Tuesday, 2019-07-02

*** yamamoto has joined #openstack-infra00:01
*** mattw4 has quit IRC00:10
*** weifan has joined #openstack-infra00:13
donnydclarkb: can you run fio in another instance to give me back some data. I need to know if I am actually making progress or not00:17
*** lseki has quit IRC00:17
donnydhttps://www.irccloud.com/pastebin/zgLxUacE/00:18
donnydThis is what I am getting in an instance now... was more like 4k iops in the vm00:18
clarkbdonnyd: ya let me see00:19
donnydIf this is not good, then I have a bit more tuning I can do to get a couple more k if need be... but that will be about all it can do00:20
clarkbdonnyd: hrm doesn't seem to like my credentials00:21
clarkbdid those get wiped/changed during the updates?00:21
donnydyea00:21
donnydThe ones I gave you for openstackci and openstackzuul should be working00:21
donnydhttps://www.irccloud.com/pastebin/3tUjxjKl/00:21
donnydChecking00:22
donnydThey are there00:22
clarkbdonnyd: did the passwords change?00:23
donnydno00:23
clarkbI updated the passwords on our side, but using the ones you set and the ones I set doesn't seem to work00:23
clarkbok let me recheck my clouds.yaml00:23
*** weifan has quit IRC00:28
clarkbyup got it was os_project_domain_name chagned from Default to default I think00:29
*** weifan has joined #openstack-infra00:29
clarkbdonnyd: so I should go ahead and reupload our image as qcow2 then boot an instance and run some tests?00:32
donnydyea you can go ahead00:32
*** rfarr_ has quit IRC00:32
donnydThe openstackci project should be good to go00:33
*** rfarr_ has joined #openstack-infra00:33
*** tdasilva has quit IRC00:33
clarkbwhat about openstackzuul?00:34
clarkb(thats where I'll upload to since it will run the test instances)00:34
*** tdasilva has joined #openstack-infra00:34
*** hamzy has quit IRC00:35
donnydstill working on the ipv6 part00:35
*** smarcet has left #openstack-infra00:36
clarkbok image is uploading00:37
*** gyee has quit IRC00:37
clarkbexcept I am doing the raw image woops00:37
clarkbI'll stop it, delete and start again00:37
*** jamesmcarthur has joined #openstack-infra00:40
*** weifan has quit IRC00:41
donnyd:)00:42
donnydThey would still have worked00:43
clarkbya but the upload takes forever because the raw image size isn't small :/00:43
ianwauristor: is there anything sensitive in the audit logs?00:43
*** ijw has quit IRC00:43
clarkbdonnyd: oh re fio are you asking us to run in one of your cloud instances or on another cloud?00:46
clarkb(I read it previously as on your cloud but not sure if that is what you emant now looking at it again)00:46
donnydyea, I was curious to see00:46
donnydIf I am 50K IOPS away, then I may need to procure some better gear for this workload00:47
clarkbI think devstack jobs actually record that /me looks for some links00:48
*** rfarr_ has quit IRC00:50
*** weifan has joined #openstack-infra00:50
*** rfarr_ has joined #openstack-infra00:50
clarkbdonnyd: if you download http://logs.openstack.org/12/665512/3/check/tempest-full/c5e5fe3/controller/logs/dstat-csv_log.txt.gz then put it in https://lamada.eu/dstat-graph/ it will give you some idea00:51
clarkbthere are spikes in the 10k ranges but mostly its much loewr than that00:51
*** jamesmcarthur has quit IRC00:51
clarkbI think the interesting thing will be to run tempest on these nodes though as there are other factors too00:52
donnydok, so I am surely in the ballpark00:52
clarkband I'll work on that. Though dinner is served here now so maybe that will have to be a morning activity00:52
donnydOk, the openstackzuul project should be good to hook00:53
donnydtesting now00:53
*** jamesmcarthur has joined #openstack-infra00:53
donnydyour image is done00:54
clarkbya I just created my ssh key too00:54
clarkbif you are happy with the network I'll boot a node, start nova unittests then eat dinner00:54
donnydit works00:56
donnydtest complete00:56
donnyd:)00:56
clarkbk booting instance then thanks00:56
donnydtry again00:57
donnydselinux policies with containers and NFS have been killling me00:57
clarkbimage has no associated data? trying again00:58
donnydyea, its an selinux thing on the glance side.. not sure why it does that yet...00:58
donnydIt still does seem to take quite a while for nova to pull down the image00:59
*** ccamacho has quit IRC01:00
clarkbI think nova may convert it to raw by default too01:00
donnydyea that could be it... but I don't think its at that stage yet.01:01
donnydhttps://www.irccloud.com/pastebin/TrMfvC4I/01:01
clarkbya not sure how it gets logged01:02
donnydi stand corrected01:02
donnyd42436     277229 46.3  0.0 476904 36284 ?        Sl   01:01   0:04 qemu-img convert -t none -O raw -f qcow2 /var/lib/nova/instances/_base/e956c6c8779473194d0964072073a2efe98d85d9.part /var/lib/nova/instances/_base/e956c6c8779473194d0964001:02
*** weifan has quit IRC01:02
clarkbthat can be changed (and force it to boot off qcow2, not sure how that will affect io though)01:02
*** ccamacho has joined #openstack-infra01:02
donnydIf performance is close enough, I won't really worry too much. Will only have to wait once a day01:03
clarkband I'm in01:04
*** yamamoto has quit IRC01:04
donnydyey... that only took 3 days of refactoring to get back to where we started01:04
donnydLOL01:04
donnydAfter this test is done, I will bring up the rest of the hypervisors and we will be in business01:09
*** tdasilva_ has joined #openstack-infra01:10
*** jamesmcarthur has quit IRC01:10
*** jamesmcarthur has joined #openstack-infra01:10
*** imacdonn has quit IRC01:10
clarkbthe nova unittests are running01:11
*** imacdonn has joined #openstack-infra01:11
*** tdasilva has quit IRC01:12
*** jamesmcarthur has quit IRC01:15
donnydenjoy your dinners01:15
*** hongbin has joined #openstack-infra01:16
clarkbhttp://paste.openstack.org/show/753733/ looks much bette01:17
clarkbparticularly the db tests that are listed with times there01:17
*** redrobot has quit IRC01:17
clarkbpreviously worst case was like 160s now 49s01:17
clarkbwe'll need to run tempest too which I can do in the morning01:18
clarkbthanks!01:18
donnydI will take a 100% improvement01:18
donnydOk cool01:18
donnydI will work on bringing up the rest of the gear then. Api's might be up and donwn01:18
clarkbtotal runtime was 501s before now 397s01:19
donnydSo that is closer to what I saw the other clouds doing01:19
donnydthey were somewhere in that 400 range01:19
clarkbyup01:20
donnydWell enjoy your evening. I am hopeful tomorrow will be the day01:21
donnydIs it ok if I take down your instance?01:22
corvusclarkb: i'd like to add dstat_graph to jobs (see https://review.opendev.org/610100 ) but that just clones from github... we need a way to avoid doing that in every post playbook... got any ideas?  add a new caching proxy for random github projects?  DIB element to cache it on our images?  beg a DD to package it?  install docker and run it in a docker container?01:22
donnydI have a couple sysctls on the nfs server that require a rebbot01:22
donnydIf its running tests, it's not a big deal... can be done any time01:23
clarkbcorvus: we could clone it into our images01:23
clarkbdonnyd: yes you can take down the instance01:23
donnydok thanks01:23
corvusclarkb: yeah, that feels like the most actionable thing i can think of.  we can write the zuul-jobs role to optionall take a path to a cached version, otherwise clone from github, so it's still univesally usable01:24
clarkb++01:24
clarkbI'll also reset passwords to their values in hiera tomorrow morning01:24
clarkbhrm maybe I hsould just go ahead and do that now?01:25
corvusk, i'll plan on doing that.  but right now, i'm halfway through "leviathan wakes" and that's not nearly far enough.01:25
clarkbha01:25
clarkbya I'll reset passwords to hiera values now so that others can interact with the fortnebula cloud if necessary01:25
clarkbthats done. Values in hiera are correct now however I also need to update clouds.yaml for the project_domain_name change from Default to default. That will happen tomorrow01:30
donnydawesome. see you tomorrow clarkb01:30
donnydI will have the rest of things up in the morning01:31
*** rfarr_ has quit IRC01:33
*** rfarr_ has joined #openstack-infra01:34
*** jamesmcarthur has joined #openstack-infra01:37
*** jamesmcarthur has quit IRC01:41
*** ccamacho has quit IRC01:57
*** ccamacho has joined #openstack-infra01:58
*** ccamacho has quit IRC01:58
*** ccamacho has joined #openstack-infra01:59
*** rfarr_ has quit IRC02:05
*** rfarr_ has joined #openstack-infra02:06
*** hwoarang has quit IRC02:09
*** rfarr_ has quit IRC02:09
*** rfarr_ has joined #openstack-infra02:10
*** ccamacho has quit IRC02:21
*** ccamacho has joined #openstack-infra02:22
*** ccamacho has quit IRC02:31
*** ccamacho has joined #openstack-infra02:31
*** bhavikdbavishi has joined #openstack-infra02:32
*** bhavikdbavishi1 has joined #openstack-infra02:35
*** ccamacho has quit IRC02:36
*** bhavikdbavishi has quit IRC02:36
*** bhavikdbavishi1 is now known as bhavikdbavishi02:36
*** jamesmcarthur has joined #openstack-infra02:37
*** ccamacho has joined #openstack-infra02:38
*** yamamoto has joined #openstack-infra02:41
*** rfarr_ has quit IRC02:42
*** tdasilva_ has quit IRC02:42
*** hwoarang has joined #openstack-infra02:42
*** rfarr_ has joined #openstack-infra02:43
*** tdasilva_ has joined #openstack-infra02:43
*** tdasilva_ has quit IRC02:44
*** tdasilva_ has joined #openstack-infra02:45
*** rfarr_ has quit IRC02:45
*** rfarr_ has joined #openstack-infra02:46
*** tdasilva_ has quit IRC02:48
*** ccamacho has quit IRC02:48
*** tdasilva_ has joined #openstack-infra02:48
*** rfarr_ has quit IRC02:50
*** rfarr_ has joined #openstack-infra02:50
*** ccamacho has joined #openstack-infra02:51
auristorianw: I'm back02:54
*** ccamacho has quit IRC02:55
*** ccamacho has joined #openstack-infra02:56
ianwauristor: hey; i hope we captured enough in logs to be helpful02:56
*** yamamoto has quit IRC02:57
auristorI'm looking at the July 1 14:42 case.  openafs "vos" needs a mode to timestamp the verbose output :(02:57
*** rfarr_ has quit IRC02:59
*** rfarr_ has joined #openstack-infra03:00
auristorwhat host is 104.130.4.160 ?03:02
ianwthat's mirror.iad.rax.opendev.org ... there's a dmseg log in the tar file that should have the utc dmesg for that and our gra1 kafs mirror03:03
auristorgrep 536870950 afs*.dfw.dafileserver.audit.log | grep -v "CODE 0"03:04
auristorwill give you what I'm looking at03:04
auristorThe 102 errors are VNOVNODE which means the cache manager was requesting a file that no longer exists.03:05
auristorThe 106 errors are the VOFFLINE messages.03:05
*** ccamacho has quit IRC03:06
auristorwhat the audit log tells me is that the kafs client is only reading from afs02.dfw and isn't failing over to afs01.dfw03:07
*** yamamoto has joined #openstack-infra03:07
auristorwhich is host 38.108.68.37 ?03:07
ianwhrmm... not sure what's up the reverse dns03:08
ianwthat looks vexxhosty ... looking around03:09
auristor38.108.68.37 received a VOFFLINE (106) from afs02.dfw and retried the request to afs01.dfw which succeeded03:10
ianwhttps://opendev.org/opendev/system-config/src/branch/master/inventory/openstack.yaml#L50903:10
ianwmirror01.sjc1.vexxhost.openstack.org03:10
auristoropenafs 1.6 ?03:11
ianwyep03:11
auristor:)03:11
*** ykarel|away has joined #openstack-infra03:11
auristorkafs is retrying sometimes.  It did at the 10:49:02 when afs02.dfw returned VOFFLINE03:13
auristorinteresting.  it retried the first vnode which received a VOFFLINE but it didn't retry the requests for the next three vnodes03:14
auristorI won't be able to discuss with dhowells until Wednesday.  I know he is racing to finish some other work for the next merge window.03:16
ianwno probs ... do you think we have enough in there for now that we should disable the audit logs?03:17
auristoryes03:17
*** hamzy has joined #openstack-infra03:18
auristorThe only other thing you could share would be the FileLog which will provide the timestamps at which the volume callback messages were scheduled for delivery.03:18
*** rfarr_ has quit IRC03:21
*** rfarr_ has joined #openstack-infra03:21
ianwauristor: ok i can grab them too, just a min03:22
*** ykarel|away is now known as ykarel03:24
*** diablo_rojo has quit IRC03:26
*** diablo_rojo has joined #openstack-infra03:27
*** ccamacho has joined #openstack-infra03:29
ianwauristor: at top level in http://people.redhat.com/~iwienand/ ... for completeness I'll regenerate the tarball with them03:30
*** diablo_rojo has quit IRC03:32
*** rfarr_ has quit IRC03:34
auristorto give you some explanation of some of the FileLog messages03:34
*** rfarr_ has joined #openstack-infra03:34
*** ccamacho has quit IRC03:35
auristorAn entry such as "CheckHost_r: Probing all interfaces of host 147.135.133.8:7001 failed, code -1" indicates that a rpc was received from the indicated cache manager but rpc the fileserver issued to the client's callback service timed out before it was responded to.03:35
auristorthat is indicative of an openafs cache manager experiencing a soft-deadlock03:36
*** tdasilva_ has quit IRC03:37
*** psachin has joined #openstack-infra03:37
auristor"BVCB: volume callback for Host 00007F6671273CC8 (23.253.125.17:7001) failed" might be a client that has shutdown, a NAT port mapping that has timed out, or other communication difficulties.03:37
*** tdasilva_ has joined #openstack-infra03:38
auristor"fssync: breaking all call backs for volume 536870983" is logged when the volserver takes a volume offline for a potential data change such as a re-clone operation.03:38
*** rfarr_ has quit IRC03:39
ianwok, so that's mostly the points of interest?03:40
*** rfarr_ has joined #openstack-infra03:40
auristorThese clients appear to be experiencing soft-deadlocks regularly03:41
auristor     10 23.253.125.17:700103:41
auristor     10 91.123.202.253:700103:41
auristor     15 37.153.172.158:700103:41
auristor     20 158.69.80.87:700103:41
auristor     21 104.239.192.5:700103:41
auristor     21 149.202.161.140:700103:41
auristor     23 38.108.68.37:700103:41
auristor     29 104.130.4.160:700103:41
auristor     30 147.135.133.8:700103:41
auristor     32 38.145.32.16:700103:41
auristor     33 199.204.45.64:700103:41
ianwheh, i'd say that's just about all of them? :)03:41
auristorthats only the top dozen :)03:42
*** whoami-rajat has joined #openstack-infra03:42
*** rfarr_ has quit IRC03:43
*** rfarr_ has joined #openstack-infra03:44
auristorI would say that the cache is too small on those clients03:44
*** jamesmcarthur has quit IRC03:49
*** jamesmcarthur has joined #openstack-infra03:50
*** igordc has quit IRC03:51
*** yamamoto has quit IRC03:54
*** tdasilva_ has quit IRC03:55
*** jamesmcarthur has quit IRC03:57
*** ykarel is now known as ykarel|afk03:59
*** kjackal has joined #openstack-infra04:01
ianwhrm04:01
*** yamamoto has joined #openstack-infra04:02
ianw23.253.125.17 is files.openstack.org ... i wonder if we've even upped the cache on that?04:02
ianwnot really ... /afs:/var/cache/openafs:1000000004:04
*** ykarel|afk has quit IRC04:06
auristorits the -stat cache size that most often leads to soft-deadlocks04:09
*** rfarr_ has quit IRC04:12
*** rfarr_ has joined #openstack-infra04:13
*** udesale has joined #openstack-infra04:14
*** yamamoto has quit IRC04:20
*** hongbin has quit IRC04:21
*** rfarr_ has quit IRC04:22
*** rfarr_ has joined #openstack-infra04:23
*** ricolin has joined #openstack-infra04:25
*** jamesmcarthur has joined #openstack-infra04:29
*** ykarel|afk has joined #openstack-infra04:34
*** jamesmcarthur has quit IRC04:36
*** ramishra has joined #openstack-infra04:36
*** rfarr_ has quit IRC04:36
*** rfarr_ has joined #openstack-infra04:37
*** yamamoto has joined #openstack-infra04:57
*** ykarel|afk is now known as ykarel05:02
*** dansmith has quit IRC05:04
*** jamesmcarthur has joined #openstack-infra05:05
*** yamamoto has quit IRC05:07
*** rfarr_ has quit IRC05:09
*** jamesmcarthur has quit IRC05:10
*** rfarr_ has joined #openstack-infra05:10
*** rfarr_ has quit IRC05:20
*** rfarr_ has joined #openstack-infra05:21
*** dansmith has joined #openstack-infra05:22
*** ykarel_ has joined #openstack-infra05:26
*** ykarel has quit IRC05:28
*** rfarr_ has quit IRC05:30
*** rfarr_ has joined #openstack-infra05:30
*** rfarr_ has quit IRC05:34
*** rfarr_ has joined #openstack-infra05:35
*** udesale has quit IRC05:35
*** ykarel_ is now known as ykarel|meeting05:36
*** udesale has joined #openstack-infra05:36
*** kjackal has quit IRC05:36
*** jamesmcarthur has joined #openstack-infra05:42
*** rfarr_ has quit IRC05:46
*** rfarr_ has joined #openstack-infra05:46
*** jamesmcarthur has quit IRC05:47
*** tdasilva has joined #openstack-infra06:00
*** rfarr_ has quit IRC06:01
*** rfarr_ has joined #openstack-infra06:01
*** rakhmerov has quit IRC06:02
ianw#status log restarted afs file servers without auditlog.  see notes in https://etherpad.openstack.org/p/opendev-mirror-afs for log recent log bundle06:05
openstackstatusianw: finished logging06:05
*** yamamoto has joined #openstack-infra06:09
*** jtomasek has joined #openstack-infra06:15
*** kjackal has joined #openstack-infra06:19
*** dchen has quit IRC06:21
*** dchen has joined #openstack-infra06:21
*** rfarr_ has quit IRC06:25
*** rfarr_ has joined #openstack-infra06:25
*** rfarr_ has quit IRC06:28
*** rfarr_ has joined #openstack-infra06:29
*** slaweq has joined #openstack-infra06:30
*** pgaxatte has joined #openstack-infra06:31
openstackgerritIan Wienand proposed opendev/zone-opendev.org master: Add mirror-update01.opendev.org  https://review.opendev.org/66861006:31
*** yamamoto has quit IRC06:33
*** Emine has joined #openstack-infra06:36
*** Emine has quit IRC06:37
*** yamamoto has joined #openstack-infra06:37
*** yamamoto has quit IRC06:37
openstackgerritIan Wienand proposed opendev/system-config master: Add mirror-update01.opendev.org server  https://review.opendev.org/66861206:38
*** rfarr_ has quit IRC06:39
*** rfarr_ has joined #openstack-infra06:40
*** yamamoto has joined #openstack-infra06:41
*** rfarr_ has quit IRC06:45
*** rfarr_ has joined #openstack-infra06:46
*** raukadah is now known as chandankumar06:46
*** dpawlik has joined #openstack-infra06:49
*** pgaxatte has quit IRC06:53
*** pgaxatte has joined #openstack-infra06:54
*** pgaxatte has quit IRC06:54
*** bhavikdbavishi has quit IRC07:01
*** pgaxatte has joined #openstack-infra07:02
*** pgaxatte has quit IRC07:03
*** pgaxatte has joined #openstack-infra07:04
*** pgaxatte has quit IRC07:04
*** ginopc has joined #openstack-infra07:08
*** xek has joined #openstack-infra07:12
*** pgaxatte has joined #openstack-infra07:15
*** tesseract has joined #openstack-infra07:15
openstackgerritIan Wienand proposed opendev/system-config master: Move rsync mirror updates to new opendev.org mirror-update host  https://review.opendev.org/66778207:17
openstackgerritIan Wienand proposed opendev/system-config master: Add mirror-update01.opendev.org server  https://review.opendev.org/66861207:17
openstackgerritIan Wienand proposed opendev/system-config master: mirror-update: update keytab testing  https://review.opendev.org/66862007:17
*** rfarr_ has quit IRC07:20
*** rfarr_ has joined #openstack-infra07:20
*** witek has joined #openstack-infra07:21
*** tosky has joined #openstack-infra07:22
*** rfarr_ has quit IRC07:22
*** rfarr_ has joined #openstack-infra07:23
openstackgerritIan Wienand proposed openstack/diskimage-builder master: Mailing lists change openstack-dev to openstack-discuss  https://review.opendev.org/66843507:23
openstackgerritMerged openstack/diskimage-builder master: install gnupg2 by default in debian-minimal  https://review.opendev.org/66855407:32
*** yamamoto has quit IRC07:33
*** ricolin_ has joined #openstack-infra07:35
*** dchen has quit IRC07:35
*** dchen has joined #openstack-infra07:37
*** ricolin has quit IRC07:37
openstackgerritIan Wienand proposed opendev/system-config master: Add mirror-update01.opendev.org server  https://review.opendev.org/66861207:38
*** yamamoto has joined #openstack-infra07:38
*** evrardjp is now known as evrardjp_on_holi07:40
*** evrardjp_on_holi is now known as evrardjp_away07:40
*** ociuhandu has joined #openstack-infra07:42
*** tosky has quit IRC07:43
*** tosky has joined #openstack-infra07:43
*** jamesmcarthur has joined #openstack-infra07:44
*** jamesmcarthur has quit IRC07:48
*** e0ne has joined #openstack-infra07:49
*** ricolin__ has joined #openstack-infra07:49
*** rfarr_ has quit IRC07:50
*** jamesmcarthur has joined #openstack-infra07:51
*** rfarr_ has joined #openstack-infra07:51
*** ricolin_ has quit IRC07:51
*** jpich has joined #openstack-infra07:53
*** rpittau|afk is now known as rpittau07:54
*** jamesmcarthur has quit IRC07:55
*** yamamoto has quit IRC07:55
*** rfarr_ has quit IRC07:55
*** rfarr_ has joined #openstack-infra07:56
*** pkopec has joined #openstack-infra07:56
*** yamamoto has joined #openstack-infra07:57
*** lucasagomes has joined #openstack-infra07:59
*** dchen has quit IRC08:00
*** pcaruana has joined #openstack-infra08:01
*** dchen has joined #openstack-infra08:01
*** ralonsoh has joined #openstack-infra08:02
*** dchen has quit IRC08:04
*** udesale has quit IRC08:05
*** dchen has joined #openstack-infra08:05
*** ykarel_ has joined #openstack-infra08:05
*** udesale has joined #openstack-infra08:06
*** ykarel|meeting has quit IRC08:07
*** dchen has quit IRC08:09
*** yamamoto has quit IRC08:09
*** dchen has joined #openstack-infra08:10
*** trident has quit IRC08:10
*** trident has joined #openstack-infra08:13
*** dchen has quit IRC08:15
*** yamamoto has joined #openstack-infra08:16
*** dchen has joined #openstack-infra08:16
*** dchen has quit IRC08:17
*** jamesmcarthur has joined #openstack-infra08:22
*** tkajinam has quit IRC08:26
*** jamesmcarthur has quit IRC08:28
*** rfarr_ has quit IRC08:28
*** rfarr_ has joined #openstack-infra08:29
*** ykarel_ has quit IRC08:32
*** rfarr_ has quit IRC08:35
*** rfarr_ has joined #openstack-infra08:35
*** sshnaidm|afk is now known as sshnaidm|ruck08:36
*** ykarel_ has joined #openstack-infra08:37
*** ykarel_ is now known as ykarel|lunc08:37
*** ykarel|lunc is now known as ykarel|lunch08:38
*** rfarr_ has quit IRC08:41
*** rfarr_ has joined #openstack-infra08:42
*** rfarr_ has quit IRC08:45
*** rfarr_ has joined #openstack-infra08:45
*** bhavikdbavishi has joined #openstack-infra08:46
*** rfarr_ has quit IRC08:48
*** rfarr_ has joined #openstack-infra08:48
*** ykarel|lunch is now known as ykarel08:54
*** jamesmcarthur has joined #openstack-infra09:00
*** yamamoto has quit IRC09:01
*** ricolin_ has joined #openstack-infra09:04
*** rfarr_ has quit IRC09:06
*** rfarr_ has joined #openstack-infra09:06
*** ricolin__ has quit IRC09:07
*** Lucas_Gray has joined #openstack-infra09:11
*** rfarr_ has quit IRC09:12
*** rfarr_ has joined #openstack-infra09:13
*** yamamoto has joined #openstack-infra09:13
*** jaosorior has joined #openstack-infra09:13
*** priteau has joined #openstack-infra09:17
*** tdasilva has quit IRC09:21
*** yamamoto has quit IRC09:22
*** rfarr_ has quit IRC09:23
*** rfarr_ has joined #openstack-infra09:23
*** rfarr_ has quit IRC09:33
*** rfarr_ has joined #openstack-infra09:33
*** ricolin_ is now known as ricolin09:33
*** bhavikdbavishi has quit IRC09:34
*** udesale has quit IRC09:35
*** udesale has joined #openstack-infra09:36
*** rfarr_ has quit IRC09:41
*** rfarr_ has joined #openstack-infra09:41
*** whoami-rajat has quit IRC09:42
*** rfarr_ has quit IRC09:46
*** rfarr_ has joined #openstack-infra09:47
*** ricolin_ has joined #openstack-infra09:47
*** yamamoto has joined #openstack-infra09:47
jrosserE:Failed to fetch http://mirror.ord.rax.opendev.org/debian-security/dists/stretch/main/binary-amd64/Packages.gz  Hash Sum mismatch09:49
*** ricolin has quit IRC09:49
*** rfarr_ has quit IRC09:52
*** yamamoto has quit IRC09:52
*** rfarr_ has joined #openstack-infra09:52
openstackgerritMerged openstack/diskimage-builder master: Mailing lists change openstack-dev to openstack-discuss  https://review.opendev.org/66843509:53
*** yamamoto has joined #openstack-infra09:55
*** priteau has quit IRC09:57
*** ykarel is now known as ykarel|meeting09:58
*** yamamoto has quit IRC09:59
*** roman_g has quit IRC10:02
*** roman_g has joined #openstack-infra10:03
*** jamesmcarthur has quit IRC10:04
*** yamamoto has joined #openstack-infra10:04
*** yamamoto has quit IRC10:04
*** roman_g_ has joined #openstack-infra10:05
*** priteau has joined #openstack-infra10:06
*** psachin has quit IRC10:08
*** roman_g has quit IRC10:08
sshnaidm|ruckwhat does mean "Triggered by: https://review.opendev.org/gitweb?p=openstack/tripleo-ci.git;a=commitdiff;h=None" ? Isn't it from patch?10:09
*** roman_g_ is now known as roman_g10:10
*** rfarr_ has quit IRC10:16
*** rfarr_ has joined #openstack-infra10:16
ianwjrosser: can you give a timestamp?  we put that into rotation yesterday, should be running openafs 1.8.3.  we haven't been seeing issues afaik with the other server running that (dfw.rax) ...10:26
*** tdasilva has joined #openstack-infra10:27
jrosserWe just had a bunch of openstack-ansible jobs fail I think all the same10:27
* jrosser just in the server room, 20mins or so and I can find a log10:28
*** rfarr_ has quit IRC10:28
*** rfarr_ has joined #openstack-infra10:29
*** yamamoto has joined #openstack-infra10:30
*** bhavikdbavishi has joined #openstack-infra10:31
ianwjrosser: ok, we'll want to correlate it i think to any possible release of a volume or logs on the host.  we can also check via logstash.  i gotta disappear now, but will be back for meeting and check on things tomorrow too10:32
*** ykarel|meeting is now known as ykarel10:35
jrosserianw: http://logs.openstack.org/87/668287/1/gate/openstack-ansible-deploy-aio_metal-debian-stable/d3a3b3d/job-output.txt.gz#_2019-07-02_08_35_08_62162510:36
jrosserhttp://logs.openstack.org/18/665218/1/gate/openstack-ansible-deploy-aio_metal-debian-stable/bf8ca5c/job-output.txt.gz#_2019-07-02_08_35_47_26574310:37
*** yamamoto has quit IRC10:37
jrosserhttp://logs.openstack.org/64/666564/2/gate/openstack-ansible-functional-debian-stable/37baf18/job-output.txt.gz#_2019-07-02_08_39_18_67733010:38
jrossertheres more but hopefully thats enough to start with10:38
*** bhavikdbavishi has quit IRC10:40
*** jpich has quit IRC10:43
*** rfarr_ has quit IRC10:43
*** rfarr_ has joined #openstack-infra10:43
*** kjackal has quit IRC10:44
openstackgerritMerged opendev/system-config master: do backports for Debian 'buster'  https://review.opendev.org/66815710:51
*** ginopc has quit IRC10:52
*** udesale has quit IRC10:54
*** ginopc has joined #openstack-infra10:59
*** rfarr_ has quit IRC10:59
*** rfarr_ has joined #openstack-infra11:00
*** udesale has joined #openstack-infra11:00
AJaegerjrosser: I just downloaed the file from your first change and it has the correct hash sum, so might have been a race that is fixed now. Still, I think ianw will want to check log files.11:02
*** kjackal has joined #openstack-infra11:07
jrosserAJaeger: ok, i'll recheck those jobs11:07
*** ykarel is now known as ykarel|afk11:09
*** rfarr_ has quit IRC11:13
*** udesale has quit IRC11:13
*** rfarr_ has joined #openstack-infra11:14
*** ykarel|afk has quit IRC11:16
*** yamamoto has joined #openstack-infra11:30
*** rfarr_ has quit IRC11:31
*** kjackal has quit IRC11:31
*** rfarr_ has joined #openstack-infra11:32
*** rfarr_ has quit IRC11:36
*** rfarr_ has joined #openstack-infra11:36
*** ykarel has joined #openstack-infra11:37
*** tdasilva has quit IRC11:38
*** rh-jelabarre has joined #openstack-infra11:39
*** psachin has joined #openstack-infra11:39
openstackgerritMonty Taylor proposed zuul/zuul master: Spec: Add a Kubernetes Operator for Zuul  https://review.opendev.org/65918011:40
*** priteau has quit IRC11:41
*** jaosorior has quit IRC11:42
*** kjackal has joined #openstack-infra11:43
openstackgerritGraham Hayes proposed zuul/nodepool master: Implement an Azure driver  https://review.opendev.org/55443211:44
*** ykarel is now known as ykarel|meeting11:52
*** jamesmcarthur has joined #openstack-infra12:01
*** pgaxatte has quit IRC12:04
*** jamesmcarthur has quit IRC12:05
*** rfarr_ has quit IRC12:06
*** pgaxatte has joined #openstack-infra12:06
*** rfarr_ has joined #openstack-infra12:06
openstackgerritMonty Taylor proposed zuul/zuul master: Spec: Add a Kubernetes Operator for Zuul  https://review.opendev.org/65918012:16
*** bhavikdbavishi has joined #openstack-infra12:18
*** electrofelix has joined #openstack-infra12:22
*** jcoufal has joined #openstack-infra12:28
*** pkopec has quit IRC12:29
*** yamamoto has quit IRC12:31
*** jamesmcarthur has joined #openstack-infra12:34
*** tdasilva has joined #openstack-infra12:35
*** rlandy has joined #openstack-infra12:35
*** pkopec has joined #openstack-infra12:37
*** jamesmcarthur has quit IRC12:39
*** jaosorior has joined #openstack-infra12:39
*** jamesmcarthur has joined #openstack-infra12:42
*** eharney has quit IRC12:43
*** udesale has joined #openstack-infra12:43
*** jpich has joined #openstack-infra12:45
*** rfarr__ has joined #openstack-infra12:50
*** sthussey has joined #openstack-infra12:53
*** rkukura has quit IRC12:53
*** rfarr_ has quit IRC12:53
*** aaronsheffield has joined #openstack-infra12:54
*** priteau has joined #openstack-infra12:55
*** ricolin_ is now known as ricolin12:56
*** ekultails has joined #openstack-infra12:59
*** ykarel|meeting is now known as ykarel13:01
*** pkopec has quit IRC13:04
*** jamesmcarthur has quit IRC13:05
*** pkopec has joined #openstack-infra13:06
*** priteau has quit IRC13:13
*** ccamacho has joined #openstack-infra13:13
*** priteau has joined #openstack-infra13:13
*** lseki has joined #openstack-infra13:14
*** mriedem has joined #openstack-infra13:14
*** adriancz has joined #openstack-infra13:19
*** ykarel is now known as ykarel|afk13:20
*** eharney has joined #openstack-infra13:26
*** smarcet has joined #openstack-infra13:27
*** rfarr__ has quit IRC13:28
*** yamamoto has joined #openstack-infra13:28
*** smarcet has left #openstack-infra13:28
*** jamesmcarthur has joined #openstack-infra13:32
*** rfarr has joined #openstack-infra13:33
*** sreejithp has joined #openstack-infra13:40
*** bhavikdbavishi has quit IRC13:44
*** liuyulong has joined #openstack-infra13:45
*** yamamoto has quit IRC13:47
*** iurygregory_ has joined #openstack-infra13:49
*** Jeffrey4l_ has joined #openstack-infra13:50
openstackgerritJames E. Blair proposed zuul/nodepool master: Update documentation for k8s in-cluster config  https://review.opendev.org/66867913:51
*** iurygregory has quit IRC13:51
*** Jeffrey4l has quit IRC13:54
mudpuppysup14:14
*** Fidde has joined #openstack-infra14:15
openstackgerritBrian Haley proposed opendev/irc-meetings master: Create a meeting for Networking OVN project  https://review.opendev.org/66801314:17
openstackgerritMerged openstack/project-config master: Load extra Zuul config from zuul-jobs  https://review.opendev.org/66856314:18
openstackgerritPete Vander Giessen proposed openstack/project-config master: Add microstack project.  https://review.opendev.org/66868114:18
sshnaidm|ruckMaybe problem with iax docker mirror: TTPError: 502 Server Error: Bad Gateway for url: http://mirror.iad.rax.opendev.org:8082/cloudflare/registry-v2/docker/registry/v2/blobs/sha256/6a/614:21
sshnaidm|ruckhttp://logs.openstack.org/71/668671/1/check/tripleo-ci-centos-7-scenario004-standalone/30b6866/logs/undercloud/home/zuul/standalone_deploy.log.txt.gz#_2019-07-02_14_07_1014:21
fricklercloudflare reported network issues earlier14:23
openstackgerritBrian Haley proposed opendev/irc-meetings master: Create a meeting for Networking OVN project  https://review.opendev.org/66801314:27
openstackgerritMerged zuul/nodepool master: Allow nodepool for using in-cluster configs  https://review.opendev.org/66817214:32
*** Lucas_Gray has quit IRC14:32
*** Lucas_Gray has joined #openstack-infra14:33
openstackgerritJames E. Blair proposed openstack/project-config master: Cache dstast_graph on nodepool images  https://review.opendev.org/66868314:33
corvusclarkb, mordred: ^ does that look right?14:33
*** redrobot has joined #openstack-infra14:33
mordredcorvus: probably?14:34
mordredcorvus: also - you know dag archived dstat14:35
mordred(it'll obviously keep working - but such is life)14:35
corvusmordred: yeah, when there's an alternative, i'm sure we can switch :)14:36
*** yamamoto has joined #openstack-infra14:39
*** yamamoto has quit IRC14:39
*** sreejithp has quit IRC14:43
*** pkopec is now known as pkopec|lunch14:46
*** zbr has quit IRC14:53
*** zbr has joined #openstack-infra14:54
*** pgaxatte has quit IRC14:57
*** evrardjp_away has quit IRC14:58
*** evrardjp has joined #openstack-infra14:59
*** mattw4 has joined #openstack-infra15:02
*** Fidde has quit IRC15:03
*** yamamoto has joined #openstack-infra15:04
*** edmondsw_ has joined #openstack-infra15:04
*** gyee has joined #openstack-infra15:04
openstackgerritLon Hohberger proposed openstack/diskimage-builder master: ironic-agent: Use targetcli & system-python-devel on rhel8  https://review.opendev.org/66868715:07
openstackgerritMerged openstack/project-config master: Cache dstast_graph on nodepool images  https://review.opendev.org/66868315:10
*** Lucas_Gray has quit IRC15:10
openstackgerritLon Hohberger proposed openstack/diskimage-builder master: ironic-agent: Use targetcli & python3-devel on rhel8  https://review.opendev.org/66868715:11
openstackgerritJeff Liu proposed zuul/zuul-operator master: Add playbooks folder and zuul job that installs k8s  https://review.opendev.org/66802915:17
*** bhavikdbavishi has joined #openstack-infra15:22
*** smarcet has joined #openstack-infra15:22
*** liuyulong has quit IRC15:22
*** smarcet has left #openstack-infra15:22
*** e0ne has quit IRC15:27
*** bhavikdbavishi1 has joined #openstack-infra15:29
*** yamamoto has quit IRC15:30
*** bhavikdbavishi has quit IRC15:30
*** bhavikdbavishi1 is now known as bhavikdbavishi15:30
openstackgerritJames E. Blair proposed zuul/zuul-jobs master: Move project config in-repo  https://review.opendev.org/66869815:31
openstackgerritJames E. Blair proposed openstack/project-config master: Remove zuul-jobs config  https://review.opendev.org/66869915:32
corvushrm, zuul did not pick up that change the way i expected; let me see if we've reloaded the tenant config15:35
openstackgerritMerged zuul/nodepool master: Update documentation for k8s in-cluster config  https://review.opendev.org/66867915:37
*** rkukura has joined #openstack-infra15:39
*** tdasilva has quit IRC15:40
*** tdasilva has joined #openstack-infra15:41
corvusah, it did work, it just had an invalid config, and didn't report it because it's not attached to the pipeline yet (we don't have a .* project in the zuul tenant)15:45
clarkbwe set a lot of IP addresses in local.conf during devstack jobs15:52
clarkbmakes pulling a local.conf from a job to run locally a major pain15:52
clarkb(if anyone is wonderig why this doesn't work for them, remove all the host ip stuff in local.conf)15:52
openstackgerritJames E. Blair proposed openstack/project-config master: Zuul-jobs: load extra config in openstack tenant  https://review.opendev.org/66870315:53
corvusclarkb, mordred, AJaeger: ^ i got ahead of myself, can you review that soon?15:53
*** ykarel|afk is now known as ykarel15:54
clarkbI can15:54
donnydclarkb: How hard would it be to build up a freebsd box for zuul?15:55
donnydoh and everything should be good to go15:55
donnydI may swap out the glance backend, but that *should* be transparent15:56
clarkbdonnyd: freebsd as the node running zuul or freebsd images for running tests on with zuul?15:56
*** yamamoto has joined #openstack-infra15:56
clarkbI think freebsd as node running zuul is likely pretty straightforward and may just work15:56
donnydThe nfs mount point is seemingly causing some major issues with podman or docker based containers for glance only15:56
donnydtesting code on a freebsd image15:57
donnydhttps://review.opendev.org/#/c/668694/115:57
clarkbI don't think diskimage-builder can build images for freebsd today. It could be updated to do that likely, possibly requiring a freebsd builder node to host those builds15:57
donnydJust curious because Tim asked15:57
donnydSo we just need resources and time is what I am hearing15:58
clarkbWe can also point zuul/nodepool at preexisting images in a cloud. There are a bunch of reasons why we like to rebuild images daily to update caches and not worry about security updates though so if we do that we'll need to think about thei mplications of it15:58
clarkbdonnyd: ya, basically nothing inherent in zuul's design should prevent it. Its more a matter of it hasn't been done by us before so likely lots to learn :)15:58
*** ginopc has quit IRC16:00
*** jamesmcarthur has quit IRC16:03
*** kjackal has quit IRC16:03
AJaegercorvus: +2A on 66870316:05
*** mattw4 has quit IRC16:06
*** yamamoto has quit IRC16:06
corvusclarkb, donnyd: my order of preference would be: 1) freebsd has something like 'debootstrap' which just happens to magically work on ubuntu and we can use the existing dib methodology [seems unlikely].  2) we have a freebsd builder which can run something like 'debootstrap' on freebsd and add dib support for that [this ought to work -- they make those release isos *somehow* right?]  3) we dig out16:08
corvusthe old nodepool 0.1 code that does snapshot based builds and make a new nodepool builder that uses snapshots instead of dib.  [a lot of work (this is mostly a nodepool coding project), but gives us an extremely flexible builder]16:08
openstackgerritMarios Andreou proposed openstack/diskimage-builder master: WIP: Adds release->rhel8 for ironic-agent pkg-map  https://review.opendev.org/66843616:09
corvusor, i gues 4) we just use upstream images for freebsd and forget about our local caching, etc.16:09
*** gyee has quit IRC16:12
*** iurygregory_ has quit IRC16:16
openstackgerritMerged openstack/project-config master: Zuul-jobs: load extra config in openstack tenant  https://review.opendev.org/66870316:17
*** jamesmcarthur has joined #openstack-infra16:17
*** rkukura_ has joined #openstack-infra16:17
*** rkukura has quit IRC16:20
*** rkukura_ is now known as rkukura16:20
*** rpittau is now known as rpittau|afk16:21
*** gyee has joined #openstack-infra16:25
*** lucasagomes has quit IRC16:32
*** jpich has quit IRC16:35
openstackgerritNate Johnston proposed openstack/project-config master: Mark networking-ovn-tempest-dsvm-ovs-release as voting  https://review.opendev.org/66870816:36
clarkbtempest is running in the test node now16:38
donnydon mine?16:38
clarkbdonnyd: yup16:38
clarkbdevstack took 27 minutes without using cached data which is inline with other clouds iirc16:38
clarkbso looking good so far16:38
donnydWell that is good news16:39
clarkbhttp://paste.openstack.org/show/753783/ compared to http://logs.openstack.org/38/662538/1/check/tempest-full/fafc02f/controller/logs/devstacklog.txt.gz#_2019-05-31_22_06_26_02216:39
clarkbit was actually quite a bit slower in the job I grabbed my local.conf from ^16:39
clarkbnow to see how tempest does16:40
*** rlandy has quit IRC16:41
clarkbthe trick for reusing a zuul job's local.conf is to comment out error on clone and all the entries with HOST in them16:41
clarkbdevstack will autodetect ip addresses if you comment out the HOST entries16:41
*** rlandy has joined #openstack-infra16:42
clarkbthis is going to take a bit of time so I am going to get a bike ride in now while waiting. Back in a bit16:42
donnydOk cool16:42
donnydLMK how it works out16:42
*** rlandy has quit IRC16:43
*** rlandy has joined #openstack-infra16:44
*** yamamoto has joined #openstack-infra16:46
*** priteau has quit IRC16:48
*** udesale has quit IRC16:49
openstackgerritMerged zuul/nodepool master: Remove devstack plugin functional test jobs  https://review.opendev.org/66715616:52
openstackgerritBrian Haley proposed openstack/project-config master: Mark networking-ovn-tempest-dsvm-ovs-release voting  https://review.opendev.org/66871416:56
*** yamamoto has quit IRC16:56
*** ykarel is now known as ykarel|away16:58
openstackgerritMerged zuul/nodepool master: Reparent nodepool-functional-openstack-src  https://review.opendev.org/66799517:01
openstackgerritJames E. Blair proposed zuul/zuul-jobs master: Move in-repo test jobs into zuul-tests.d  https://review.opendev.org/66871617:01
*** tesseract has quit IRC17:01
*** jamesmcarthur has quit IRC17:02
*** jamesmcarthur has joined #openstack-infra17:02
openstackgerritJames E. Blair proposed zuul/zuul-jobs master: Move in-repo test jobs into zuul-tests.d  https://review.opendev.org/66871617:04
*** goldyfruit has quit IRC17:04
*** ricolin_ has joined #openstack-infra17:05
*** ykarel|away has quit IRC17:05
*** goldyfruit has joined #openstack-infra17:05
openstackgerritJames E. Blair proposed zuul/zuul-jobs master: Add zuul-jobs-tox-linters  https://review.opendev.org/66805217:07
*** ricolin has quit IRC17:07
*** weifan has joined #openstack-infra17:14
*** ricolin_ has quit IRC17:14
*** ociuhandu_ has joined #openstack-infra17:14
*** ociuhandu_ has quit IRC17:15
*** mattw4 has joined #openstack-infra17:15
*** ardaguclu has joined #openstack-infra17:15
*** witek has quit IRC17:15
*** weifan has quit IRC17:15
*** ardaguclu has left #openstack-infra17:16
*** ociuhandu has quit IRC17:18
*** electrofelix has quit IRC17:26
*** hemna has joined #openstack-infra17:30
*** sshnaidm|ruck is now known as sshnaidm|afk17:31
hemnahey guys, working on a ci job definition for ceph-iscsi.  how to I get zuul to allow me to clone a repo not owned/run by openstack as part of the job?17:31
hemnahttps://review.opendev.org/#/c/667108/4/.zuul.yaml@16717:32
hemnaI just want devstack to allow cloning of a couple github.com/open-iscsi projects17:32
*** ociuhandu has joined #openstack-infra17:33
*** ociuhandu has quit IRC17:38
*** ralonsoh has quit IRC17:42
*** goldyfruit has quit IRC17:43
corvushemna: if it's not something that we're gating, then it should be installed from a release rather than from source17:44
hemnacorvus: there is no dist package for them17:44
hemnathey are installed from their clone directly17:44
hemnathat's the only option for tcmu-runner, targetcli-fb and ceph-iscsi projects17:45
corvushemna: openstack software should only depend on released software17:45
hemnathat doesn't help me test at all17:46
corvushemna: are you planning on releasing something that depends on unreleased software?17:46
hemnathe official ceph-iscsi docs say to install from github17:46
*** mattmceuen has left #openstack-infra17:46
hemnahttp://docs.ceph.com/docs/mimic/rbd/iscsi-target-cli-manual-install/17:47
hemnathat's just how it's done17:47
corvushemna: we're sort of in the realm of policy here, and previously openstack has established this policy that to be friendly to users and delpoyers, it would depend only on released software, so that our install instructions didn't say "download this source code from github and install it yourself".  so i think if you want to expand the scope to cover depending on software without formal releases,17:48
corvusthat's probably a good topic for the TC17:48
clarkbdonnyd: so tempest ran to completion but with 24 failed jobs. Some of the problems are related to ambiguous network specifications booting instances. There seems to have been some OOMing too :/17:48
hemnawell, if I can't get a gate job for this, then the ceph-iscsi driver is dead for cinder17:48
clarkbdonnyd: I'll specify the IP addrs that devstack should auto figure out in case taht is necessary for the networking pieces and the OOMing might be fallout of tempest going sideways17:48
donnydOOMing?17:49
donnydOut of Memory?17:49
AJaegerconfig-core, could you review corvus' change https://review.opendev.org/668699 to move jobs to zuul-jobs, please?17:49
clarkbdonnyd: ya17:49
corvushemna: well, don't let it die here -- i'm just trying to point you to the right venue for the conversation17:49
clarkbdonnyd: I don't think the cloud is to blame for that as the amount of memory on the host matches our other instances17:50
hemnaso there is no way to even test this then17:50
corvushemna: maybe an outcome is one of the distros sees this as a problem they can fix and packages it up.17:50
AJaegerconfig-core, and also review https://review.opendev.org/#/c/668052  , please17:50
hemna:(17:51
donnydHrm.. well the nodes they run on have plenty of the memory17:51
clarkbdonnyd: ya I'll rerun and make sure this wasn't fallout of the network changes I made in local.conf17:51
corvushemna: there are definitely ways to test this.  i can tell you several ways to make that job work, but if you pushed up those patches, we'd have to -1 them because they are out of policy.  so, please, please, please, just talk to the TC about this, okay? :)17:51
hemnaI can't change distros17:51
donnydwhen you say "ambiguous network specifications", what do you mean17:52
donnydalso does nested virt need to be enabled?17:52
AJaegerhemna: openSUSE has these three packages, so test using openSUSE-1517:52
*** gyee has quit IRC17:52
clarkbdonnyd: nova was failing to boot instances because a specific network wasn't specified and it was ambiguous according to the nova logs17:52
clarkbdonnyd: its the error you get if there are several possible networks that can be used and nova doesn't know which to use (note this is in the nested cloud not your cloud)17:53
hemnahow do I change the job definition to force using openSUSE then ?17:53
clarkbdonnyd: no we disable nested virt in these jobs as it is unreliable so it is fine to leave that off17:53
clarkbhemna: you can specify a new nodeset. If you look in devstack's zuul config there are examples of this17:54
AJaegerhemna: devstack-single-node-opensuse-150 is in devstack repo defined, keystone has keystone-dsvm-functional-federation-opensuse15, have a look how they do it17:55
AJaegerhemna: there are other jobs defined as well... if you need a review, I'm sure dirk and cmurphy can help17:55
donnydIs there anything you need me to do?17:55
clarkbdonnyd: at this point I don't think so. I'll work on narrowing down the failures if they are reproduceable and let you know what I find17:56
hemnaok, so just specify nodeset: devstack-single-node-opensuse-150 as part of the job?17:56
*** jamesmcarthur has quit IRC17:56
donnydok sounds great17:56
openstackgerritMerged zuul/zuul-jobs master: Move project config in-repo  https://review.opendev.org/66869817:57
AJaegerhemna: that is a job name, check how the job does it17:57
AJaegerbbl17:58
*** ijw has joined #openstack-infra17:58
*** weifan has joined #openstack-infra18:02
openstackgerritMerged zuul/zuul-jobs master: Move in-repo test jobs into zuul-tests.d  https://review.opendev.org/66871618:03
*** tdasilva has quit IRC18:03
*** armax has joined #openstack-infra18:04
openstackgerritMerged zuul/zuul master: Filter out unprotected branches from builds if excluded  https://review.opendev.org/66666418:05
*** gyee has joined #openstack-infra18:06
*** goldyfruit has joined #openstack-infra18:10
*** ykarel|away has joined #openstack-infra18:13
*** igordc has joined #openstack-infra18:13
hemnaok well, it looks like the devstack-plugin-ceph isn't supported on opensuse18:20
hemnaso that kills it.18:20
hemna:(18:20
*** weifan has quit IRC18:21
*** weifan has joined #openstack-infra18:22
*** ykarel|away has quit IRC18:23
donnydclarkb: when you are done with your tempest testing please LMK18:23
donnydI am going to swap from nfs to swift for the backend18:23
*** weifan has quit IRC18:23
donnydI can upload your image with a quickness, so downtime will be minimal18:23
*** weifan has joined #openstack-infra18:24
clarkbdonnyd: will do18:24
clarkbits probably another hour away? Not 100% sure18:24
donnydNo rush. I am still doing a little tuning18:25
clarkbI've discovered that tempest should run as a tempest user (at least it does in the gate) and I don't have a tempest user at the end of my devstack run so sorting that out18:25
clarkbtempest was a lot easier to run a couple years ago :/18:26
*** ardaguclu has joined #openstack-infra18:28
*** ardaguclu has quit IRC18:28
*** ardaguclu has joined #openstack-infra18:29
clarkbanyone know where the tempest user comes from? I see d-g adds it in the legacy jobs but not fiding it in tempest or devstack v3 roles18:31
clarkbaha there it is18:31
clarkband we do that because the user running tempest needs to be able to sudo?18:32
clarkbAlso another piece of data, I've realized I didn't create a swap file which our normal jobs will do. This may explain the OOMing18:34
clarkbbut also does that mean tempest can't run without digging into swap?18:35
clarkbthat seems like a pretty bad regresion18:35
mordredclarkb: one would think that devstack would create the tempest user - or that there would be a tempest plugin for devstack that would create the tempest user ... but one would think many things18:36
clarkbmordred: nope devstack does not do it. The devstack jobs do though18:37
clarkband I agree that seems backwards18:37
clarkbif tempest wants its own user then give it one during devstacking18:37
clarkbfwiw I've decided I don't ened to make a tempest user because the zuul user can do passwordless sudo so just running it as zuul18:37
clarkbI'm also going to be quite sad if I have to create a swapfile just to run devstack successfully18:38
clarkbI think base devstack + tempest should not be swapping18:39
clarkbI guess I can go ahead and preempitvely create that file while tempest is running18:39
clarkbbecause more interested in performance results than cloud memory use18:39
* clarkb does this18:39
*** weifan has quit IRC18:40
*** weifan has joined #openstack-infra18:40
*** weifan has quit IRC18:40
*** weifan has joined #openstack-infra18:41
*** weifan has quit IRC18:41
*** bhavikdbavishi has quit IRC18:41
*** weifan has joined #openstack-infra18:41
*** weifan has quit IRC18:42
*** weifan has joined #openstack-infra18:42
*** weifan has quit IRC18:43
*** weifan has joined #openstack-infra18:43
*** bhavikdbavishi has joined #openstack-infra18:43
*** weifan has quit IRC18:44
clarkbSKIPPED: nova-network is gone. I wonder if we can delete those tests then?18:44
*** weifan has joined #openstack-infra18:44
*** armax has quit IRC18:44
*** weifan has quit IRC18:44
*** weifan has joined #openstack-infra18:44
clarkb~15 minutes to our weekly meeting. Join us in #openstack-meeting18:45
*** weifan has quit IRC18:45
*** ardaguclu has quit IRC18:45
*** weifan has joined #openstack-infra18:45
*** ardaguclu has joined #openstack-infra18:46
*** weifan has quit IRC18:46
*** weifan has joined #openstack-infra18:46
*** weifan has quit IRC18:47
*** ardaguclu has left #openstack-infra18:48
clarkbhemna: there are other ways to address it. I think corvus' point was to ensure that going down the road of other methods was not going to surprise the TC18:49
clarkb(or our users)18:49
*** weifan has joined #openstack-infra18:50
*** weifan has quit IRC18:50
*** weifan has joined #openstack-infra18:51
openstackgerritNicholas Shobe proposed opendev/bindep master: Add dependency version examples to README  https://review.opendev.org/66874018:54
*** weifan has quit IRC18:55
*** smarcet has joined #openstack-infra18:56
clarkbya test node is almost 200MB into swap now18:56
clarkbjust cross 200MB swap now18:57
clarkbso if anyone wants ot improve zuul throughput: fixing the swapping in tempest/devstack jobs will likely significantly speed up all of those jobs thus freeing up resources for other tests18:57
* clarkb tucks that away for next time peopel complain18:57
clarkbI should write an email about it after our meeting /me makes a note to do this18:58
*** weifan has joined #openstack-infra19:00
*** pkopec|lunch has quit IRC19:05
*** xek has quit IRC19:08
clarkbnow up to 360MB swappage19:09
donnydShould we include swap in the flavor?19:16
*** kmarc has quit IRC19:16
clarkbdonnyd: it isn't necessary our jobs will create a swap file if there is no device. (we have to do that in a bunch of clouds)19:16
*** kmarc has joined #openstack-infra19:16
clarkbit should do the right thing if you add swap to the falvor though19:16
donnydNot sure if that will actually make it perform better though19:18
clarkbya I don't think we are too worried about swap performance. It is there as a safety valve but should be avoied19:18
*** tosky has quit IRC19:19
ianwclarkb: yep, i think you're on the right track, ord seems not to have 1.8.3 packages.  investigating ...19:20
yoctozepto2019-07-02 19:17 Err:15 http://mirror.ord.rax.opendev.org/ubuntu bionic-updates/universe amd64 Packages19:21
yoctozepto  File has unexpected size (1320734 != 1321741). Mirror sync in progress? [IP: 2001:4801:7824:103:be76:4eff:fe10:9c2 80]19:21
yoctozepto  Hashes of expected file:19:21
yoctozepto   - Filesize:1321741 [weak]19:21
yoctozepto   - SHA256:8b11598039ba0cb9f3a9717e3119feaa865f113ed9e51d173935b38763cf1aea19:21
yoctozepto   - SHA1:4f2e9579a49005777803b51eb4fb905881ae0db1 [weak]19:21
yoctozepto   - MD5Sum:820f33b9f056e1c9e2579db15d997b7d [weak]19:21
yoctozepto  Release file created at: Tue, 02 Jul 2019 16:32:17 +000019:21
yoctozeptosorry for multiline19:21
*** jamesmcarthur has joined #openstack-infra19:21
yoctozeptoseems to be failing still19:22
clarkbyoctozepto: ya that is related to ianw's message just prior to yours19:23
clarkbyoctozepto: it is using the wrong version of openafs unexpectedly19:23
yoctozeptoclarkb: thanks, I skimmed the message, lone 'ord' did not catch my eye ;/19:24
*** panda has quit IRC19:27
*** jamesmcarthur has quit IRC19:28
*** panda has joined #openstack-infra19:30
openstackgerritIan Wienand proposed opendev/system-config master: openafs-client: ensure latest package and reorder install  https://review.opendev.org/66874519:33
*** bhavikdbavishi has quit IRC19:33
ianwyoctozepto / clarkb: ^ i think ord.rax was in an odd pocket where i'd created the server a while ago, but we didn't put it into rotation until just a few days ago.  i think a totally new server would have got 1.8.3 packages installed19:34
ianwhowever, i'll still need to manual reboot after that anyway19:34
donnydThis is just a random bit of info in my testing19:36
donnydcpu utilization for my controllers when using NFS for glance were at 60(ish) percent19:37
donnydno real reason.. just was19:37
donnydjust swapped out the backend and it dropped by down to where it should at... at 8%19:38
openstackgerritMonty Taylor proposed zuul/zuul master: Spec: Add a Kubernetes Operator for Zuul  https://review.opendev.org/65918019:46
*** jamesmcarthur has joined #openstack-infra19:47
ijwSo it seems that networking-vpp's mirror on github is present but no longer synced, and the date looks supiciously like it was the date of conversion from openstack/networking-vpp to x/networking-vpp.  Is there something I'm supposed to have done, there?19:48
clarkbijw: we stopped syncing all x/.* repos to github beacuse they are no logner part of openstack. The idea there being we reduce confusion over what is and isn't openstack by being specific. What projects can do instead of relying on the default openstack/ replication is setup zuul jobs that replicate to arbitrary git remotes including github19:50
clarkbthere was an email thread about it on the openstack-discuss list which I can dig up if you give me a couple minutes19:50
*** jamesmcarthur has quit IRC19:50
yoctozeptoianw: should not you leave linux-headers-... with present instead of latest? (understanding-wise because it is pinned anyway)19:50
fungisoon the openstack github org admins will go through and mark all those stale repos as "archived"19:50
clarkbijw: http://lists.openstack.org/pipermail/openstack-discuss/2019-April/005007.html19:51
ijwYeah, ok - the change doesn't surprise me, but leaving the repos present but orphaned caught someone out19:51
clarkbianw: re devstack + tempest swapping I'll try to put an email together to openstack-discuss about it after lunch19:52
ianwyoctozepto: my theory is that should pull in the latest kernel updates, which will be applied with the latest openafs package when a host reboots ... theory anyway :)19:52
fungiijw: the alternative was to delete them all immediately, but then that makes it harder t transfer them to other orgs later19:52
clarkbdonnyd: infra-root http://logs.openstack.org/38/662538/1/check/tempest-full/fafc02f/job-output.txt.gz#_2019-05-31_22_56_53_731181 compared to http://paste.openstack.org/show/753791/ on fortnebula. I think we are in the same ballpark especially given that devstack runs quicker on fortnebula than that job ran it19:53
*** Fidde has joined #openstack-infra19:53
fungiseems so19:53
corvusclarkb, mordred: https://review.opendev.org/668699 can merge now (and then the rest of the changes can be self-testing)19:53
*** factor__ has joined #openstack-infra19:54
ianwyoctozepto: actually, no i guess you're right because it's using {{ ansible_kernel }}19:54
*** bnemec has quit IRC19:54
ianwso it will only ever install the version currently running19:54
yoctozeptoianw: exactly19:54
openstackgerritIan Wienand proposed opendev/system-config master: openafs-client: ensure latest package and reorder install  https://review.opendev.org/66874519:55
*** icarusfactor has quit IRC19:56
clarkbyoctozepto: ianw I think they are equivalent in this case beacuse the version is specified19:56
clarkb+2 either way19:56
ianwyeah, but it might try re-installing every run, i dunno19:56
donnydWell after some tweaking, it's likely to get even faster19:57
yoctozeptoianw, clarkb: yeah, they are functionally equivalent, though present will not trigger check for update (which is unnecessary) and better reflects the task action19:58
*** tdasilva has joined #openstack-infra19:58
donnydclarkb: So the test was good then?20:00
clarkbdonnyd: ya20:03
donnydSo what is next?20:04
clarkbI'll boot a mirror after lunch and we'll add it to nodepool20:05
clarkbthat will build images then wwe can tell zuul to use testnodes there20:05
clarkbdonnyd: I dont think you need to do anything at this point but I'll let you know if that changes20:06
*** bnemec has joined #openstack-infra20:07
donnydok cool20:07
donnydWell I will just keep humming away over here then20:07
openstackgerritMerged openstack/project-config master: Remove zuul-jobs config  https://review.opendev.org/66869920:11
donnydclarkb: can I reboot your instance while i finish off my last bit of testing?20:12
clarkbyes20:13
donnydok thanks20:13
ianwclarkb/corvus : in https://review.opendev.org/#/q/topic:mirror-update-ansible+(status:open+OR+status:merged) i think the dns change can merge, then there's the follow-on to add to inventory.  if i could get some eyes, i can merge that today with babysitting20:13
ianwoh and a follow-on for testing slightly longer keytabs, to make sure i'm handling line-breaks ok20:14
*** factor__ has quit IRC20:14
openstackgerritMerged zuul/zuul-jobs master: Add zuul-jobs-tox-linters  https://review.opendev.org/66805220:14
*** eharney has quit IRC20:15
*** factor has joined #openstack-infra20:16
*** factor has quit IRC20:18
*** factor has joined #openstack-infra20:24
*** ijw has quit IRC20:32
*** Lucas_Gray has joined #openstack-infra20:34
*** Lucas_Gray has quit IRC20:37
*** ijw has joined #openstack-infra20:37
openstackgerritMatt McEuen proposed opendev/irc-meetings master: Adjust Airship meeting time  https://review.opendev.org/66875520:40
*** psachin has quit IRC20:42
*** evrardjp has quit IRC20:51
*** Lucas_Gray has joined #openstack-infra20:52
*** evrardjp has joined #openstack-infra20:52
openstackgerritIan Wienand proposed opendev/system-config master: openafs-client: ensure latest package and reorder install  https://review.opendev.org/66874520:52
ianwyoctozepto / clarkb: ^ i really do like our testing finding issues that hadn't thought of :)  although i should have put a comment first time around because i remember now why it was done like that20:53
*** ekultails has quit IRC20:58
*** pcaruana has quit IRC20:58
*** eharney has joined #openstack-infra20:59
*** Lucas_Gray has quit IRC21:03
*** Lucas_Gray has joined #openstack-infra21:04
*** sshnaidm|afk has quit IRC21:06
*** priteau has joined #openstack-infra21:07
*** priteau has quit IRC21:07
*** sshnaidm has joined #openstack-infra21:08
openstackgerritClark Boylan proposed opendev/system-config master: Add fortnebula cloud to cloud launcher  https://review.opendev.org/66825121:09
*** sshnaidm is now known as sshnaidm|afk21:09
openstackgerritClark Boylan proposed opendev/system-config master: Use qcow2 images on fortnebula  https://review.opendev.org/66875721:09
clarkbinfra-root ^ those are two small housekeeping changes for fortnebula cloud21:09
clarkbI'm going to work on booting the mirror now21:09
*** jcoufal_ has joined #openstack-infra21:12
clarkbhrm all-clouds.yaml hasn't updated on bridge.o.o?21:13
*** ijw has quit IRC21:13
*** ijw has joined #openstack-infra21:14
*** mattw4 has quit IRC21:15
*** jcoufal has quit IRC21:16
donnydclarkb: Do I have 5 minutes?21:17
donnydjust wanted to run one more test... i promise21:17
*** mattw4 has joined #openstack-infra21:17
donnydIf you are ready to pull the trigger it can wait21:17
clarkbdonnyd: yup21:18
clarkbI've discovered a flaw in our config of the bastion21:18
clarkbso figuring that out first21:18
*** ijw has quit IRC21:19
donnydOk cool21:19
clarkbinfra-root the issue with bridge clouds.yaml files is service-bridge.yaml is only ever run during testing21:20
clarkbdo we need to add that to run_all.sh or does someone else understand this better than me?21:21
clarkbI wasn't around when we split that stuff up so may have missed some details21:21
mordredclarkb: hrm21:22
*** jcoufal_ has quit IRC21:22
*** eharney has quit IRC21:22
*** jamesmcarthur has joined #openstack-infra21:22
mordredclarkb: I'm pretty sure corvus groks it all - but is also afk atm21:22
mordredclarkb: looking21:23
mordredclarkb: can you say more words about what issue you're observing?21:23
clarkbmordred: I updated the bridge all-clouds.yaml and clouds.yaml files to include the new fortnebula cloud. But those chagnes in opendev/system-config are not applied to bridge.o.o21:23
clarkbmordred: I think they are not applied because service-bridge.yaml which makes those changes is not run in production but is run in testing21:24
clarkbI'm not sure if that is or was an intentional change as part of the base.yaml split up21:25
mordredclarkb: well - ok, yeah - I'm also not sure what the difference between bridge and service-bridge is21:25
*** bnemec has quit IRC21:25
*** jamesmcarthur has quit IRC21:27
*** bnemec has joined #openstack-infra21:27
clarkbif anyone else recalls let me know. We'll need that sorted so that cloud launcher works and configures security groups and keys and stuff21:31
clarkbI guess I can modify the clouds.yaml files manually while we search memory21:31
donnydwell... my power just went out... so it may be a few minutes before things come back up21:32
donnydmy generator is sitting outside waiting to be installed21:32
clarkboh no21:33
clarkbI'll work on something else for a bit then21:34
donnydI just can't win21:34
donnydLOL21:34
clarkbwhen it rains it pours21:34
clarkbwe had some really fun weather here yesterday. A tornado (really weak one) even knocked over some trees in portland21:35
donnydwell... I guess I will find out how things are handled when stuff goes bad21:35
clarkbfungi: http://logs.openstack.org/62/666162/1/check/openstackci-beaker-puppet-4/3e13a54/job-output.txt.gz#_2019-07-02_19_16_53_145830 the error appears to be a lack of zlib dev? I'm gonna update bindep to see if that makes it happy21:36
openstackgerritClark Boylan proposed opendev/system-config master: Add zlib dev packages to bindep  https://review.opendev.org/66875921:39
donnydmy ups was 2 minutes 2 short21:39
donnydLOL21:39
openstackgerritClark Boylan proposed opendev/puppet-mediawiki master: Canonicalize clone URLs  https://review.opendev.org/66616221:40
clarkbfungi: ^ lets see if that is happier21:40
fungiclarkb: thanks! i wonder why it stopped working21:48
clarkbfungi: well and its odd that the job doesn't seem to fail on all system-config changes? I don't quite undersatnd that aspect of it yet21:48
clarkboh wait21:50
clarkbits because they are different repos21:50
clarkbI bet the change was dropping global bindep fallback21:51
fungioh, indeed... if it had no bindep.txt then the timing is about right21:52
clarkbya so system-config change likely completely unnecessary21:52
fungisuppose we can just add it in the same puppet-mediawiki change21:55
fungibut i'll have to do that latter... being dragged away again now21:55
fungi(or feel free to push it yourself if you want)21:55
ianwclarkb: i feel like service-bridge should probably run as part of run-all.sh ; it was running as part of base.yaml in https://review.opendev.org/#/c/656871/ ... my feeling it was it was just overlooked21:56
openstackgerritClark Boylan proposed openstack/openstack-zuul-jobs master: Make infra puppet beaker jobs parent to legacy-base  https://review.opendev.org/66876121:56
clarkbinfra-root ^ that should fix puppet jobs21:57
clarkbfungi: ^ I'm fixing it via zuul job config then allpuppet modules will be happy21:57
clarkbianw: ok I'll work on that change now then21:57
openstackgerritClark Boylan proposed opendev/puppet-mediawiki master: Canonicalize clone URLs  https://review.opendev.org/66616222:00
clarkbianw: and other infra-roots do you think all of playbooks/bridge should go into playbooks/service-bridge?22:03
clarkbI'm not understanding why they are different22:03
clarkbmost of the content of playbooks/bridge was moved into playbooks/service-bridge in the change ianw linked22:03
clarkbI'm happy to propose running them as separate playbooks for now too22:03
clarkboh I think maybe it was because we want all of the service stuff to happen after base.yaml22:04
clarkbbut some bridge stuff has to happen first22:04
clarkbin that case I'll push it up as two separate things22:04
donnydok, its back up22:05
openstackgerritClark Boylan proposed opendev/system-config master: Apply service-bridge.yaml in run_all.sh  https://review.opendev.org/66876222:06
clarkbianw: ^ there22:06
ianw+2, thanks22:09
ianwjust need to do morning runs, then i'll come back and make sure ord is fixed first priority22:09
clarkbok now to write that email about devstack+tempest swapping22:14
donnydgoing to dinner... be back later22:15
corvusianw, clarkb: +3 668762 -- we probably just forgot it, or it's possible that we thought we were going to get to the point where we would actually have a service-bridge job by the time that merged22:20
corvusanyway, easy enough to revert that along with the addition of the service-bridge job when we're ready :)22:20
clarkbcorvus: ++22:23
clarkbianw: I've identified at least one cause of memory pressure on tempest jobs: cinder-backup consumes 785MB of memory according to dstat22:24
clarkbadding that into my email22:24
clarkbhrm dstat may not look at resident memroy but virtual memory?22:32
clarkbthe peakmem tracker shows different numbers22:32
clarkbcorvus: can you review https://review.opendev.org/#/c/668761/ too? the -1 from zuul there is due to the mirror problem that ianw is fixing22:38
*** jcoufal has joined #openstack-infra22:39
corvusclarkb: +3 should i recheck?22:39
corvus(i think the answer is yes so i did)22:40
corvus(we can at least roll the dice again while ianw is out)22:40
clarkbya recheck is fine, thanks22:40
*** rfarr_ has joined #openstack-infra22:43
*** rfarr has quit IRC22:46
*** panda has quit IRC22:52
openstackgerritJames E. Blair proposed zuul/zuul-jobs master: Add base integration roles  https://review.opendev.org/66806122:54
openstackgerritJames E. Blair proposed zuul/zuul-jobs master: Shard project-pipeline config  https://review.opendev.org/66876322:54
openstackgerritJames E. Blair proposed zuul/zuul-jobs master: Add base integration roles  https://review.opendev.org/66806122:57
*** panda has joined #openstack-infra22:59
*** tkajinam has joined #openstack-infra23:02
openstackgerritMerged openstack/openstack-zuul-jobs master: Make infra puppet beaker jobs parent to legacy-base  https://review.opendev.org/66876123:02
clarkbfungi's wiki change passes tests now https://review.opendev.org/#/c/666162/3 if anyone else wants to review taht real quick23:06
*** slaweq has quit IRC23:07
*** Fidde has quit IRC23:11
*** dchen has joined #openstack-infra23:12
*** aaronsheffield has quit IRC23:13
*** mriedem has quit IRC23:20
*** rfarr_ has quit IRC23:22
*** rfarr_ has joined #openstack-infra23:23
*** mattw4 has quit IRC23:23
openstackgerritLogan V proposed openstack/diskimage-builder master: Remove immutable attributes before cleanup  https://review.opendev.org/66876623:26
openstackgerritLogan V proposed openstack/diskimage-builder master: Remove immutable attributes before cleanup  https://review.opendev.org/66876623:26
*** rfarr_ has quit IRC23:27
*** rfarr_ has joined #openstack-infra23:28
ianwbuilding 1.8.3 on ord so this gets fixed up now23:30
*** sthussey has quit IRC23:32
*** lseki has quit IRC23:33
*** dchen has quit IRC23:37
*** dchen has joined #openstack-infra23:39
*** dchen has joined #openstack-infra23:41
*** dchen has quit IRC23:42
*** dchen has joined #openstack-infra23:44
ianw#status log mirror.ord.rax.opendev.org rebooted and running openafs 1.8.3; should fix up periodic "hash mismatch" errors seen in jobs23:46
openstackstatusianw: finished logging23:46
*** dchen has quit IRC23:46
*** dchen has joined #openstack-infra23:47
*** dchen has quit IRC23:49
*** dchen has joined #openstack-infra23:51
*** weifan has quit IRC23:53
*** weifan has joined #openstack-infra23:53
*** weifan has quit IRC23:57
ianwinfra-root: if i could get some eyes on https://review.opendev.org/#/c/668612/ that will allow me to do mirror-update migration; the host is up and ready23:58

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