530857/ but after over 6 hours the job still seems to run on stable branches, just idle curiosity00:50
armaxthanks00:50
clarkbarmax: the job being tempest-full?00:51
armaxyes00:51
armaxclarkb: hi and happy new year!00:51
clarkbarmax: hello!00:52
clarkbarmax: changes like that should apply immediately when merged as zuul is aware it merged and can apply it immediately00:52
clarkbarmax: my guess is something else is making it not work, will look at it00:52
armaxI still see it on http://zuulv3.openstack.org/ and change 53085700:52
armaxalso, I don’t seem to see the tempest.conf/log files being collected anymore I was tracing andreaf’s recent changes around log collection00:53
armaxbut couldn’t spot anything obvious00:53
clarkbthe regex seems to work as expected (just testing in python interpreter)00:55
armaxclarkb: yeah, I was looking at that00:55
clarkbmy next guess is that some other variant is applying and overriding the job selection00:58
clarkbI think it logs that stuff for us /me looks00:59
*** cuongnv has joined #openstack-infra00:59
armaxclarkb: right, but looking at http://codesearch.openstack.org/?q=tempest-full&i=nope&files=&repos= I see no smoking gun01:00
armaxassuming I’d be looking in the right place01:00
clarkbarmax: http://logs.openstack.org/57/530857/1/check/tempest-full/6a17ce3/zuul-info/inventory.yaml has the path under inheritance_path01:00
armaxah, that’s good to know01:00
clarkbarmax: the top of the list is the base job and then the bottom is the last config for the job01:00
armaxpretty cool01:01
armaxoh01:01
armaxtempest being branchless maybe that’s the source of our problem01:01
clarkbarmax: I think I see it01:02
clarkbarmax: project-config/projects.yaml line 1034801:02
clarkbarmax: that defines tempest-full there with irrelevant files but no branch matchers01:02
clarkbI think if you add the branch matchers to that too it will do what you expect01:02
armaxoh01:03
armaxI can give it a go01:03
armaxthough, I wasn’t expecting it to behave like that01:03
armaxpatch coming right up01:05
armaxI guess I’d have to do the same for the gate part01:14
clarkbya anywhere you are making a new variant will need it I Think01:14
armaxaye01:14
clarkbsince this isn't inheritance its a different definition01:14
openstackgerritArmando Migliaccio proposed openstack-infra/project-config master: Make sure we only run tempest-full on master  https://review.openstack.org/53104501:14
*** bandini has quit IRC01:14
*** fanzhang has left #openstack-infra01:15
*** lastmikoi has quit IRC01:17
*** mpjetta has quit IRC01:17
*** Swami has quit IRC01:17
*** bandini has joined #openstack-infra01:17
openstackgerritEmilien Macchi proposed openstack-infra/system-config master: Add Puppetlabs mirror for Ubuntu Xenial  https://review.openstack.org/53103001:19
*** mpjetta has joined #openstack-infra01:19
armaxclarkb: I hope I got it right, thanks for the tip01:20
*** aviau has joined #openstack-infra01:21
*** lastmikoi has joined #openstack-infra01:23
corvusremote:   https://review.openstack.org/531046 Only run tempest-full on the master branch01:27
corvusarmax, clarkb, andreaf: ^ that may be a simpler alternative01:27
armaxcorvus: indeed01:28
corvussince the current issue is fundamental to the job, if we can keep the fix as close to that as possible, it should be easier to unwind later01:28
corvusarmax: can you make a change Depends-On: the one i just pushed up to verify that fixes it for yoU?01:28
armaxcorvus: sure01:29
armaxcorvus: looks like that did the trick01:32
armaxI don’t see the job running for 53104801:33
armaxcorvus: does this mean that https://review.openstack.org/#/c/530915/1/zuul.d/zuul-legacy-project-templates.yaml could be partially reverted?01:34
armaxat least for the parts that touch the branches for tempest-full01:34
corvusarmax: yes i think so01:49
armaxcorvus: OK, let me see if I can earn some brownie points :)01:49
*** smatzek has joined #openstack-infra01:53
*** threestrands has joined #openstack-infra01:55
*** smatzek has quit IRC01:57
*** kaisers has quit IRC01:58
openstackgerritMatt Riedemann proposed openstack-infra/project-config master: Don't run legacy-tempest-dsvm-neutron-full on non-code nova changes  https://review.openstack.org/53105202:10
*** smcginnis has quit IRC02:13
*** zhurong has joined #openstack-infra02:14
openstackgerritTristan Cacqueray proposed openstack-infra/nodepool feature/zuulv3: license: remove dos line break  https://review.openstack.org/53105702:17
mnaserDo we wanna drop +r?02:18
corvusmnaser: ++02:19
openstackgerritArmando Migliaccio proposed openstack-infra/openstack-zuul-jobs master: Restore branch conditional on tempest-full  https://review.openstack.org/53105802:19
armaxcorvus: ^02:20
*** rlandy|bbl is now known as rlandy02:20
openstackgerritshangxdy proposed openstack-infra/gear master: Modify connection timeout process  https://review.openstack.org/53105902:22
*** rlandy has quit IRC02:23
*** ijw has joined #openstack-infra02:26
*** rfolco has quit IRC02:29
*** ijw has quit IRC02:30
*** bobh has joined #openstack-infra02:36
*** namnh has joined #openstack-infra02:47
*** bobh has quit IRC02:53
*** RuiChen has joined #openstack-infra03:01
*** kiennt26 has joined #openstack-infra03:10
*** ijw has joined #openstack-infra03:14
*** coolsvap has joined #openstack-infra03:26
yamamotois there an equivalent of stats.timers.nodepool.job.xxx these days?03:31
*** mriedem has quit IRC03:48
*** xarses_ has joined #openstack-infra04:08
openstackgerritMerged openstack-dev/pbr master: Deprecate testr and nose integration  https://review.openstack.org/51822404:09
*** lbragstad has quit IRC04:11
*** gyee has quit IRC04:29
*** armax has quit IRC04:32
*** armax has joined #openstack-infra04:33
*** armax has quit IRC04:33
*** armax has joined #openstack-infra04:33
*** armax has quit IRC04:34
*** armax has joined #openstack-infra04:34
*** armax has quit IRC04:34
*** armax has joined #openstack-infra04:35
*** armax has quit IRC04:35
*** udesale has joined #openstack-infra04:37
*** nicolasbock has quit IRC04:53
*** ramishra has joined #openstack-infra05:03
*** yangzhenyu has joined #openstack-infra05:08
openstackgerritMerged openstack-infra/project-config master: Remove devstack-gate specific tox jobs  https://review.openstack.org/52855405:10
openstackgerritMerged openstack-infra/openstack-zuul-jobs master: Only copy openafs centos rpms if we built them  https://review.openstack.org/53101805:13
*** ijw has quit IRC05:13
*** ramishra has quit IRC05:20
*** pots has quit IRC05:21
openstackgerritMerged openstack-infra/openstack-zuul-jobs master: Remove converted devstack-gate tox jobs  https://review.openstack.org/52855505:26
*** ramishra has joined #openstack-infra05:39
*** janki has joined #openstack-infra05:43
*** eumel8 has joined #openstack-infra05:50
*** markmcclain has quit IRC05:51
*** markmcclain has joined #openstack-infra05:54
openstackgerritYAMAMOTO Takashi proposed openstack-infra/project-config master: grafana: Update networking-midonet  https://review.openstack.org/53022005:57
openstackgerritOpenStack Proposal Bot proposed openstack-infra/project-config master: Normalize projects.yaml  https://review.openstack.org/53107806:06
*** swest has joined #openstack-infra06:14
*** armaan has quit IRC06:21
*** hongbin has joined #openstack-infra06:22
*** hongbin has quit IRC06:22
*** dbecker has quit IRC06:29
openstackgerritMerged openstack-infra/project-config master: Normalize projects.yaml  https://review.openstack.org/53107806:34
*** 07EAAP3JW has joined #openstack-infra06:43
*** 07EAAP3JW has quit IRC06:43
*** hongbin has joined #openstack-infra06:43
*** zhurong has quit IRC06:47
*** zhurong has joined #openstack-infra06:52
*** jaosorior has quit IRC07:01
*** ramishra has quit IRC07:10
openstackgerritMerged openstack-infra/zuul feature/zuulv3: Strip \r from build UUID in fingergw  https://review.openstack.org/53097307:10
*** jaosorior has joined #openstack-infra07:11
*** ramishra has joined #openstack-infra07:12
*** jaosorior has quit IRC07:13
*** jaosorior has joined #openstack-infra07:14
*** jaosorior has quit IRC07:14
*** jaosorior has joined #openstack-infra07:14
*** sbezverk has quit IRC07:21
*** benj_ has quit IRC07:22
*** threestrands has quit IRC07:25
*** annp has joined #openstack-infra07:26
*** jrist has quit IRC07:29
*** pcaruana has joined #openstack-infra07:32
*** AJaeger has quit IRC07:44
*** AJaeger has joined #openstack-infra07:47
*** jaypipes has quit IRC07:53
*** markvoelker has quit IRC07:55
*** rcernin has quit IRC08:04
*** shardy has joined #openstack-infra08:16
*** sshnaidm|afk is now known as sshnaidm08:28
*** liusheng has quit IRC08:30
*** ramishra has quit IRC08:38
*** dtantsur|afk is now known as dtantsur08:49
*** lucas-afk is now known as lucasagomes08:49
*** mnencia has quit IRC08:56
*** mnencia has joined #openstack-infra08:56
*** jpich has joined #openstack-infra09:02
*** larainema has quit IRC09:03
*** hongbin has quit IRC09:03
*** e0ne has joined #openstack-infra09:05
*** jrist has joined #openstack-infra09:13
*** mnencia has quit IRC09:19
*** mnencia has joined #openstack-infra09:20
*** derekh has joined #openstack-infra09:24
stephenfinMorning o/ Would anyone be able to advise me on what's happening here with the openstack-tox-functional + privsep test? https://review.openstack.org/#/c/530961/09:33
fricklerstephenfin: iiuc sudo access is removed for tox tests, so starting the privsep daemon fails here http://logs.openstack.org/61/530961/1/check/openstack-tox-functional/b033cc9/job-output.txt.gz#_2018-01-03_16_51_01_31452509:36
stephenfinfrickler: Ah, it's a tox thing09:37
* stephenfin initially thought permissions issues, but then realized you'd need sudo for DevStack stuff09:37
*** tosky has joined #openstack-infra09:38
AJaegerstephenfin: mordred has been working on some tox tests with sudo enabled - best discuss with him the status there and whether this would help your case09:38
stephenfinAJaeger: Sounds like a good call. Yet more work for mordred 🎉09:41
*** kiennt26 has quit IRC09:49
openstackgerritMerged openstack-infra/openstack-zuul-jobs master: Remove integrated-gate-nova-net  https://review.openstack.org/52985809:55
*** markvoelker has joined #openstack-infra09:56
*** greghaynes has quit IRC09:58
*** sshnaidm is now known as sshnaidm|afk10:00
*** erlon has joined #openstack-infra10:00
*** cuongnv has quit IRC10:06
*** pbourke has quit IRC10:10
*** annp has quit IRC10:12
*** toabctl has quit IRC10:16
*** ijw has joined #openstack-infra10:17
*** ijw has quit IRC10:21
*** toabctl has joined #openstack-infra10:23
danpawlikdoes soneome know what is the new kernel for Ubuntu xenial, trusty with a patch for Meltdown and Spectre?10:26
*** namnh has quit IRC10:27
*** yangzhenyu has quit IRC10:30
*** markvoelker has quit IRC10:30
openstackgerritMerged openstack-infra/openstack-zuul-jobs master: Remove legacy-irc-meetings-tox-ical  https://review.openstack.org/52850110:31
openstackgerritMerged openstack-infra/openstack-zuul-jobs master: Remove legacy swift tox jobs  https://review.openstack.org/52865010:31
fricklerdanpawlik: still pending, see https://wiki.ubuntu.com/SecurityTeam/KnowledgeBase/SpectreAndMeltdown10:46
openstackgerritJens Harbott (frickler) proposed openstack-infra/openstack-zuul-jobs master: Remove obsolete interop jobs  https://review.openstack.org/52864510:48
chandankumarAJaeger: Hello10:49
chandankumarAJaeger: Please point me an example for common template for writting python-tempest-plugin jobs10:50
AJaegerchandankumar: check openstack-zuul-jobs/zuul.d/project-templates.yaml10:50
chandankumarAJaeger: sure!10:50
*** markmcclain has quit IRC10:51
chandankumarAJaeger: i think we can merge this one https://review.openstack.org/#/c/528674/ then i will create the common template and update in one go, what you say?10:51
AJaegerfrickler: Remove legacy-irc-meetings-tox-ical was jsut merged - I'll update the introop change10:51
AJaegerchandankumar: works for me10:52
fricklerAJaeger: ah, you are right, that one needs to be removed, too, sorry for mixing that up10:52
chandankumarfrickler: needs one +2 and +w https://review.openstack.org/#/c/528674/10:52
chandankumarAJaeger: frickler Thanks :-)10:53
openstackgerritAndreas Jaeger proposed openstack-infra/openstack-zuul-jobs master: Remove obsolete interop jobs  https://review.openstack.org/52864510:53
AJaegerfrickler: thanks for trying to fix it ;) new version ^10:53
AJaegerfrickler: because of those merge conflicts, I'd like to get those changes in and through our backlog. Thanks for helping to reduce it. Care to review https://review.openstack.org/#/c/530240/ as well, please?10:54
*** larainema has joined #openstack-infra10:57
*** markmcclain has joined #openstack-infra10:59
AJaegerwe should ask asettle to add pkovar ...11:00
AJaegerstephenfin: see https://docs.openstack.org/infra/system-config/irc.html#basic-channel-operator-commands11:01
AJaegerwrong channel ;(11:01
frickleruh oh, I think I approved stuff too fast, now zuul starts swapping ... :-(11:01
fricklerhope it will get itself together in a bit11:01
openstackgerritSam Betts proposed openstack-infra/devstack-gate master: Move location of devstack-gate unstack to post-tempest  https://review.openstack.org/53072611:01
fricklerinfra-root: zuulv3.o.o web service isn't responding for me and http://cacti.openstack.org/cacti/graph.php?action=view&local_graph_id=63979&rra_id=all looks like memory usage exploded. I'll try to take a look at the logs on the node11:15
openstackgerritAdam Spiers proposed openstack-infra/infra-manual master: promote IRC channel creation info into its own section  https://review.openstack.org/53113311:16
*** nicolasbock has joined #openstack-infra11:17
fricklerseems this is the patch that got merged last and caused memory usage to increase. not sure whether there is anything specific to that patch, or just that the n-th reconfiguration was too much for zuul https://review.openstack.org/51341811:24
frickler#status notice zuul seems to have gotten stuck and will probably need a restart, please be patient11:26
openstackstatusfrickler: sending notice11:26
*** kopecmartin has joined #openstack-infra11:26
-openstackstatus- NOTICE: zuul seems to have gotten stuck and will probably need a restart, please be patient11:27
*** markvoelker has joined #openstack-infra11:27
fricklerinfra-root: /me will go and have lunch and give zuul a bit of time to recover, planning to restart in half an hour unless the situation improves11:27
*** openstackstatus has quit IRC11:28
*** openstack has quit IRC11:28
*** openstack has joined #openstack-infra13:08
*** ChanServ sets mode: +o openstack13:08
mnaserbot is late for work today at 8:08 am :>13:08
mordredfrickler: fwiw, if you're talking with folks in #freenode, corvus put together a plan yesterday to get us to a better consolidated bot situation13:09
AJaegermordred: apparently not anymore, see 11:28 the killed by Sigyn ;(13:09
*** openstackstatus has joined #openstack-infra13:09
*** ChanServ sets mode: +v openstackstatus13:09
mordredAJaeger: yah - it's possible it's just the openstack bot that's on the whitelist and not statusbot13:09
mordred\o/13:10
* mordred waves to openstackstatus13:10
openstackgerritDmitry Tantsur proposed openstack-infra/project-config master: Update jobs for ironic-tempest-plugin  https://review.openstack.org/52952813:11
dtantsurAJaeger: removed controversial bit ^^^13:11
AJaegerdtantsur: not controversial - I just don't know yet how to do it ;( And nobody digged into it yet...13:12
* frickler is talking to sigyn admin now, do we have a list of all our bots somewhere?13:12
dtantsurAJaeger: somebody will have to - it's blocking creating of new projects that publish to pypi13:13
openstackgerritDmitry Tantsur proposed openstack-infra/project-config master: Publish ironic-tempest-plugin releases to pypi  https://review.openstack.org/53114813:13
dtantsurand then we can merge this ^^^13:13
AJaegerdtantsur: LGTM13:13
mordredfrickler: openstack, openstackstatus and openstackgerrit - and lemme check what accessbot is using13:13
AJaegerdtantsur: the first one I mean - didn't see the second one yet13:13
dtantsurk13:13
dtantsurthe 2nd one can wait13:13
AJaegerdtantsur: do you really need the plugin on pypi? chandankumar, did we put tempest plugins to pypi?13:14
mordredfrickler: although the plan corvus wrote up yesterday will consolidate us down to just the openstack bots assocated with the 'openstack' nickserv account13:14
AJaegerdtantsur: we should solve it anyhow - just curious...13:14
dtantsurAJaeger: we currently recommend pip install in our docs :) since it's branchless, I don't see big problems with people using it from pypi in a venv13:14
AJaegerdtantsur: ah, then you need to release it as well...13:14
toskyfrickler: re failures on zuul, should we wait a bit before rechecking?13:14
dtantsurAJaeger: in any case, this can wait13:15
openstackgerritAndreas Jaeger proposed openstack-infra/project-config master: Add neutron-vpnaas docs and release notes pub job to zuul layout  https://review.openstack.org/52269513:15
fricklertosky: I think things should be mostly fine now, maybe unless you are making changes to zuul.yaml rechecking should be fine13:16
evrardjpmordred: frickler ptgbot?13:16
toskyfrickler: ack, thanks13:16
evrardjpthinking ahead just in case13:16
chandankumarAJaeger: currently we have not putted there but it would be good to put there13:16
evrardjpand happy new year to those I haven't said it yet.13:17
fricklerinfra-root: so according to the sigyn op we should be fine now bot-wise. just tell them when we invite sigyn to more channels, probably dmsimard did that yesterday?13:17
mordredfrickler: accessbot is openstackinfra13:18
mordredfrickler: ah - cool13:19
Shrewscan an infra-root make my workday have less typing by +3'ing https://review.openstack.org/530918 for me?13:20
*** trown|outtypewww is now known as trown13:21
AJaegerShrews, infra-root, I think we first need to restart Zuul - I'm seeing too many stuck changes and no move forward13:21
ShrewsAJaeger: oh that's fun13:21
fricklerAJaeger: oh, indeed, swapping activity has returned, it looked like it had stabilized until 10 minutes ago13:22
fricklerstill, looking at the time of the day, I'd say we can wait for corvus or maybe fungi to show up and decide13:25
openstackgerritMerged openstack-infra/project-config master: Added initial jobs for blazar-tempest-plugin  https://review.openstack.org/52867413:25
openstackgerritMerged openstack-infra/project-config master: Add Zuul project layout for ara-web  https://review.openstack.org/52836013:25
openstackgerritMerged openstack-infra/project-config master: Add noop job for django_openstack_auth master branch  https://review.openstack.org/52932113:25
AJaegerwow, some merges ^13:25
*** dhill_ has joined #openstack-infra13:30
openstackgerritAndreas Jaeger proposed openstack-infra/project-config master: Remove legacy job for python-heatclient  https://review.openstack.org/51357813:32
*** lucas-hungry is now known as lucasagomes13:32
*** rosmaita has joined #openstack-infra13:39
*** dtantsur is now known as dtantsur|brb13:49
sshnaidm|afkwe see a lot of post_failure in jobs, is it known issue?13:58
*** sshnaidm|afk is now known as sshnaidm|mtg13:58
*** bobh has joined #openstack-infra13:58
arxcruz|ruckmtreinish: around ?14:03
arxcruz|ruckmtreinish: I'm noticing some tripleo jobs are not showing up in openstack-health, wondering if something has changed, in order to have status updated there14:04
arxcruz|ruckwe do have the testrepository.subunit file14:04
*** bobh has quit IRC14:08
openstackgerritStibbons proposed openstack-dev/pbr master: Support v<semver> version  https://review.openstack.org/53116114:18
*** jokke_ has quit IRC14:21
*** hjensas has joined #openstack-infra14:22
*** dtantsur|brb is now known as dtantsur14:22
fricklersshnaidm|mtg: yes, zuul has some memory issues (again), please bear with us for a bit14:25
*** dansmith has quit IRC14:25
*** smatzek has quit IRC14:25
AJaegerinfra-root, anybody in the US already awake? Do you want to restart zuul - or wait and debug? We have so many timeouts that it's unusable IMHO...14:26
*** hongbin_ has joined #openstack-infra14:27
*** therve_ is now known as therve14:29
*** mriedem has joined #openstack-infra14:29
*** hongbin_ has quit IRC14:29
*** hongbin has joined #openstack-infra14:30
fricklerAJaeger: o.k., I'm convinced, proceeding to save the queues in preparation for a restart now14:30
fricklerinfra-root: ^^14:30
*** kgiusti has joined #openstack-infra14:31
*** rosmaita has quit IRC14:32
*** gouthamr has joined #openstack-infra14:34
fricklerinfra-root: failed to receive any output for the queues, restarted zuul-scheduler anyway now14:37
*** bobh has joined #openstack-infra14:38
*** esberglu has joined #openstack-infra14:39
*** yamamoto has quit IRC14:43
*** yolanda has quit IRC14:45
fricklero.k., zuul-web seems to have needed a restart, too. /me is curious what the bots will do when I send the next status notice now14:45
frickler#status notice zuul has been restarted, all queues have been reset. please recheck your patches when appropriate14:45
openstackstatusfrickler: sending notice14:45
*** xarses_ has quit IRC14:46
-openstackstatus- NOTICE: zuul has been restarted, all queues have been reset. please recheck your patches when appropriate14:47
openstackstatusfrickler: finished sending notice14:48
*** lbragstad has joined #openstack-infra14:54
*** eharney has joined #openstack-infra14:54
*** ramishra has joined #openstack-infra14:56
*** esberglu has quit IRC14:57
openstackgerritMerged openstack-infra/project-config master: Remove legacy job for python-heatclient  https://review.openstack.org/51357815:00
openstackgerritMerged openstack-infra/project-config master: Add neutron-vpnaas docs and release notes pub job to zuul layout  https://review.openstack.org/52269515:01
*** yamamoto has joined #openstack-infra15:01
*** smatzek has joined #openstack-infra15:01
*** smatzek has quit IRC15:02
*** smatzek has joined #openstack-infra15:02
openstackgerritMatt Riedemann proposed openstack-infra/project-config master: Run legacy-tempest-dsvm-nova-libvirt-kvm-apr in devstack experimental  https://review.openstack.org/53117015:10
openstackgerritMonty Taylor proposed openstack-infra/zuul feature/zuulv3: WIP Rework log streaming to use logging  https://review.openstack.org/53117115:10
openstackgerritAndreas Jaeger proposed openstack-infra/openstack-zuul-jobs master: Remove legacy jobs for python-heatclient  https://review.openstack.org/51357915:17
ttxmordred: just saw your message. I think going through Jean-Daniel is still a good way15:21
mtreinisharxcruz|ruck: if subunit data is getting picked up and written to the subunit2sql db then the jobs will show up in o-h15:21
thingeehey can any infra core do a quick review for the success and thanks bot so I can get a super user post out please https://review.openstack.org/#/c/510699/315:21
arxcruz|ruckmtreinish: so, how can I check if it's being picked up? because we had a lot of failures recently that are not showing up in o-h15:21
mtreinisharxcruz|ruck: https://github.com/openstack-infra/system-config/blob/master/modules/openstack_project/files/logstash/jenkins-log-client.yaml#L6-L2715:22
arxcruz|ruckmtreinish: I assume testrepository.subunit.gz is caught also right ?15:23
mtreinisharxcruz|ruck: yeah it trys the file and the file +.gz15:23
arxcruz|ruckmtreinish: thanks, i'll check on my side15:24
*** smatzek has quit IRC15:24
mordredttx: ah! thanks for reminding me15:25
openstackgerritMatt Riedemann proposed openstack-infra/elastic-recheck master: Add query for cinder CapacityFilter fail bug 1741275  https://review.openstack.org/53117515:25
openstackbug 1741275 in Cinder "CapacityFilter intermittently fails in the gate due to insufficient virtual space on the host" [Undecided,Confirmed] https://launchpad.net/bugs/174127515:25
mordredinfra-root: jean-daniel repsponded, our OVH account should be fixed now15:26
dmsimarddo we know what was the problem ?15:26
pabelangerfrickler: I am going to try and help on the zuul memory front, will be a good learning experience, I think corvus wanted to do that before bumping to 30GB15:26
pabelangermordred: ack15:26
*** smatzek has joined #openstack-infra15:26
pabelangerhttps://review.openstack.org/530945 reverts clouds.yaml removal of OVH15:27
*** smatzek_ has joined #openstack-infra15:27
*** smatzek_ has quit IRC15:28
*** smatzek_ has joined #openstack-infra15:28
AJaegermordred: did you test that the servers are up again? Can we now merge https://review.openstack.org/#/c/530818/ to enable it?15:29
fricklerpabelanger: I just not sure whether it will be possible to make short-term progress there and we have the end of the cycle coming up, so upping the flavor might still be a good intermediate solution15:30
AJaegerpabelanger, dmsimard, could either of you do quick cleanup review for devstack-gate, please? https://review.openstack.org/#/c/52881715:31
*** smatzek has quit IRC15:31
dmsimard+315:32
pabelangerAJaeger: we first need 530945 for OVH revert15:32
* frickler goes to enjoy rush hour traffic now, bbl15:32
dmsimardI'll be going through my review dashboard today btw15:32
dmsimardpabelanger: let's wait on mordred's ack from AJaeger's question15:32
*** smatzek_ has quit IRC15:33
AJaegerthanks, dmsimard15:33
fungimordred: echoing dmsimard's question... did jd say what went wrong with the account?15:33
AJaegerpabelanger: ok, will wait with +2...15:33
openstackgerritMerged openstack-infra/system-config master: Update SSH key for shrews  https://review.openstack.org/53091815:35
*** smatzek has joined #openstack-infra15:36
*** smatzek has quit IRC15:38
*** AJaeger has quit IRC15:40
*** xarses_ has joined #openstack-infra15:42
*** AJaeger has joined #openstack-infra15:44
*** eharney has quit IRC15:46
openstackgerritMerged openstack-infra/elastic-recheck master: Add query for cinder CapacityFilter fail bug 1741275  https://review.openstack.org/53117515:50
openstackbug 1741275 in Cinder "CapacityFilter intermittently fails in the gate due to insufficient virtual space on the host" [Undecided,Confirmed] https://launchpad.net/bugs/174127515:50
*** smatzek has joined #openstack-infra15:51
*** smatzek has quit IRC15:55
*** armax has joined #openstack-infra15:56
*** smatzek has joined #openstack-infra15:57
*** smatzek_ has joined #openstack-infra16:00
*** hemna_ has joined #openstack-infra16:01
*** udesale has joined #openstack-infra16:01
*** smatzek has quit IRC16:01
*** smatzek_ has quit IRC16:05
*** smatzek has joined #openstack-infra16:06
*** smatzek has quit IRC16:07
mordredfungi, dmsimard: he did not - he said they were going to look in to it16:08
*** smatzek has joined #openstack-infra16:08
mordredfungi, dmsimard: also, fwiw, he said to not hesitate to contact him - so I thnk he's still our contact there16:08
mordredAJaeger, dmsimard, pabelanger: I have not verified that anything works yet16:09
dmsimardmordred: so we can enable nodepool ? did you confirm it was okay ?16:09
dmsimardok16:09
mordreddmsimard: sorry, about 5 items deep in my morning stack atm - http://mirror.bhs1.ovh.openstack.org/pypi/ seems to be working16:10
mordredhttp://mirror.gra1.ovh.openstack.org/pypi/ not as much16:11
dmsimardpossible he fixed bhs1 but not gra1 ?16:11
openstackgerritMike Perez proposed openstack-infra/puppet-statusbot master: Add configuration for thanks/success page url  https://review.openstack.org/53118516:13
*** udesale has quit IRC16:14
*** edmondsw has joined #openstack-infra16:15
*** smatzek has quit IRC16:15
*** smatzek_ has joined #openstack-infra16:15
*** kopecmartin has quit IRC16:16
*** smatzek_ has quit IRC16:16
openstackgerritMike Perez proposed openstack-infra/system-config master: Set the success/thanks page url for status bot  https://review.openstack.org/53118716:16
*** smatzek has joined #openstack-infra16:22
*** snuffkin has quit IRC16:22
*** eharney has joined #openstack-infra16:24
*** nicolasbock has quit IRC16:24
*** nicolasbock has joined #openstack-infra16:25
*** smatzek has quit IRC16:27
openstackgerritMike Perez proposed openstack-infra/statusbot master: Thanks & Success bot provide confirmation site url  https://review.openstack.org/51069916:27
*** smatzek has joined #openstack-infra16:30
openstackgerritMiguel Lavalle proposed openstack-infra/project-config master: Update Neutron legacy jobs definition for master  https://review.openstack.org/53050016:32
*** smatzek_ has joined #openstack-infra16:34
openstackgerritGhanshyam Mann proposed openstack-infra/openstack-zuul-jobs master: Skip tempest 35 job for doc and reno only changes  https://review.openstack.org/53119416:35
*** smatzek has quit IRC16:35
*** e0ne has quit IRC16:36
openstackgerritGhanshyam Mann proposed openstack-infra/openstack-zuul-jobs master: Skip legacy-grenade-dsvm-neutron job for doc and reno only changes  https://review.openstack.org/53119516:40
corvusdmsimard: do you want to remove +r ?16:41
*** dtantsur is now known as dtantsur|afk16:41
dmsimardyes, I can do that16:41
dmsimardtesting something first16:42
*** ramishra has quit IRC16:45
dmsimardoh yeah I remember now.. ugh16:45
dmsimardsince you can't change modes with chanserv, I have to actually run mlock -r, and then an empty mlock.16:45
dmsimardif you don't see me talking for the next 6 minutes you know why :D16:46
*** ChanServ sets mode: -r 16:48
*** gcb has quit IRC16:49
*** gcb has joined #openstack-infra16:50
clarkbdmsimard: re gra1 mirror also possoble they vouldnt recover the instance after whatever happened16:52
*** pcaruana has quit IRC16:52
openstackgerritDavid Shrewsbury proposed openstack-infra/zuul feature/zuulv3: Allow Ansible 2.4  https://review.openstack.org/53100916:52
openstackgerritDavid Shrewsbury proposed openstack-infra/zuul feature/zuulv3: Don't remove builds that don't exist  https://review.openstack.org/53119816:52
clarkbwe can probably go ahead and reenable bhs1 then sort out gra116:52
*** smatzek_ has quit IRC16:53
pabelangerwfm16:54
fungiyeesh, the irc spammers are mostly gone from freenode at the moment, but are picking up activity levels on oftc now16:54
dmsimardyuck ..16:55
dmsimardsomeone has too much time on their hands16:55
fungigot a few on the gnome community's irc network as well16:55
*** sdague has joined #openstack-infra16:56
*** smatzek_ has joined #openstack-infra16:56
*** jpich has quit IRC16:56
*** danpawlik has quit IRC16:56
*** danpawlik has joined #openstack-infra16:57
Shrewswhat we need is a paywall16:57
Shrewsdeposited directly to my bank account, of course16:57
dmsimardmirror01.gra1.ovh.openstack.org shows up as SUSPENDED, I'll try and figure that out16:59
openstackgerritMerged openstack-infra/devstack-gate master: Remove run-tests and py3-run-tests  https://review.openstack.org/52881716:59
*** smatzek_ has quit IRC17:00
dmsimardwe have ourselves a mirror: http://mirror.gra1.ovh.openstack.org/pypi/17:01
pabelangerdmsimard: clarkb: https://review.openstack.org/530945/ will add back OVH to clouds.yaml so we can start poking at APIs17:01
pabelangernice17:01
dmsimarduptime is showing up as 116 days o_O17:03
fungisuspended huh?17:03
*** esberglu has joined #openstack-infra17:03
dmsimardso it was really just suspended17:04
fungineat17:04
fungiresume ftw!17:04
dmsimardI had to do a double take on osc because I expected "unsuspend" but nope17:04
clarkbdmsimard: did you have to reboot it?17:06
dmsimardno, just a resume17:06
clarkbah17:06
dmsimardhence the 116 day uptime17:06
fungiinfra-root: unrelated to the cpu pipelining exploits disclosed today, be careful with recursive chown/chgrp and avoid following symlinks: http://lists.gnu.org/archive/html/coreutils/2017-12/msg00045.html17:06
dmsimardlooks like it might have been suspended around Dec 29th/Dec 30th17:06
dmsimardon the topic of cpu exploits, are we planning to upgrade kernels and reboot ?17:07
clarkbdmsimard: yes but no kernels available yet as far sa I can tell17:07
clarkbubuntu's story is the embargo was supposed to lift on the 9th17:07
clarkbbut it lifted a week early and now they are scrambling as a result17:08
fungidmsimard: we probably should, though i recommend coordinating it in #-incident once packages are up17:08
dmsimardfungi: +117:08
dmsimardJust learned that the packages are up for centos so I'll be busy for a bit updating RDO's stuff17:10
*** caphrim007 has joined #openstack-infra17:10
dmsimardI guess we can do the git farm on -infra17:10
*** slaweq has joined #openstack-infra17:10
fungiyeah, rolling restarts on those should be safe if we disable them in haproxy one at a time17:11
fungirestarting git.o.o itself will be disruptive though17:11
clarkbpabelanger wanted to do that anyways for general patching17:11
clarkb(but they are also not servers that I'd worry about given the vulnerability)17:12
clarkbI'm more concerned about my laptop and infracloud17:12
fungiyeah, our biggest exposure is anywhere we run arbitrary payloads. job nodes and zuul executors are my largest worry17:13
clarkbno kernels for either yet :/17:13
fungiand our personal workstations/laptops/devices of course, right17:13
clarkbhttps://wiki.ubuntu.com/SecurityTeam/KnowledgeBase/SpectreAndMeltdown fwiw17:14
*** dhill_ has quit IRC17:14
fungiin other news, hurricane-force winds coupled with a thunderstorm and blizzard is... an interesting combo17:14
corvusfungi: i woke up to a 4.4 today.17:14
funginice!17:15
corvusi'm walking distance from the epicenter17:15
fungii basically didn't sleep last night because the wind was deafening17:15
*** slaweq has quit IRC17:15
dmsimardThey're announcing around 16 inches of snow here17:15
corvuseverything is awesome!17:16
funginow if we could just combine them all into a thundersnowcanequake17:16
pabelangercorvus: yah, I can look again today how to rotate out a git server from LB17:17
*** gyee has joined #openstack-infra17:17
dmsimardso I didn't read the meltdown/spectre docs in depth but.. regardless of if we patch our stuff or not -- given we're on virtual machines in different clouds, that means we're still vulnerable if the cloud provider hasn't updated, correct ?17:17
corvusdmsimard: yes, though i'm assuming that's what was happening over december.... has rax mentioned anything publicly yet i wonder?17:17
dmsimardhmm, there was something specific about xen actually, I think only PV was impacted ? /me looks17:18
*** spzala has joined #openstack-infra17:18
*** spzala has quit IRC17:18
danpawlikcoreycb: Hi. Do you know what is the state of cannonical team to create new Qemu package for CVE-2017-5715 ?17:18
fungidmsimard: at least in rackspace, i read the xen advisories and they make it sound like the hypervisors are mostly safe from guest breakout with pvhvm (it's mostly straight pv 64-bit guests you have to worry about being able to address the hypervisor's pagetable)17:18
corvusdmsimard, fungi: nope, apparently that's not what was happening in december17:18
clarkbdmsimard: yes my understanding is that particularly for meltdown (easier to exploit) hypervisors can have their memory read which would expose other guests17:18
corvushttps://techcrunch.com/2018/01/03/cloud-infrastructure-vendors-begin-responding-to-chip-kernel-vulnerability/17:18
*** smatzek has joined #openstack-infra17:18
corvusrax is at the end there17:18
clarkbdmsimard: but short of shutting down all our servers I don't think we can do much there17:19
corvusapparently they weren't cool enough to be part of the nda17:19
fungiyeah17:19
dmsimardyeah.... this sucks17:19
clarkbspectre is a more widespread problem but is much harder to take advantage of17:20
danpawlikor maybe someone else have some contacts with canonical17:20
dmsimardI expected Intel's stock price to tank way more than that17:20
fungixen advisories mention a hypervisor shim you can insert to safely protect the host from pv guests17:20
fungiso i wouldn't be surprised if rax mitigates that way17:20
*** eharney has quit IRC17:20
coreycbdanpawlik: i believe this has the latest status: https://wiki.ubuntu.com/SecurityTeam/KnowledgeBase/SpectreAndMeltdown17:21
*** smatzek has quit IRC17:21
*** smatzek has joined #openstack-infra17:21
corvushere's an idea: shut everything down and extend winter holidays another week.17:21
dmsimard+117:21
pabelangerI could get behind that17:21
fungicorvus: i'll get started by shutting down my systems now ;)17:22
dmsimardalternatively justify 25% lack of productivity due to "performance impact"17:22
*** eharney has joined #openstack-infra17:22
fungii should go out and play in the snow, except for these 50mph wind gusts (at least we're below hurricane velocities again now)17:22
corvusdmsimard: yeah, you get to use that for the next 10 years too17:22
danpawlikcoreycb: ok, thanks.17:23
danpawlikcoreycb: its really sad, that suse and red hat has a new package but all patches are confidential17:23
fungithe kernel patches are public17:24
clarkbdanpawlik: suse has a new package? my tumbleweed has yet to get 4.14.11 (maybe 4.14.9 got a backported patch?)17:24
coreycbdanpawlik: it may have something to do with the embargo, i'm not on the security team though17:24
fungionly the discussion around the kernel patches was embargoed17:24
danpawlikcoreycb: and maybe you will know, why debian is telling that package related is just linux (https://security-tracker.debian.org/tracker/CVE-2017-5715), on ubuntu you have (https://people.canonical.com/~ubuntu-security/cve/2017/CVE-2017-5715.html) just kernel but red hat and suse has other opinion: https://www.suse.com/security/cve/CVE-2017-5715/ and17:24
danpawlikhttps://access.redhat.com/security/cve/cve-2017-571517:24
danpawlikhmm17:25
danpawlikfungi: but its partialy related to the kernel17:25
danpawlikfungi: so its possible that emgargo is also on that bug too17:25
*** dhill_ has joined #openstack-infra17:26
danpawlikas I read on the Internet that patch is for: "qemu update to pass the new MSR and CPU flags from the microcode update"17:26
fungiyeah, there are of course other mitigating patches being implemented in, e.g., virtualization systems for similar reasons17:26
openstackgerritStephen Finucane proposed openstack-dev/pbr master: doc: Rework features doc  https://review.openstack.org/51224817:27
openstackgerritStephen Finucane proposed openstack-dev/pbr master: doc: Minor rework of usage doc  https://review.openstack.org/51224917:27
fungiit's hard to say that these bugs "affect" any particular piece of software. they're hardware design flaws which need to be mitigated or worked around by software. kernels are in the best position to do most of that17:27
persiaExcitingly, there is a decent chance that hypervisors can be made immune to this before hardware, making a strong argument for running sensitive materials on clouds (assuming the substrate maintainers make efforts to control code run on hardware)17:27
fungipersia: totally agree17:28
stephenfindhellmann, mordred: Fancy shoving these pbr docs patches through? https://review.openstack.org/#/q/topic:doc-rework+(status:open+OR+status:merged)17:29
dmsimardfungi: yeah for Red Hat, there's updates to kernel microcode_ctl qemu-kvm qemu-kvm-rhev libvirt dracut linux_firmware17:30
dmsimard(that's from an openstack perspective on a RHEL deployment)17:30
*** stephenfin has left #openstack-infra17:30
*** stephenfin has joined #openstack-infra17:30
EmilienMclarkb, pabelanger : I would appreciate some review on https://review.openstack.org/#/c/531030/ - thanks17:32
EmilienMclarkb, pabelanger : I'll do centos as well17:32
fungipersia: well, not necessarily clouds, but virtual servers at the very least17:32
persiafungi: Well, yes, but that isn't quite as exciting, except for the Qubes folk.17:34
fungiheh, indeed ;)17:34
dmsimardoh, a demo exploiting meltdown for capturing realtime data.. https://twitter.com/misc0110/status/94870638749178675217:36
*** erlon has quit IRC17:36
clarkbdmsimard: ya meltdown is much more trivial to exploit17:38
fungicorvus: i particularly enjoyed "we have not yet heard from ibm"17:38
clarkbdmsimard: but is also the attack with patches (kpti)17:38
clarkbnow just need new kernels17:38
clarkbdo we have a revert of the ovh disable in nodepool yet?17:38
*** rlf1 has joined #openstack-infra17:39
*** greghaynes has joined #openstack-infra17:42
coreycbdanpawlik: i'm not sure. #ubuntu-hardened may be a good place for that question.17:43
clarkbmordred: there are quite a few items with your name attached to them on https://etherpad.openstack.org/p/zuulv3-issues anything I can help with as far as moving things along? But also would be good to clear out things that have been fixed17:43
clarkbmordred: the tox with sudo thing came up again too (and is on the etherpad) is that sorted yet?17:43
dmsimardclarkb: I saw a revert of the clouds.yaml, not for nodepool17:44
danpawlikcoreycb: thanks. I can't find such irc group to ask  so I was thinking that I will ask you  / community and someone will know something more17:44
danpawlikjoin #ubuntu-hardened17:44
clarkbdmsimard: ok I'llget a nodepool revert up now17:44
fungiclarkb: curious if you happen to have seen mriedem's post to -dev about issues serving logs from node-named directories, and what precipitated the regressions there17:45
*** lucasagomes is now known as lucas-afk17:46
openstackgerritClark Boylan proposed openstack-infra/project-config master: Revert "Disable OVH"  https://review.openstack.org/53120217:46
clarkbdmsimard: pabelanger fungi mordred ^ use ovh again in nodepool17:46
*** slaweq has joined #openstack-infra17:47
clarkbfungi: os-loganalyze is very specific about what files it will bother annotating since logs come in all the formats. I'm betting the new job changing file paths (and possibly file names?) no longer match os loganalyze's expectations17:47
clarkbfungi: I can take a look though17:48
fungiclarkb: thanks, i'll dig into osla internals then. i looked at the apache vhost config and it didn't appear to treat certain subpaths specially17:49
clarkbandreaf: looking into ^ the first thing I notice is that none of the log files are compressed. That should be addressed a relatively urgent thing to fix17:49
*** armax has quit IRC17:50
*** armax has joined #openstack-infra17:50
mordredclarkb, frickler: I feel like someone was discussing the need for tox without sudo revoked earlier17:50
clarkbfungi: ^ may actually be the problem if the filename osla is looking for include the .gz17:50
clarkbmordred: ya they were in sb17:50
clarkbmordred: for a functional job for os privsep17:50
*** sshnaidm|mtg is now known as sshnaidm17:51
clarkbfungi: hrm no ^.*\.txt(\.gz)?$ is the regex in the wsgi conf and that should match either version17:54
mordredhttps://review.openstack.org/#/c/512653/ <-- there's a patch for a that (from scanning the etherpad)17:55
clarkbbut we enforce the .gz in apache. RewriteRule ^/(.*\.txt\.gz)$ /htmlify/$1 [QSA,L,PT,NS]17:55
clarkbfungi: mriedem andreaf ^ pretty sure that is the issue. If the job starts compressing the log files again it should work17:55
jlvillalAny known issues with unit tests jobs? Or is it just us (Ironic)?17:55
jlvillalNever mind. Seems to be just us.17:56
clarkbjlvillal: zuuls status page builds list should help you filter and see if others are havingtrouble17:56
clarkbbut I'm not aware of anything17:56
jlvillalclarkb: Thanks!17:57
clarkbfungi: I'm writing a response to the dev list now unless you have already started one17:58
corvusi think we want the job performing the compression.  i'd like for us to get to the point where we can rely on that because it's in a post playbook, and stop having the cron17:59
*** Apoorva has joined #openstack-infra17:59
clarkbcorvus: yes absolutely, the cron already has a hard time keeping up so pushing that into the scale out job workers is a big win17:59
corvushopefully that's something we can put in the new genericized log archiving role18:00
*** david-lyle has quit IRC18:00
openstackgerritMerged openstack-infra/system-config master: Revert "Remove ovh from clouds.yaml files"  https://review.openstack.org/53094518:01
*** david-lyle has joined #openstack-infra18:01
fungiclarkb: i had not started one, go for it18:02
fungii had only just started to dig into the problem18:02
*** derekh has quit IRC18:03
*** Goneri has joined #openstack-infra18:05
*** slaweq has quit IRC18:10
AJaegerclarkb: I run 4.14.11 from http://download.opensuse.org/repositories/Kernel:/stable/standard - always latest stable kernel18:12
*** jascott1 has joined #openstack-infra18:13
AJaegermordred: it was stephenfin18:14
*** SumitNaiksatam has joined #openstack-infra18:14
dmsimardmordred: does http://paste.openstack.org/raw/638205/ ring you a bell ? image is http://paste.openstack.org/show/638215/18:17
AJaegermordred: I think we can merge the tox-with-sudo change - after removing the duplicated job, see my comment on https://review.openstack.org/#/c/51265318:18
mordreddmsimard: no - I'm not sure I've seen a conflict on delete before18:18
dmsimardmordred: haven't tried deleting it manually yet, going to try and reproduce with shade proper18:18
mordredAJaeger: actually - I feel like corvus and I had a brief conversation on this topic right before the break ...18:18
openstackgerritXiang Chen proposed openstack-infra/git-review master: fix bug  #!/story/2001438  https://review.openstack.org/53120918:19
corvusmordred: that was so last year... what are we talking about? :)18:19
mordreddmsimard: if you do, can you put shade.simple_logging(http_debug=True) at the top of your script and save the output? the json decode error bothers me, I think there is an error message we're missing18:19
mordredcorvus: revoke-sudo and the unittest base job18:20
corvusmordred: oh yes, i believe i said something which caused you to discover a very clever solution to the problem.18:20
AJaegermordred: perhaps in context of https://review.openstack.org/529662 ?18:20
corvusmordred: you may have written it into the margins of a proof or something.18:20
mordredcorvus: yes - that is what I remember too18:20
AJaegermordred: check 529662, that moves the sudo call...18:21
mordredcorvus: but I don't remember the solution or what proof I wroteit in the margins of18:21
clarkbcorvus: mordred andreaf mriedem fyi just pushed https://review.openstack.org/531208 to confirm the log annotation problems are related to gzipping18:21
clarkbI expect that the general gzipping of log files will be more involved so didn't start with that18:21
mordredAJaeger: yes!18:21
fungicorvus: i suspect the margins were too small to contain it18:21
clarkbfungi: oh ^ you too18:21
mordredAJaeger, corvus: https://review.openstack.org/529662 is, in fact, what we came up with18:21
AJaegermordred: so, plan is to merge 529662  and then rework the tox-with-sudo to just have a new run.yaml?18:22
mriedemclarkb: huh, how did that regresss? didn't devstack's post-test host cleanup code do that before?18:22
corvusmordred: you wrote it in the margins of a git commit! clever :)18:22
AJaegerLet me +3 529662. mordred will you rework the tox-with-sudo or shall I?18:22
mordredAJaeger: yup. lemme do that real quick18:22
clarkbmriedem: tempest-full no longer uses devstack-gate so its entirely new cleanup code that wasn't ported properly18:22
*** slaweq has joined #openstack-infra18:23
mriedemok18:23
mriedemfigured it was something like that18:23
clarkbmriedem: basically we ported devstack-gate into zuulv3 native job config and this was something missed18:23
fungidevstack-gate essentially being a legacy framework obsoleted by the available options in zuul v318:23
openstackgerritXiang Chen proposed openstack-infra/git-review master: fix bug  #!/story/2001438  https://review.openstack.org/53120918:23
openstackgerritXiang Chen proposed openstack-infra/git-review master: fix bug  #!/story/2001438  https://review.openstack.org/53120918:24
dmsimardmordred: reproduced with debugging enabled: http://paste.openstack.org/show/638223/18:24
mordreddmsimard: thanks!18:24
openstackgerritXiang Chen proposed openstack-infra/git-review master: fix bug  #!/story/2001438  https://review.openstack.org/53120918:24
dmsimardmordred: looks like the http 409 is text/plain, not JSON18:25
dmsimard"RESP BODY: Omitted, Content-Type is set to text/plain; charset=UTF-8. Only application/json responses have their bodies logged."18:25
fungithinking about it, i guess we've been deprecating pieces of devstack-gate and moving them into our services for years... who remembers when nodepool was actually some cron'd scripts in devstack-gate?18:25
*** dhill_ has quit IRC18:25
*** shardy has quit IRC18:25
AJaegerteam, three quick reviews for job removals in openstack-zuul-jobs: https://review.openstack.org/#/c/528645/ , https://review.openstack.org/#/c/530240/  , https://review.openstack.org/#/c/513579/ , please18:25
*** dhill_ has joined #openstack-infra18:26
corvusfungi: i remember when nodepool was a jenkins job :)18:26
corvus(there was a job to create nodes, and another job to delete them.  this design is not advised.)18:27
mordreddmsimard: ok. cool. I thnk there is a patch lurking somewhere to fix that18:27
dmsimardmordred: running openstack image delete yields a slightly more useful error message: Image 4bbd529d-83d3-4d6f-a69e-df7e1e69aa6b could not be deleted because it is in use: The image cannot be deleted because it is in use through the backend store outside of Glance. (HTTP 409)18:27
*** slaweq has quit IRC18:27
fungicorvus: indeed, though the scripts those jobs ran were in the devstack-gate repo was i guess my point18:28
mordreddmsimard: ++ that is the error message we should be producing in shade too18:28
corvusfungi: ah yep :)18:28
mordredthat will be fixed by the thing I'm thiking about18:28
dmsimardmordred: I'll file a bug just so I don't forget and can refer to it if need be18:28
mordreddmsimard: so - as for the other thing, *blink-blink*18:28
mordreddmsimard: thanks18:29
mordreddmsimard: I think that may be an issue to raise with the tripleo cloud team18:29
*** yamamoto has quit IRC18:29
mordred(the in-use-by-something-else error)18:29
dmsimardmordred: oh, yeah sure I'll take care of that18:29
dmsimardthey upgraded to ocata recently and it was a bumpy ride18:30
dmsimard2 weeks of headaches :(18:30
clarkbmordred: corvus is stage-output the magical zuul role that should do the gzipping?18:30
fungidmsimard: but... hooray for being on latest openstack i guess?18:32
AJaegermordred: shouldn't we move the revoke-sudo in project-config as well?18:32
fungidmsimard: d'oh, wait, that's pike ;)18:32
* dmsimard frowns18:33
*** yamamoto has joined #openstack-infra18:33
dmsimardocata actually EOLs like, next month18:33
*** rlf1 has quit IRC18:33
clarkbhrm stage-output claims to already compressing things but that doesn't appear to be happening18:34
*** erlon has joined #openstack-infra18:34
*** numans_afk is now known as numans18:36
clarkboh! there is a comment in the file I am updating that says it bypasses stage-output18:37
AJaegermordred: thinking further - no need to change anything for project-config18:37
clarkbso maybe my patch is the actual fix afterall18:37
clarkbcorvus: ^ fyi18:37
*** yamamoto has quit IRC18:37
corvusclarkb: mordred and andreaf have been the most involved in the log copying rework.  i know the general shape of the end-result, but i don't have the current context.18:41
clarkbcorvus: thanks, mordred andreaf can you review https://review.openstack.org/#/c/531208/ to fix mriedem's problems with os-loganalyze? I now think that that fix may actually be the fix until we stop bypassing stage-output with service logs18:41
openstackgerritMerged openstack-infra/zuul-jobs master: Move revoke-sudo from unittest pre to job run  https://review.openstack.org/52966218:42
openstackgerritJeremy Stanley proposed openstack-infra/system-config master: Limit Gerrit SSH API connections to 100 per addr  https://review.openstack.org/52971218:43
dmsimardmordred: fyi image bug if you want to refer to it: https://storyboard.openstack.org/#!/story/200144018:45
openstackgerritMerged openstack-infra/project-config master: Revert "Disable OVH"  https://review.openstack.org/53120218:47
AJaegeryeah, another 238 nodes ;) ^18:47
AJaegercorvus: do we still need the merge-check pipeline or can we remove it from project-config/zuul.d/pipelines.yaml ?18:50
dmsimardAJaeger: if it works18:53
dmsimardshould keep an eye on nodepool grafana just in case18:53
clarkbAJaeger: I think we can remove it if we are willing to rely on gerrits merge checking18:54
AJaegerclarkb: we're not using it currently, let me propose the change...18:55
*** ijw has joined #openstack-infra18:55
openstackgerritAndreas Jaeger proposed openstack-infra/project-config master: Remove unused merge-check pipeline  https://review.openstack.org/53121518:57
AJaegerclarkb, corvus ^18:57
clarkbAJaeger: zuul caught an error with it that you'll need to address first18:57
AJaegerclarkb: yeah - seems that we do use it.18:58
* AJaeger loves zuul's self-checks18:58
clarkbbut in a trivial way I think we can drop that use of it18:58
clarkbunless that makes the system-required template invalid18:58
*** caphrim007_ has joined #openstack-infra18:59
*** Goneri has quit IRC19:00
*** slaweq has joined #openstack-infra19:01
*** pcrews has joined #openstack-infra19:01
*** caphrim007 has quit IRC19:03
AJaegerclarkb: I fear it will break it. Let's leave it for now...19:04
*** slaweq has quit IRC19:05
*** pots has joined #openstack-infra19:08
mnaserdmsimard: mordred is the in-use glance error by any chance running against an openstack cloud with rbd storage?19:10
dmsimardmnaser: yeah, I know what the 409 is, there is a volume using that image19:12
mnaserok cool (or a nova instance too)19:12
* mnaser goes back to reboot party19:12
clarkbdmsimard: so you have external tools using nodepool images causing nodepool's image deletions to fail?19:13
clarkbspeaking of reboot parties19:13
dmsimardNope, nodepool boot from volume19:13
clarkbpabelanger: ^ do you want to work through the git backends today?19:13
*** slaweq has joined #openstack-infra19:14
eumel8clarkb, ianw_pto: It seems we're ready to update translate-dev to Zanata 4.3.3. Awaiting a valid download URL from the Zanata guys and then I will send out a new proposal the next days.19:16
clarkbeumel8: sounds good, this is to address the bugs that were found right?19:17
eumel8clarkb: yes, the wrong landing page after openid login19:18
openstackgerritMerged openstack-infra/openstack-zuul-jobs master: Remove obsolete interop jobs  https://review.openstack.org/52864519:19
*** slaweq has quit IRC19:19
clarkbthere are 144 ovh instances in use right now19:20
clarkbelastic-recheck not showing a spike in the pip failures so far19:20
eumel8clarkb: regarding the wrong accounts in the dev database we decided to drop the complete database and started with a new one. I think it's easier for you instead to puzzle the wrong accounts with the good one. This caused last year after changing the openid provider.19:21
clarkbeumel8: you'll lose the translation history and stuff but thats probbaly ok for the dev instance I guess19:22
openstackgerritAndreas Jaeger proposed openstack-infra/project-config master: WIP: Use native Zuul v3 bindep fallback tests  https://review.openstack.org/53010419:23
eumel8clarkb: sure, but nobody can work on the dev server which registered before 2016. It's easier to install some test repos again19:23
openstackgerritMerged openstack-infra/openstack-zuul-jobs master: Remove legacy-grenade-publish-docs  https://review.openstack.org/53024019:24
*** jascott1 has quit IRC19:26
*** jascott1 has joined #openstack-infra19:26
openstackgerritAndreas Jaeger proposed openstack-infra/project-config master: WIP: Use native Zuul v3 bindep fallback tests  https://review.openstack.org/53010419:27
*** jascott1 has quit IRC19:28
*** jascott1 has joined #openstack-infra19:29
*** jascott1 has quit IRC19:30
openstackgerritMerged openstack-infra/openstack-zuul-jobs master: Remove legacy jobs for python-heatclient  https://review.openstack.org/51357919:30
*** jascott1 has joined #openstack-infra19:33
EmilienMclarkb: thanks for your review on https://review.openstack.org/#/c/531030/ - I'll start the mirror for puppetlabs on centos19:34
*** slaweq has joined #openstack-infra19:35
*** yamamoto has joined #openstack-infra19:35
openstackgerritAndreas Jaeger proposed openstack-infra/project-config master: Add native Zuul v3 bindep fallback tests  https://review.openstack.org/53010419:38
*** smatzek has quit IRC19:39
*** slaweq has quit IRC19:39
openstackgerritAndreas Jaeger proposed openstack-infra/project-config master: Add native Zuul v3 bindep fallback tests  https://review.openstack.org/53010419:40
*** armax has quit IRC19:40
*** yamamoto has quit IRC19:42
openstackgerritAndreas Jaeger proposed openstack-infra/project-config master: Use new bindep tests  https://review.openstack.org/53122219:44
dmsimardTIL certbot uses pbr https://github.com/certbot/certbot/blob/d039106b687f7300a41e11088412f16dd6f023d5/certbot-auto#L99719:46
openstackgerritAndreas Jaeger proposed openstack-infra/project-config master: DNM: Test new bindep tests  https://review.openstack.org/53122519:47
AJaegerconfig-core, could you review  https://review.openstack.org/530104 , please? Like to know whether the native bindep test is fine. Once that is merged, we can test it with 531225 and then use by default with 53122219:48
*** sshnaidm is now known as sshnaidm|off19:53
*** eharney has quit IRC20:01
*** jascott1 has quit IRC20:02
*** jascott1 has joined #openstack-infra20:02
*** jascott1 has quit IRC20:03
openstackgerritJeremy Stanley proposed openstack-infra/system-config master: Limit Gerrit SSH API connections to 100 per addr  https://review.openstack.org/52971220:03
*** jascott1 has joined #openstack-infra20:04
clarkbfungi: can you multiline puppet like that?20:04
clarkbI guess we'll know soon enough20:04
*** spzala has joined #openstack-infra20:04
fungiyeah, i thought we'd done it elsewhere20:05
fungiat any rate, the missing $ characters were a no-go for it20:06
fungibizarre that assigning collections like that requires you to have $ on the variable being assigned20:06
fungii feel like i still don't grok some of these syntax nuances in puppet20:07
openstackgerritEmilien Macchi proposed openstack-infra/system-config master: Add Puppetlabs mirror for CentOS7  https://review.openstack.org/53122920:07
*** Goneri has joined #openstack-infra20:07
*** jascott1 has quit IRC20:08
*** jascott1 has joined #openstack-infra20:10
*** markmcclain has quit IRC20:10
*** markmcclain has joined #openstack-infra20:11
*** jascott1 has quit IRC20:13
*** slaweq has joined #openstack-infra20:13
*** nicolasbock has quit IRC20:13
*** SumitNaiksatam has quit IRC20:13
corvusit's easy, you always need a $ except when it's not required20:14
*** slaweq_ has joined #openstack-infra20:15
clarkbmy ubderstanding of it is you need the $ anywhere it is a variabke and drop the $ if it is a parameter name20:16
*** slaweq has quit IRC20:18
fungiahh. that sort of makes sense20:19
mordredtry one. wait for puppet-lint to complain, then try the other20:19
mordredthat's my approach20:19
fungiseems to have been my approach today as well20:19
dmsimardis there a way in nodepool v2 to re-upload an image to a provider ?20:25
clarkbdmsimard: yes run the image-upload command20:25
dmsimarddoesn't seem to exist :/20:26
clarkbdid you mean v3?20:26
clarkbv3 doesnt have it but the pre v3 stuff should20:27
clarkbv3 instead attempts to recocnile state more aggressively and upload things when necessary20:27
clarkbso you can force a reupload by deleting older images20:27
dmsimardhttp://paste.openstack.org/raw/638343/ is the help I get from nodepool -h20:28
dmsimardit's weird cause I remember image-upload existing ..20:29
openstackgerritJames E. Blair proposed openstack-infra/infra-specs master: Add IRC bot improvement spec  https://review.openstack.org/53124220:29
corvusclarkb, fungi, dmsimard, mnaser, mordred: ^ that's the brainstorming etherpad from yesterday dumped into spec form20:30
dmsimardsubscribed ty20:30
* mnaser adds $self and goes back to reboot party20:30
clarkbdmsimard: if there is a zookeeper instead of a gearman I believe that is expected20:30
clarkbdmsimard: but if your builders speak gearman instead I would expectn upload command20:31
dmsimardit's kinda weird because we're in this weird mix of things where we're actually still running jenkins and we'll jump straight to zuul v320:33
dmsimardI'll figure something out20:33
clarkbthe builder + zk work was independent of that though20:33
*** jascott1 has joined #openstack-infra20:34
dmsimardafaict our nodepool builder uses gearman20:34
dmsimardeh, maybe not.. seeing zookeeper stuff in the logs20:35
openstackgerritMerged openstack-infra/zuul feature/zuulv3: Don't remove builds that don't exist  https://review.openstack.org/53119820:36
dmsimardI'll just run a new build and see what happens20:36
*** jascott1 has quit IRC20:38
*** gagehugo has quit IRC20:39
*** jascott1 has joined #openstack-infra20:41
*** jascott1 has quit IRC20:43
*** jascott1 has joined #openstack-infra20:47
*** jascott1 has quit IRC20:50
*** slaweq has joined #openstack-infra20:50
*** smatzek has joined #openstack-infra20:50
openstackgerritMerged openstack-infra/zuul-jobs master: Fix pep8/yamllint errors  https://review.openstack.org/53068520:53
*** edmondsw has quit IRC20:54
*** jascott1 has joined #openstack-infra20:54
*** eharney has joined #openstack-infra20:56
*** jascott1 has quit IRC20:56
*** jascott1 has joined #openstack-infra20:58
*** smatzek has quit IRC21:05
*** smatzek has joined #openstack-infra21:06
*** smatzek has quit IRC21:10
*** caphrim007_ has quit IRC21:10
*** caphrim007 has joined #openstack-infra21:11
clarkbcan an infra root be second review on https://review.openstack.org/#/c/516009/ to fix infra publications publishing? Thta will clear out an item from the zuulv3 issues etherpad21:17
*** eumel8 has quit IRC21:19
*** edmondsw has joined #openstack-infra21:19
mordredclarkb: done21:19
dmsimardThere will be a brief moment of unavailability of the RDO trunk repositories soon, hopefully the reverse proxy cache will hold us through the reboot for the kernel update.21:19
mordreddmsimard: I believe in the reverse proxy cache21:20
openstackgerritMerged openstack-infra/publications master: Update make-index for Zuul v3  https://review.openstack.org/51600921:20
dmsimardmordred: thank you for believing :D21:20
clarkbfungi: http://logs.openstack.org/12/529712/3/check/legacy-infra-puppet-apply-3-ubuntu-trusty/7ad69cd/openstack-infra/system-config/applytest/puppetapplytest01 apply tested your multiline variable assignment and didn't explode so +2 from me21:22
*** slaweq has quit IRC21:23
clarkbhttps://review.openstack.org/#/c/523937/1 is a simple docs fix too related to storyboard and project creation21:24
*** smcginnis has joined #openstack-infra21:27
*** jascott1 has quit IRC21:29
*** jascott1 has joined #openstack-infra21:29
*** slaweq_ has quit IRC21:29
*** jascott1 has quit IRC21:30
*** olaph has quit IRC21:30
*** jascott1 has joined #openstack-infra21:30
*** olaph has joined #openstack-infra21:31
*** slaweq has joined #openstack-infra21:32
mriedemclarkb: so on https://review.openstack.org/#/c/531208/ the logs look good,21:33
mriedemhttp://logs.openstack.org/08/531208/1/check/tempest-full/80532f7/controller/logs/screen-n-api.txt.gz?level=WARNING#_Jan_04_19_18_07_55286721:33
mriedemfiltering, colors, all works21:33
mriedemBUT!21:33
mriedemhttp://logs.openstack.org/08/531208/1/check/tempest-full/80532f7/controller/logs/etc/nova/ still requires a download - but maybe that's a different fix21:34
clarkbmriedem: ya I think thats a separate but related issue21:34
*** jascott1 has quit IRC21:34
clarkbmriedem: separate because this fix won't fix it but related to the porting of devstack-gate and changes in how things are compressed21:34
dmsimardthe vhost mime types handles .txt.gz but not .ini.gz and etc21:35
*** jascott1 has joined #openstack-infra21:35
dmsimardin puppet-openstackci21:35
dmsimardwe'd need to add .conf.gz, .ini.gz and whatever else21:35
mriedemclarkb: ack; +1 on getting the logs fixed though, that was much more annoying than the config files21:35
*** threestrands has joined #openstack-infra21:35
*** threestrands has quit IRC21:35
*** threestrands has joined #openstack-infra21:35
mriedemmtreinish: sdague: https://review.openstack.org/#/c/531208/21:36
dmsimardmriedem: http://git.openstack.org/cgit/openstack-infra/puppet-openstackci/tree/templates/logs.vhost.erb#n2421:36
*** slaweq has quit IRC21:37
*** jascott1 has quit IRC21:37
mriedemdmsimard: so conf and ini files need to be in that list too?21:37
* mriedem doesn't puppet21:37
dmsimardwhatever extension that ought to be text/plain (when decompressed) should be there, yeah21:37
dmsimardthere's little to no puppet in that file, it's mostly an apache vhost :)21:38
mriedemheh, i don't do that either21:38
mriedemthere isn't much that i do, do21:38
*** jascott1 has joined #openstack-infra21:38
dmsimardyou do nova21:38
dmsimardthat's good enough :D21:38
clarkbah ya so I think the change here from devstack-gate to logs is we used to rename the inis to .txt21:38
clarkbwhich is also another option we could do21:39
*** dhill_ has quit IRC21:39
clarkbbut might be more accurate and easier to grok if we .ini and handle that as txt21:39
dmsimardclarkb: either we rename or we add the mimetypes21:39
mriedemlooks like it would be at least ini, conf and filters21:39
mtreinishmriedem: +A21:39
dmsimardrenaming everything has a cost21:39
corvusi think renaming is weird and we should stop21:39
mriedemcan we just hit everything that's under etc/?21:39
*** dhill_ has joined #openstack-infra21:39
dmsimardcorvus: +121:39
*** jcoufal has quit IRC21:40
dmsimardTrunk RDO repositories have been rebooted after kernel update btw21:41
*** dhill_ has quit IRC21:41
*** dhill_ has joined #openstack-infra21:42
*** jascott1 has quit IRC21:42
openstackgerritClark Boylan proposed openstack-infra/openstack-zuul-jobs master: Create artifacts dir regardless of afs rpm builds  https://review.openstack.org/53124921:42
clarkbok ^ is a fix to my fix for the openafs rpm builds21:42
*** sbezverk has joined #openstack-infra21:43
clarkbreviews much appreciated21:43
*** dmellado has quit IRC21:43
clarkbanyone know if the releasenote build issues with tox siblings is all working happily now? (I kind of assume it does simply because relese notes not working would be a big issue people would be complaining about)21:44
dmsimardbtw insight on spectre/meltdown from a qemu perspective https://www.qemu.org/2018/01/04/spectre/21:45
*** bobh has quit IRC21:45
*** wolverineav has joined #openstack-infra21:46
*** dmellado has joined #openstack-infra21:46
*** dhill_ has quit IRC21:47
*** dhill_ has joined #openstack-infra21:48
EmilienMclarkb: https://review.openstack.org/#/c/531229/ is ready for review when you have time21:48
*** dhill_ has quit IRC21:49
*** slaweq_ has joined #openstack-infra21:50
openstackgerritJames E. Blair proposed openstack-infra/zuul feature/zuulv3: WIP: Support cross-source dependencies  https://review.openstack.org/53080621:52
*** dhill_ has joined #openstack-infra21:53
*** dhill_ has quit IRC21:54
*** esberglu has quit IRC21:54
*** slaweq_ has quit IRC21:54
dmsimardbetween two kernel updates, reading on the news coming out today.. this from the NY Times: "The worldwide community of coders that oversees the open-source Linux operating system, which runs about 30 percent of computer servers worldwide, has already posted a patch for that operating system." .. only 30% ?21:55
*** smatzek has joined #openstack-infra21:57
*** dhill_ has joined #openstack-infra21:58
fungidmsimard: you forget that most servers are now serverless21:59
dmsimardoh, is that "the cloud" ? :D22:01
fungisoon it will even be softwareless22:01
clarkbit will just be an aggregate of thought like in the matrix22:03
fungisounds about right22:04
fungiwake up, neo22:04
dmsimardquick, get keanu reeves22:04
fungistrange things are afoot in the cloud22:05
*** slaweq has joined #openstack-infra22:05
*** esberglu has joined #openstack-infra22:06
*** slaweq has quit IRC22:10
*** markmcclain has quit IRC22:10
clarkbpabelanger: are you around today? Thinking it owuld be a good idea to get the centos updates working and reboots for meltdown in place as soon as possible and know you were interested in how to make haproxy do the right thing around that22:11
*** markmcclain has joined #openstack-infra22:11
dmsimardI'm almost finished patching all of RDO's stuff, I'd take a look but today was pretty rough22:12
*** rcernin has joined #openstack-infra22:12
clarkbI still don't see anything from ubuntu so I think we are in a holding pattern there22:13
dmsimardyeah still no update on the wiki22:14
clarkbdmsimard: its fairly straightforward to do the git backends but there is a process and I know pabelanger wanted to go through it before we all holidayed so wanted to give him the chance22:14
clarkbI'm going to go ahead and disable git08 in haproxy and use it as the first node to be updated22:15
*** smatzek has quit IRC22:15
clarkbdmsimard: you don't happen to know off the top of your head what version of th ekernel we want do you? (centos 7)22:18
dmsimardyeah one sec22:18
dmsimardkernel-3.10.0-693.11.6.el7.x86_64: Linux trunk-primary.rdoproject.org 3.10.0-693.11.6.el7.x86_64 #1 SMP Thu Jan 4 01:06:37 UTC 2018 x86_64 x86_64 x86_64 GNU/Linux22:19
*** markmcclain has quit IRC22:19
clarkbtyty22:19
dmsimardthere is another previous dot release within -69322:19
dmsimardlike 693.2.2 for example, just need to make sure it's the .1122:19
clarkbgotcha22:19
dmsimardgotta love rhel kernel versions22:20
*** markmcclain has joined #openstack-infra22:20
clarkbjust waiting for git daemon processes to drop off then will yum update and reboot22:21
*** aeng has joined #openstack-infra22:21
*** smatzek has joined #openstack-infra22:21
*** Goneri has quit IRC22:21
*** smatzek has quit IRC22:22
*** smatzek has joined #openstack-infra22:23
dmsimardwhere is the haproxy for those ?22:23
dmsimarddoesn't seem documented in https://docs.openstack.org/infra/system-config/git.html22:24
clarkbif anyone is wondering haproxy command socket commands are `echo "disable server balance_git_http/git08.openstack.org" | socat /var/lib/haproxy/stats stdio` `echo "disable server balance_git_https/git08.openstack.org" | socat /var/lib/haproxy/stats stdio` and `echo "disable server balance_git_daemon/git08.openstack.org" | socat /var/lib/haproxy/stats stdio`22:24
clarkbdmsimard: git.openstack.org22:24
fungion git.o.o22:24
dmsimardah, that was obvious actually22:24
clarkbthen when done with git08 rplace disable with enable make sure its happy then go to the next one22:24
clarkbthat won't kill existing connections which is why I have to wait forthings to settle on git08 first (but no new connections)22:24
dmsimardansible has a haproxy module but never tried it http://docs.ansible.com/ansible/latest/haproxy_module.html22:25
dmsimardapparently supports draining22:26
*** bobh has joined #openstack-infra22:26
clarkbyum update is in progress and kernel version in yum output looked good22:27
*** smatzek has quit IRC22:27
clarkbI'll walk through all of the git backends unless someone else wants to get exposure to haproxy too. Then probably just yolo reboot git.o.o later tonight when things go quiet?22:28
openstackgerritJames E. Blair proposed openstack-infra/zuul feature/zuulv3: WIP: Support cross-source dependencies  https://review.openstack.org/53080622:28
clarkbopen to better suggestions re yolo reboot22:29
*** jistr has quit IRC22:29
clarkbrebooting git08 now22:30
dmsimardgit.o.o is the backend for review.openstack.org, right ?22:30
dmsimardor review.openstack has it's own git repos ?22:30
clarkbGerrit is the authoritative canonical source of git repos and it does its own hosting. It then mirrors to our official mirrors, git.openstack.org22:31
clarkblots of things point at git.o.o to take load off of gerritthough22:31
clarkbLinux git08.openstack.org 3.10.0-693.11.6.el7.x86_64 #1 SMP Thu Jan 4 01:06:37 UTC 2018 x86_64 x86_64 x86_64 GNU/Linux22:32
dmsimardlgtm22:32
clarkbthat is the same version thta dmsimard posted above. Apache is running I am adding the server back to the haproxy backends list22:32
pabelangerclarkb: dmsimard: sorry, just back online for a moment, my turn not to feel well, should have mentioned something a little sooner22:32
clarkbpabelanger: no problem just didn't want the opportunity to be wasted. If you are unwell you should work on feeling better andI'm sure we can figure out another time to haproxy later22:33
dmsimardclarkb: reboot takes like a minute if there's nothing silly like an overdue fsck or a selinux relabel22:33
dmsimardclarkb: is there a reason why git.o.o runs centos if it's just for haproxy ?22:34
clarkbdmsimard: we had to use centos for cgit on the backends and I think it was easier to keep everything the same for a logical service rather than mixing22:34
pabelangerclarkb: yah, I am hoping it is only a 24 hour thing, but I can read up on backscroll for haproxy commands after you deal with kernels22:34
clarkbdmsimard: however we could run it on ubuntu or other distros for sure22:34
*** ihrachys has quit IRC22:34
*** ihrachys has joined #openstack-infra22:35
dmsimardalso, git.o.o hasn't been updated to centos 7.422:35
clarkbdmsimard: oh also it may hvae been that haproxy grew out of load demands later22:35
dmsimardso the update is larger than just a few packages22:35
clarkbdmsimard: 95 packages22:35
clarkbI skimmd them nothing too crazy22:35
dmsimardI see 387 packages on a yum update here22:35
corvusif we think there might be a big enough downtime, we could replace the server22:35
clarkboh this was git08 at 95 packages22:35
clarkbgit.o.o could've drifted different22:35
dmsimardclarkb: ah, yeah, git.o.o hasn't been updated to 7.4 and there are other CVEs fixes that landed in 7.4, we want that updated asap :)22:36
corvusbring up new one, test it, add cname record, watch for issues, drop old cname, delete old server22:36
dmsimardthat sounds like a plan22:36
clarkbcorvus: I don't expect it to be very long but unexpected fsck could certainly ruin that ssumption22:36
clarkband ya its a stateless node so doing the move rather than reboot is probably simplest22:37
clarkbI'll continue on with the backends for now22:37
dmsimardI believe there's a systemd/systemctl command to make sure there's no unexpected fsck but I'm always paranoid about that kind of stuff22:37
corvusi have limited time to contribute, so i'm not going to object to yolo reboot :)22:37
clarkb07 is now disabled, just waiting for it to settle22:38
*** Apoorva has quit IRC22:38
*** bobh has quit IRC22:38
*** Apoorva has joined #openstack-infra22:39
corvusi'm suddenly reminded there seems to be some kind of bug with the mta on those servers; we may be about to get a lot of queued mail.22:39
*** wolverineav has quit IRC22:39
dmsimardclarkb: I don't believe there'd be an unexpected fsck -- there's no check interval on the root partition22:40
clarkbcorvus: exciting22:41
dmsimardhttp://paste.openstack.org/raw/638471/22:41
clarkbwe probably do want to keep an eye on git performance as I expect its quite a bit syscall nound22:43
clarkb*bound22:43
clarkbbut I also don't know that we have much choice22:43
*** slaweq has joined #openstack-infra22:44
*** nicolasbock has joined #openstack-infra22:44
dmsimardclarkb: the kernel flag might not be enabled, you can check with /proc/cpuinfo22:44
dmsimardclarkb: it depends on how the hypervisor exposes the cpu, it might not be detected as a vulnerable (and thus the patch isn't applied)22:44
clarkb07 rebooting now22:44
dmsimardif the patch is applied, you'll have a line in cpuinfo with the insecure bug thing22:45
clarkbdmsimard: I don't think the cpu flags determine it, its any intel cpu22:45
clarkboh you mean to check it /me looks22:45
*** jascott1 has joined #openstack-infra22:46
*** armax has joined #openstack-infra22:47
dmsimard"bugs        : cpu_insecure" is what should be shown if the patch is effective iiuc22:47
clarkbits not, which now begs the question why22:47
clarkbit thinks it has an intel cpu22:47
clarkbor is this the thing like qemu was saying where they need to expose things up thorugh before it knows?22:48
dmsimardlet me look at a bare metal centos machine, sec22:48
clarkb07 is up enabling it in haproxy22:48
dmsimardhmm, I'm not seeing that flag on a bare metal machine either22:50
clarkbdmsimard: maybe the backported patch to centos didn't pull that stuff in?22:51
clarkbor maybe we aren't actually booting with pti :)22:51
corvushttps://www.reddit.com/r/sysadmin/comments/7o1769/go_check_your_proccpuinfo_it_will_contain_cpu/ds6xpq8/22:51
corvussome person on reddit says rhat masks the flag22:51
dmsimardbleh22:51
dmsimardthat kind of sucks22:51
clarkbya thats unfortunate22:51
corvuswould be nice to have better confirmation than that, but that seems like a thing worth exploring22:51
dmsimardthere has to be a way to see if it's loaded :/22:52
clarkbI'm continuing to 06, patching in general not a bad thing and until we get to git.o.o isn't impactful22:52
dmsimardsure22:52
clarkbbut yes I agree we sould sort out some way of confirming we are patched22:52
clarkb(and not just in the package but in actively running kernel)22:52
corvushttps://askubuntu.com/questions/992137/how-to-check-that-kpti-is-enabled-on-my-ubuntu22:54
corvusevery one of those things fails22:54
clarkbcorvus: even the /proc/config.gz grep?22:55
dmsimardred hat put out a script but it's for ovirt and it does postgresql things..22:55
corvusclarkb: that path doesn't exist22:56
clarkbah must be ubuntu things22:56
corvusclarkb: /boot/config-3.10.0-693.11.6.el7.x86_64 does but doesn't have the key22:56
dmsimardI have this:22:56
dmsimard# grep -i isolation /boot/config-3.10.0-693.11.6.el7.x86_6422:56
dmsimardCONFIG_MEMORY_ISOLATION=y22:56
*** edmondsw has quit IRC22:56
dmsimardnot sure if it's the same as the page table thing22:56
clarkbdmsimard: its not CONFIG_PAGE_TABLE_ISOLATION is what we want22:57
*** edmondsw has joined #openstack-infra22:57
clarkbunless during patching things got renamed or merged together22:57
dmsimardyeah that's what I'm wondering22:57
dmsimardthis is kernel 3.10 so..22:57
dmsimardcould be something else entirely too22:58
corvusmemory_isolation is an old flag22:58
clarkb06 is rebooting now22:58
dmsimardnothing in dmesg either ..22:59
clarkbanyone know anyone at red hat? >_>23:00
dmsimardlol23:01
*** edmondsw has quit IRC23:01
dmsimardok I'll go ahead and ask in a secret channel -_-23:02
clarkbI'm on to 05 now23:03
*** smatzek has joined #openstack-infra23:04
* dmsimard is back from the secret channel23:06
dmsimardThis gives us what we want: /sys/kernel/debug/x86/pti_enabled23:07
*** dave-mccowan has joined #openstack-infra23:08
clarkbdmsimard: if the file is present that means we are good?23:09
dmsimardit should be present and 123:09
dmsimard(do a meow on it)23:10
clarkboh ya its a 1 thanks23:10
*** ijw has quit IRC23:10
clarkbok so this did work just not in an obvious way23:10
dmsimardyeah.23:10
clarkbthat is reassuring though, thank you for figuring it out /me goe sback to patching23:10
dmsimardI should probably hang out in that secret channel more often, bunch of interesting discussions.23:11
clarkb05 rebooting now23:11
dmsimardcorvus: your suspicion of email spam was correct23:12
clarkband now on to 0423:13
*** erlon has quit IRC23:14
*** flwang has quit IRC23:14
*** ijw has joined #openstack-infra23:15
fungidebian has posted their advisory and corresponding kernel packages now23:16
dmsimardDoes ubuntu get their kernels from debian ? I'm not up to date on the relationship between debian and ubuntu anymore23:17
*** hongbin has quit IRC23:18
clarkbI want to say the kernel effectively forks off when ubuntu grabs it from debian23:18
clarkbI'm going to go ahead and start on 03 in parallel23:19
*** slaweq has quit IRC23:19
*** smatzek has quit IRC23:19
*** smatzek has joined #openstack-infra23:19
*** jistr has joined #openstack-infra23:19
clarkbno usn updates23:21
*** smatzek has quit IRC23:21
fungiyeah, debian and ubuntu work on their kernel packages somewhat indepndently23:21
*** smatzek has joined #openstack-infra23:21
clarkbif they don't get them up today those of you up early in the morning should check and get ansible to apply them then we can work on rolling reboots through the day23:21
clarkbactually if they get them up before 0600 then automated updates will install it iirc23:22
fungias clarkb indicates, i too think ubuntu inherits a kernel from debian/testing at the time the next ubuntu release is stabilized and then forks from there23:22
fungilike most packages in ubuntu23:22
fungibut they _may_ be treating the kernel specially, since ubuntu tends to enable a lot of additional bits23:23
fungiso not certain23:23
*** flwang has joined #openstack-infra23:24
*** smatzek has quit IRC23:26
clarkbwell thats exceptionally curious23:28
clarkbgit05 got 693-11.1 not 693-11.623:28
clarkband does not show pti enabled23:28
clarkbit thinks it is up to date too23:28
dmsimardyay.. I think I just finished the last important update on our infra23:29
dmsimards/our/rdo/23:29
clarkbI'm going to finish up 04-01 then go back and figure out hwat is up with 0523:29
clarkbunless someone wnts to look into 0523:29
dmsimardI'll take a look at 0523:29
openstackgerritStibbons proposed openstack-dev/pbr master: Support v<semver> version  https://review.openstack.org/53126223:30
clarkbdmsimard: maybe it is pointing at stale mirrors?23:30
*** dave-mccowan has quit IRC23:31
openstackgerritStibbons proposed openstack-dev/pbr master: Support v<semver> version  https://review.openstack.org/53116123:32
clarkb03 and 04 are done and have pti enabled23:33
clarkbdoing 01 and 02 now23:33
dmsimardclarkb: yeah, stale mirror.. doing a yum clean all and a yum update yields the update we want.. is git05 still out of haproxy ?23:33
clarkbdmsimard: it is not, I will remove it and let you know23:34
dmsimardok23:34
clarkbdmsimard: it is out now23:34
dmsimardok I'll reboot after the update completes23:34
dmsimardWhen ubuntu updates are out, it'd be cool to check if "/sys/kernel/debug/x86/pti_enabled" is also a thing on Ubuntu so we could have a quick ansible task that checks which machines we have updated and which are still pending23:35
dmsimardsince I guess we still have a mix of 14.04 and 16.0423:36
clarkb+123:36
fungiyup23:36
fungigreat idea, dmsimard23:36
fungithough we could probably || chain the shell commands in the task23:37
fungiif the methods of checking between them do have to differ23:37
dmsimardeh, either that or a proper playbook that detects centos/14.04/16.0423:37
dmsimardhopefully they're all the same23:37
*** tosky has quit IRC23:39
*** mtreinish has quit IRC23:39
dmsimardclarkb: git05 is up to date and pti enalbed23:39
clarkbdmsimard: thanks23:39
clarkbI will reenable it in haproxy now23:40
clarkbdone23:40
clarkbstill waiting on yum on 01 and 0223:40
clarkbthen I'll double check all 8 are patched and haproxy has them all then we can think about a yolo reboot on 0823:41
dmsimardI'm going to take a break and eat some nutrients23:41
clarkber not on 08, git.o.o23:41
*** mtreinish has joined #openstack-infra23:42
openstackgerritStibbons proposed openstack-dev/pbr master: Support v<semver> version  https://review.openstack.org/53116123:45
clarkb02 and 01 reenabled in haproxy23:45
openstackgerritStibbons proposed openstack-dev/pbr master: Support v<semver> version  https://review.openstack.org/53116123:46
*** rkukura has quit IRC23:46
clarkbconfirmed all 8 have pti enabled23:47
*** rkukura has joined #openstack-infra23:47
*** slaweq has joined #openstack-infra23:49
openstackgerritJames E. Blair proposed openstack-infra/zuul feature/zuulv3: WIP: Support cross-source dependencies  https://review.openstack.org/53080623:49
clarkband haproxy show stat doesn't show any in MAINT mode23:49
clarkbany other infra-roots around for possible yolo update and reboot of git.o.o?23:49
clarkbthe alternative is to build a new server and update dns records. The two downsides to this are the hour long dns record ttls and potential corporate firewall rule problems23:50
clarkbbut downside to yolo reboot is any jobs fetching git content during the couple minutes it takes to reboot will have a sad23:50
openstackgerritMike Perez proposed openstack-infra/puppet-statusbot master: Add configuration for thanks/success page url  https://review.openstack.org/53118523:50
corvusclarkb: have we accepted git.o.o as a service that needs special firewall rules?23:51
corvusi mean, it's all standard ports23:51
corvusclarkb: i have to run now and can't help today, but can pitch in some tomorrow23:52
clarkbcorvus: thats a good point http(s) should all be proxied at $corp23:52
clarkbwhereas 29418 on gerrit is special23:52
mtreinishclarkb: tempest also pulls the openstack logo from git.o.o for glance image create from remote location23:53
*** slaweq has quit IRC23:53
mtreinishso if any runs hitting those tests during the reboot window will fail23:53
corvusclarkb: it does have 9418, but heck, we've talked about dropping that altogether.  i'm not sure it's worth worrying about.23:53
clarkbya23:53
*** jistr has quit IRC23:54
clarkbI think I will go ahead and just build a new server23:54
corvusmtreinish: i like the overloading of the term image23:54
clarkbI'm going to update the dns ttls to 5 minute snow as step 923:54
clarkber step 023:54
*** jistr has joined #openstack-infra23:55
mtreinishcorvus: heh, yeah I always like that too23:55
clarkbits more cloudy to do it this way too23:55
*** stakeda has joined #openstack-infra23:55
clarkbdns updated, off to boot a new server now23:56
*** nicolasbock has quit IRC23:58

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!