Sunday, 2017-02-12

*** tonanhngo has quit IRC00:00
*** huzhengchuan has joined #openstack-kolla00:06
*** caowei has joined #openstack-kolla00:13
*** mannamne has quit IRC00:31
openstackgerritSteven Dake proposed openstack/kolla-kubernetes master: DNM bisecting kolla-ansible to find regression  https://review.openstack.org/43269700:38
*** caowei has quit IRC00:41
sbezverksdake: already found commit which broke b300:47
sbezverksdake: it is this one f31930ec92c1927db2de260fadd7083eb1c4b3d500:47
sbezverksdake: do you know how to revert it for testing?00:48
*** ssurana has joined #openstack-kolla00:51
*** ssurana1 has joined #openstack-kolla00:54
*** eaguilar has joined #openstack-kolla00:56
*** ssurana has quit IRC00:58
*** mnaser has quit IRC00:58
*** mnaser has joined #openstack-kolla01:00
*** mnaser has joined #openstack-kolla01:00
*** FiveBroDeepBook has joined #openstack-kolla01:13
*** FiveBroDeepBook has left #openstack-kolla01:13
*** salv-orlando has joined #openstack-kolla01:22
*** salv-orlando has quit IRC01:26
*** tonanhngo has joined #openstack-kolla02:01
*** tonanhngo has quit IRC02:02
*** ssurana1 has quit IRC02:08
*** goldyfruit has quit IRC02:09
sdake_sbezverk i disagree - that is a merge commit02:18
sdake_sbezverk merge commits are "empty" - they contain no code02:18
sdake_sbezverk what are the symtomsp of the failure02:18
sdake_git revert hash02:18
sdake_sbezverk you can use git show hash to see that it is empty02:20
*** ssurana has joined #openstack-kolla02:38
*** MasterOfBugs has joined #openstack-kolla03:05
*** sdake has joined #openstack-kolla03:09
*** sdake_ has quit IRC03:09
*** sp__ has joined #openstack-kolla03:14
*** rhallisey has quit IRC03:15
*** MasterOfBugs has quit IRC03:16
*** salv-orlando has joined #openstack-kolla03:23
*** eaguilar has quit IRC03:23
*** salv-orlando has quit IRC03:27
*** fragatina has joined #openstack-kolla03:29
*** fragatina has quit IRC03:34
*** Pavo has joined #openstack-kolla03:34
sbezverksdake_: I take this commit and 100% see exactly the same failure03:35
sbezverkI take one commit before it, everything runs fine03:35
sdakewhat failure do you see - the resolve dns timeout?03:35
sbezverkno03:36
sdakelink me an example gate run failure pls03:36
sbezverkI do not log handy but it is nova that gets broken03:36
sbezverkbasically you cannot launch VM03:36
sdakenova in ERROR state no hosts found?03:36
sdakeacross how many of the gate jobs?03:37
sbezverkfound the log http://logs.openstack.org/08/432408/12/check/gate-kolla-kubernetes-deploy-centos-binary-3-ceph-multi-nv/3d02fbb/console.html#_2017-02-11_15_22_48_29459303:37
sbezverkit is 100% consistant with that commit, one step back and the same job runs clean03:37
sdakeso patch set 11 is all green?03:38
sdakesbezverk the gating jobs seem to fail consistently in networking here: https://review.openstack.org/#/c/432697/603:39
sbezverkps 16 is all green you can ignore one failure it is not realted03:39
sbezverkhere is hash for commit http://logs.openstack.org/08/432408/12/check/gate-kolla-kubernetes-deploy-centos-binary-3-ceph-multi-nv/3d02fbb/console.html#_2017-02-11_15_22_48_29459303:39
sdakeok moment let me have a look03:39
sbezverke121ecb380d7f7e962fbd1afe5c402ef6b27717103:40
sdakethat unsupported operand type is what your worried about then?03:40
sdakecommit e121ecb380d7f7e962fbd1afe5c402ef6b27717103:40
sdakeAuthor: Sayantani Goswami <sayantani.goswami@intel.com>03:40
sdakeDate:   Thu Feb 2 19:29:57 2017 +000003:40
sdake    Update links in documents03:40
sdake    Update documents reference image-building documents correctly after03:40
sdake    image-building.rst was deleted from kolla-ansible directory.03:40
sdake    Change-Id: Ib66475e1f0f3055f2e38fc33ae649cae300c9c3603:40
sdakeso this works03:40
sbezverkyes03:41
sbezverkbut it fails on the next one03:41
sdakeand the commit after returns that operand error?03:41
sbezverkyes03:41
sdakelet me run that with an actual git rebase operation - i can't get teh gate to pass particularly well as is03:41
sdakeis it only 3.0.2 images?03:41
sbezverkyes03:42
sbezverkI think placement is ignored for 2.0.2 images03:42
sbezverkso no impact, but 3.0.2 probably starts using it or be aware of it03:42
sbezverknot sure03:42
sbezverkbut it is 100% clear03:42
sdakeso this commit then comes out clean in the gate? commit f31930ec92c1927db2de260fadd7083eb1c4b3d503:43
sbezverkthis is bad commit03:43
*** Pavo has quit IRC03:43
sbezverkthe one that breaks03:43
sdaketry running what you think works with a git revert facfabf3bbf57d5302f1b53f29396c532ea87352 afterwards03:43
sbezverklast good is e121ecb380d7f7e962fbd1afe5c402ef6b27717103:44
sdakeyou may have ot do git revert -m nomsg facfabf3bbf57d5302f1b53f29396c532ea8735203:44
sbezverksdake: I tried to do it manually and it gives me error03:44
sdakereverting that commit?03:44
sbezverkyeah03:44
sdakefrom a master checkout?03:44
sbezverkyes03:44
sdakelet me try -moemnt03:45
sbezverk git revert -m nomsg facfabf3bbf57d5302f1b53f29396c532ea8735203:45
sbezverkerror: switch `m' expects a numerical value03:45
sdakeya - running git revert wont work - there is a collsion in the commit which requires manual intervention03:46
sdakesbezverk run these operations locally:03:48
sdakegit reset --hard03:48
sdakegit show facfabf3bbf57d5302f1b53f29396c532ea8735203:48
sdakeyo uwill see there is no kubernetes conditional around the placement api (to set the proper endpoints)03:49
financeMansdake "APIError: 500 Server Error: Internal Server Error (\"{\"message\":\"maximum retry count cannot be used with restart policy \\'unless-stopped\"03:50
financeMantrying to deploy03:50
sdakefinanceMan busy03:50
sdakefinanceMan would love to help - limited time frame atm03:50
sdakefinanceMan try ask.openstack.org or try at another time :)03:51
sdakesbezverk you tell me what the nova.conf.j2 should look like03:51
sbezverksdake: I run these commands on master right?03:52
sdakesbezverk right03:52
sdakesbezverk I think a simple way to validate your theory is to just comment out the placement API in nova-api.conf.j203:53
sdakesbezverk the way to do that is to do that in your local copy of the repo, then git format-patch -1, then git am patchname in the gate script03:53
sbezverksdake: man that is big patch03:54
sdakefocus on nova.conf.j203:54
sdakewhat is happening is [placement] is being setup in nova.conf03:54
sbezverkI am not sure if just commenting out placement thing will work03:55
sbezverkthere are some wsgi changes as well03:55
sdakethat is for the placement api server03:55
sdakethat is independent of the nova api server03:55
sdakewe can't implement the placement api in 3.0.2 - there are no container images for it03:56
sdakesbezverk comment on lines 218-ish (where [placement] starts in a local copy of kolla-ansible03:57
sdakegit commit it03:57
sdakegit format-patch -103:57
sdakecopy that patch into a new gate script03:57
financeManHow do i shutdown a specific container? I think my deploy is failing because the heka container is still running after i tried to stop everything03:57
sdakerather a new gate test03:57
sdakefinanceMan docker stop [first 2-3 characters of container name from docker ps]03:58
sdakerather container id03:58
sdakefinanceMan still -busy :)03:58
sdakefinanceMan are you using master or a stable branch?03:58
Jeffrey4l_sbezverk, pong03:59
sdakeJeffrey4l_ i've got it03:59
sbezverksdake: I need it at the gate, I can try to kill it in the init container03:59
Jeffrey4l_what's happened?03:59
sdakesbezverk would you like me to do it :)03:59
sbezverkJeffrey4l_: Hey03:59
financeMansdake I ran "kolla-ansible pull" which I image is master and not a stable branch?03:59
sbezverksdake: sure04:00
sdakesbezverk cool give me 3-5 mins moment04:00
Jeffrey4l_sbezverk, what's wrong?04:00
sbezverkJeffrey4l_: based on git bisect that placement PS breaks our gate04:00
sbezverkf31930ec92c1927db2de260fadd7083eb1c4b3d504:01
sbezverkeven that fix I pushed recently does not help04:01
Jeffrey4l_sbezverk, hrm placement is required in Ocata. so i thin k8s need support it.04:01
sdakeJeffrey4l_ these are 3.0.2 images being gated04:01
sbezverksdake and I are trying to figure out how to revert that PS at least temporarely04:01
sdakeJeffrey4l_ so that wont work - as thre are no placement images in mitaka04:01
sdakesbezverk we dont need to revert it04:02
sdakewe need to fix it04:02
Jeffrey4l_i am confused.04:02
sdakeJeffrey4l_ wait for patchset - yo uwill understand04:02
Jeffrey4l_3.0.2 do not require  placement.04:02
sdakei can't fix it if i'm typing in irc :)04:02
Jeffrey4l_ok.04:02
sbezverkJeffrey4l: check this message04:02
sbezverkhttp://logs.openstack.org/08/432408/12/check/gate-kolla-kubernetes-deploy-centos-binary-3-ceph-multi-nv/3d02fbb/console.html#_2017-02-11_15_22_48_29459304:02
Jeffrey4l_wait for it.04:02
sbezverkJeffrey4l_: we get this failure constantly with only 3.0.2 images04:03
Jeffrey4l_are u using 4.0.0 and 3.0.2 in mix?04:03
sbezverkJeffrey4l_: 3.0.2 images, kolla-ansible master04:03
Jeffrey4l_sbezverk, that is bad.04:04
financeMansounds like someone is going to get fired, i better go to bed!04:05
Jeffrey4l_why not use 3.0.2 kolla ansible?04:05
sdakeok dudes dont panic04:05
Jeffrey4l_http://logs.openstack.org/08/432408/12/check/gate-kolla-kubernetes-deploy-centos-binary-3-ceph-multi-nv/3d02fbb/logs/openstack/kolla-nova-compute-x6grh.txt  check this.04:05
Jeffrey4l_seems nova compute try to connect placement api. but failed.04:05
sdakefinanceMan community project - nobody gets fired here :)04:05
Jeffrey4l_this is the root cause.04:05
Jeffrey4l_before placement api,  compute node resource is tracked by nova database.  after placement api, the reources are tracked by placement api.04:06
*** huzhengchuan has quit IRC04:10
sbezverkJeffrey4l_: we need to disable this behavior for 3.0.2 images04:10
sdakerelax04:10
Jeffrey4l_kolla-ansible 4.0.0 do not promise to deploy 3.0.2 images.04:11
sdakeall your questions will be answered04:11
sbezverkJeffrey4l_: ok we can pin 3.0.2 ansible04:11
sdakeno we can't04:11
Jeffrey4l_why not try to used kolla?04:11
sbezverksdake: why?04:11
openstackgerritSteven Dake proposed openstack/kolla-kubernetes master: Comment out placement API in all gate jobs  https://review.openstack.org/43272104:12
sdakethis is why ^^04:12
sdakeread the commit log and the patch pls04:12
sdakeand then ask qs :)04:12
*** crushil has joined #openstack-kolla04:13
Jeffrey4l_why not try to use 4.0.0 images?04:14
openstackgerritSteven Dake proposed openstack/kolla-kubernetes master: Comment out placement API in all gate jobs  https://review.openstack.org/43272104:14
Jeffrey4l_this patch works in gate.04:14
sdakeJeffrey4l_ - check out this list of gates:04:14
sdakehttps://review.openstack.org/#/c/432697/604:14
Jeffrey4l_but when some one want to try kolla-k8s, he will get the same error. ( because he do not add such patch)04:14
sdakegate-kolla-kubernetes-deploy-centos-binary-2-ceph-multi-nv04:15
sdakeJeffrey4l_ - right it needs to go into master of kolla-ansible04:15
sdakewith an if orchestration-engine == kubernetess then ignore04:15
sdakegate-kolla-kubernetes-deploy-centos-binary-2-ceph-multi-nv04:15
sdakesee this gate job - it has 2 in the name04:15
Jeffrey4l_this is not a good thing ;(04:15
sdakethe gate job with 2 in the name tests 2.0.2 images against master of kolla-kubernetes04:16
sdakethe gate job with 3 in the name tests 30.0.2 images against master of kolla-kubernetes04:16
sdakethere is also a gate job with -t- which is experimental which doesn't work which tests master (4.0.0 images)04:16
*** sudipto has joined #openstack-kolla04:17
*** sudipto_ has joined #openstack-kolla04:17
sdakeJeffrey4l_ i think tis a break from our tradition of having one version kolla-ansible deploy one version of kolla containers04:17
*** huzhengchuan has joined #openstack-kolla04:17
sdakeJeffrey4l_ however now that they are separate deliverables, it is a thing that can be done04:17
sbezverkJeffrey4l_: we need to supprot multiple versions because we want to test upgrade at the gate04:17
Jeffrey4l_this is different to support different deliverables.04:17
sdakeJeffrey4l_ i didn't parse that last statement - mind restating04:18
Jeffrey4l_it is supporting different version images.04:18
sdakeyup04:18
Jeffrey4l_we should avoid this.04:18
sdakethose are the gates that have been developed04:18
Jeffrey4l_it make things works.04:18
Jeffrey4l_worth.04:18
sdakethe reality is we are not destroying all those gate jobs in kolla-kubernetes because we don't want a conditional in kolla-ansible04:19
Jeffrey4l_i am OK to fix this in kolla-ansible with if  orchestration-engine == kubernetess.04:19
sdakesounds good :)04:19
Jeffrey4l_but we need avoid this.04:19
sdakei think we need to discuss further04:19
sdakei have worked with kfox1111 and sbezverk on this exact point and it seems a point of contention04:19
sdakei'm not really sure what to do about it other then roll with it04:20
sdakeJeffrey4l_ by pike genconfig should be gone from kolla-ansible entirely04:20
sdakeJeffrey4l_ and it will no longer be a kolla-ansible deliverable problem04:20
Jeffrey4l_kolla deliverables should not support different version images in one branch.04:20
sdakeJeffrey4l_ i'd challenge you on that - why not?04:21
Jeffrey4l_no one and job promise that.04:21
sdakedidn't parse - pls restate?04:21
Jeffrey4l_why we need support?04:21
sbezverkJeffrey4l_: upgrades04:21
sdakeJeffrey4l_ i dont totally understand the why it is needed04:21
Jeffrey4l_i.e. kolla-ansible ocata should not support kolla newton images.04:22
Jeffrey4l_ kolla-k8s ocata should not support kolla newton images.04:22
Jeffrey4l_if so, the end-user will want to try kolla-ansible deploy kolla newton images.04:22
sdakeJeffrey4l_ your making a statement but not explaining why04:22
Jeffrey4l_it will be end-less support.04:23
sdakei see - i feel the same way04:23
Jeffrey4l_and openstack do not promise this too.04:23
sbezverkJeffrey4l_: we do not have such rigid rules.. kolla-kub code can run with either versions04:23
Jeffrey4l_openstack nova ocata work with neuton mitaka? who knows?04:23
sdakehowever i didn't write the gates - other people did04:23
sdakeJeffrey4l_ that does not work at all04:24
sdakeJeffrey4l_ the idea is choose newton, or choose ocata, or choose pike04:24
Jeffrey4l_OK.04:24
Jeffrey4l_it is still wired.04:24
sdakei agree seems semi-odd04:24
Jeffrey4l_if so, why should we have kolla-ansible ocata branch?04:25
sbezverksdake: I left you a comment04:25
Jeffrey4l_rather then only have one master is enough.04:25
sdakeJeffrey4l_ really good question04:25
Jeffrey4l_if so, we will have end-less if-else in kolla-ansible's code.04:25
Jeffrey4l_which to handle if ocata ehn04:25
Jeffrey4l_if newton then04:25
sbezverkJeffrey4l_: Because you cannot force people deploying always use master04:25
sdakeJeffrey4l_ right - that was the decision that was made 18 months ago04:25
sbezverkoperators want to deploy stable releases04:26
sdakeJeffrey4l_ i get the endless if/else/else/else thing04:26
sdakeJeffrey4l_ we have a ptg design session just about this topic (configuration)04:26
Jeffrey4l_sbezverk, what i wanna to say is: if kolla-ansible master branch support newton, what's the different with kolla-ansible newton.04:27
sdakeJeffrey4l_ remote dial-in available04:27
Jeffrey4l_branch*04:27
Jeffrey4l_sdake, ok.04:27
sdakeJeffrey4l_ we are using the kolla-ansible deliverable for ocata and part of pike to bootstrap the config gen for kolla-kubernetes04:27
sdakei expect by the end of pike, genconfig will be reomved from kolla-ansible entirely04:28
sbezverksdake: I hope earlier04:28
sdakeand instead config gen will be handled directly in kolla-kubernetes04:28
Jeffrey4l_sdake, i know the genconfig part. i hope that too. ;)04:28
sdakeits a transitory problem which makes the jinja2 a little untidy04:28
sdakesadly ocata was a 4 month cycle and the repo split madness - well you lived that04:29
sdakevery disruptive04:29
sbezverksdake: are you going to fix your ps?04:30
sdakesbezverk looking moment04:30
sdakesbezverk good point - how do I do that precisely?04:31
sdakeso we can verify a conditona lwill actually work since we don't have cross-repo gating enabled04:32
sbezverkpushd kolla-ansible;04:32
sbezverkrun yout command04:32
sbezverkpopd04:32
openstackgerritSteven Dake proposed openstack/kolla-kubernetes master: Comment out placement API in all gate jobs  https://review.openstack.org/43272104:33
sdakesbezverk can you file a blueprint to implement the placement api please04:34
sdakewe will need it for the 4.0.0 images once they are on hub04:34
sdakeotherwise nova no worky for pike at all04:35
sbezverksdake: in the patch there are a couple of whitespaces, I hope it will not break the patch application..04:35
sdakeaccording to how Jeffrey4l_ explained it04:35
sdakesbezverk if it does i can edit it04:35
* sdake dispises whitespace behavior of git04:35
sbezverksdake: never heard about placement before :) until we hit this issue04:35
sdakesbezverk me either04:36
sdakeJeffrey4l_ do you knwo why teh merge commit is empty in git - yet includes the actual patch?04:36
sdakegit show on the merge commit (where sbezverk bisected to) turns up nothing04:36
sdakegit show of the actual placement api patch shows the real problem04:36
*** crushil has quit IRC04:39
sbezverksdake: yeah it looks like it is show stopper for Ocata04:39
sdakefortunately easy to fix04:39
Jeffrey4l_sdake, no idea.04:40
*** newmember has joined #openstack-kolla04:44
sbezverksdake: try to add $WORKSPACE before tools/04:45
openstackgerritSteven Dake proposed openstack/kolla-kubernetes master: Comment out placement API in all gate jobs  https://review.openstack.org/43272104:45
openstackgerritSteven Dake proposed openstack/kolla-kubernetes master: Comment out placement API in all gate jobs  https://review.openstack.org/43272104:46
sbezverksdake: with pushd/popd pwd does not show correct location in the file system if I rememeber right04:47
sdakei see - so why am I pushding kolla-ansible again?04:47
sdakeshouldn't I pushd kolla-kubernetes?04:47
sbezverkso even if it show wrong location, it might still work ;)04:47
sdakethe patch is in kolla-kubernetes/tools dir04:48
openstackgerritSteven Dake proposed openstack/kolla-kubernetes master: Comment out placement API in all gate jobs  https://review.openstack.org/43272104:49
sdakewhere is kolla-kubernetes repo in teh gate, $WORKSPACE?04:49
sdakewho is on second :)04:50
sbezverksdake: it is not there yet04:50
sdakesbezverk please file a bug for this problem against kolla-ansible deliverable in launchpad and link in channel04:50
sbezverkthat is why you do not see it, I think it gets cloned later04:50
*** tonanhngo has joined #openstack-kolla04:50
*** crushil has joined #openstack-kolla04:53
*** tonanhngo has quit IRC04:53
sdakezuul jobs on ipv6 = annoying04:53
sdakesince my network at home doesn't roll with ipv604:53
sbezverksdake: https://bugs.launchpad.net/kolla-ansible/+bug/166396104:54
openstackLaunchpad bug 1663961 in kolla-ansible "nova compute fails when placement is configured" [Undecided,New]04:54
sdakebbin10 mins04:57
sdakewhile the gate does it thing04:57
sbezverksdake: btw, compute kit ps is all ready and green05:13
sbezverkhttps://review.openstack.org/#/c/431156/05:14
*** caowei has joined #openstack-kolla05:16
sdakesbezverk cool05:21
sdakeare you done with the gate refactor?05:21
sdakei have a seminar on wednesday at 2pm MST05:21
sbezverkthat is the last ps for refactor05:21
sdakeand I need a working environment that I can demonstrate or record a video of05:21
sdakefor both kolla-ansible and kolla-kubernetes05:21
*** sacharya has joined #openstack-kolla05:22
sdakesbezverk can we put a pause on gate refacotrs for a week so the dev env can merge05:23
sdakesbezverk making the gate look perfect is secondary to having a working development environment05:23
sdakebmace 's patch is in merge conflict and in need of a rebase05:24
*** salv-orlando has joined #openstack-kolla05:24
sbezverkthis last one does not change previous 2 PS05:24
sbezverkI just add compute kit and experiemntal job to tests it05:25
sbezverkI do not think it is right to wait for wip when there is one ready to get merged05:25
sbezverksdake: so you fix helped05:26
sbezverks/you/your/05:26
sdakewait for wip regardig what?05:27
sbezverkbmace's PS is still in wip05:27
sdakeya - he is waiting for me to test it05:27
sdakeand i can't do that because it is in merge conflict05:27
sdakeits not wip05:28
sdakeits ready to merge if it weren't in merge conflict05:28
*** salv-orlando has quit IRC05:28
sbezverkmore reason to finish completely with the gate and then move on05:28
sdakeif by finish with the gate - you mean finish today - then yes05:29
sdakeif you mean finish next week then no05:29
sdakethe dev env is criitical - it has been blocking a whole slew of people for months!05:29
sdakei don't know how many times i have to repeat myself on this point...05:29
sbezverksdake: it can be finished today if you find second code to ack it ;)05:31
sdakezuul says that patch fixed it05:31
sdakesbezverk link plz - i'll review05:31
sbezverksdake:  https://review.openstack.org/#/c/431156/05:31
sdakewe need to make the gate a little more reliable05:32
sdakeall of the rechecks are disturbing :(05:32
*** sudipto has quit IRC05:34
*** sudipto_ has quit IRC05:34
*** sudipto has joined #openstack-kolla05:34
*** sudipto_ has joined #openstack-kolla05:34
sbezverksdake: I am looking into fixing one frequent failure cause, but still in research mode05:34
*** zhangqiankun has quit IRC05:36
*** zhangqiankun has joined #openstack-kolla05:37
sdakesbezverk https://review.openstack.org/#/c/431156/24 m0ar whitespace problems and a question results in -105:37
sdakesbezverk turn it around and i'll +2 it tonight05:38
sdakewhat frequent cause out of curiosity?05:38
sdakesbezverk there is an etherpad where the gate work is tracked correct?  Mind linking?05:40
sdakeJeffrey4l_ this seems to work for kolla-kubernetes - mind having a look: https://review.openstack.org/#/c/432721/605:40
sdakeJeffrey4l_ remember -3 images are newton05:41
openstackgerritSerguei Bezverkhi proposed openstack/kolla-kubernetes master: Adding compute kit based on iscsi backend  https://review.openstack.org/43115605:41
sdakeJeffrey4l_ and sbezverk is going to file a blueprint for api-placement in kolla-kubernetes and link it :)05:41
sbezverksdake: blue prit is already done05:42
sdakesbezverk cool can you link so i can target it05:42
sdakesbezverk btw this looks good: https://review.openstack.org/#/c/431156/24..2505:42
sbezverksdake: https://blueprints.launchpad.net/kolla-kubernetes/+spec/kolla-kubernetes-placement05:43
sbezverksdake: our gate fails often on generating resolv.conf workaround05:43
sbezverkI am looking into a way to make it less painful05:44
sdakesbezverk is this necessary? https://blueprints.launchpad.net/kolla-kubernetes/+spec/health-checks05:47
sbezverksdake: https://etherpad.openstack.org/p/kolla-kubernetes-gate-failure-tracking05:47
*** sp__ has quit IRC05:47
sbezverksdake: no it is not needed05:48
sbezverkok I am off for now, good night..05:51
sdakehttps://blueprints.launchpad.net/kolla-kubernetes/+spec/gate-continous-improvement05:51
sdakethanks sbezverk05:51
sdakehave a nice evening05:51
*** unicell has joined #openstack-kolla05:57
*** FiveBroDeepBook has joined #openstack-kolla06:00
*** FiveBroDeepBook has left #openstack-kolla06:03
*** unicell has quit IRC06:07
*** unicell has joined #openstack-kolla06:07
bmacesdake: my change can't merge until kubernetes python lib is supported in global requirements06:17
sdakebmace ok - can you rebase anyway - so I can actually test it out - I have sunday allocated to producing a working demo of it for a seminar on wednesday06:18
bmacelooks like the only issue is with the keystone rc change06:19
bmacesdake: also.. looking at review.openstack.org it doesn't show as in conflict, until the iscsi backend changes from sbezverk get merged in06:20
sdakebmace wierd gerrit showed it as -1 - let me check it moment06:20
bmacesdake: -1 from me and -1 from jenkins due to requirements failure06:21
sdakeoh requirements06:21
sdakeright06:21
bmacesdake: so, have at it and let me know how it goes :)06:21
sdakei see it is failing the gate06:21
sdakei will have at it -although its midnight here and i am out of steam06:21
bmacesdake: from what i have seen some of those fail semi-randomly.. if we really have an issue with pv / pvc allocation i'm not surprised.. which seems why the devenv doesn't come up clean sometimes also06:22
sdakeok - recheck gets the semi-random failures to a green state eventually06:22
bmacesdake: i hate inconsistent failures... the worst kind :/06:23
sdakeagreed06:23
sdakewe have a blueprint to fix that06:23
sdakebmace https://blueprints.launchpad.net/kolla-kubernetes/+spec/gate-continous-improvement06:24
*** huzhengchuan has quit IRC06:30
*** huzhengchuan has joined #openstack-kolla06:30
openstackgerritSteven Dake proposed openstack/kolla-ansible master: Fixes problem with kolla-kubernetes 3.0.2 images  https://review.openstack.org/43272506:32
openstackgerritSteven Dake proposed openstack/kolla-kubernetes master: Test a critical kolla-ansible defect resolution  https://review.openstack.org/43272106:39
*** crushil has quit IRC06:50
*** tonanhngo has joined #openstack-kolla07:03
*** sacharya has quit IRC07:08
*** sacharya has joined #openstack-kolla07:09
*** ssurana has quit IRC07:14
*** sacharya has quit IRC07:14
*** salv-orlando has joined #openstack-kolla07:25
*** salv-orlando has quit IRC07:29
*** crushil has joined #openstack-kolla07:37
*** tonanhngo has quit IRC07:39
*** unicell has quit IRC07:43
*** unicell has joined #openstack-kolla07:43
*** ipsecguy has quit IRC07:49
*** unicell1 has joined #openstack-kolla07:50
*** sdake has quit IRC07:56
*** unicell1 has quit IRC08:01
*** tonanhngo has joined #openstack-kolla08:01
*** tonanhngo has quit IRC08:04
*** crushil has quit IRC08:23
openstackgerritJeffrey Zhang proposed openstack/kolla-ansible master: Fix horizon pull error  https://review.openstack.org/43273808:46
openstackgerritJeffrey Zhang proposed openstack/kolla-ansible master: Fix horizon pull error  https://review.openstack.org/43273809:16
openstackgerritJeffrey Zhang proposed openstack/kolla-ansible master: Install ansible without upper constraints  https://review.openstack.org/43273909:16
*** kong has joined #openstack-kolla09:17
*** salv-orlando has joined #openstack-kolla09:26
*** salv-orlando has quit IRC09:30
openstackgerritJeffrey Zhang proposed openstack/kolla-ansible master: Install ansible without upper constraints  https://review.openstack.org/43273909:34
openstackgerritJeffrey Zhang proposed openstack/kolla-ansible master: Fix horizon pull error  https://review.openstack.org/43273809:34
*** hrw has quit IRC09:41
*** hrw has joined #openstack-kolla09:43
*** mannamne has joined #openstack-kolla10:19
*** ipsecguy has joined #openstack-kolla10:23
openstackgerritJeffrey Zhang proposed openstack/kolla-ansible master: Install ansible without upper constraints  https://review.openstack.org/43273910:33
openstackgerritJeffrey Zhang proposed openstack/kolla-ansible master: Fix horizon pull error  https://review.openstack.org/43273810:33
*** gema has quit IRC11:03
*** gema has joined #openstack-kolla11:03
*** huzhengchuan has quit IRC11:07
*** kelv has quit IRC11:36
*** sp__ has joined #openstack-kolla11:48
*** sp__ has quit IRC11:58
*** tonanhngo has joined #openstack-kolla12:03
*** tonanhngo has quit IRC12:05
*** pbourke has quit IRC12:08
*** pbourke has joined #openstack-kolla12:08
*** FiveBroDeepBook has joined #openstack-kolla12:09
*** FiveBroDeepBook has left #openstack-kolla12:09
*** caowei has quit IRC12:27
*** kong has quit IRC12:29
*** caowei has joined #openstack-kolla12:33
*** sp__ has joined #openstack-kolla12:50
*** salv-orlando has joined #openstack-kolla13:06
*** sudipto_ has quit IRC13:09
*** sudipto has quit IRC13:09
financeManHow do I deploy kolla from a stable branch? I am not seeing it in the documentation -> http://docs.openstack.org/developer/kolla/newton/quickstart.html13:09
*** salv-orlando has quit IRC13:10
*** salv-orlando has joined #openstack-kolla13:16
*** zhubingbing has joined #openstack-kolla13:17
*** salv-orlando has quit IRC13:22
*** haplo37 has quit IRC13:23
*** haplo37 has joined #openstack-kolla13:31
*** financeMan has quit IRC13:49
*** sdake has joined #openstack-kolla13:50
*** Pavo has joined #openstack-kolla13:54
*** ipsecguy has quit IRC13:54
*** ipsecguy has joined #openstack-kolla13:54
sdakeJeffrey4l_ morning -  around?13:55
sdakeany cores around - seeing this problem in the gate: http://logs.openstack.org/25/432725/1/check/gate-kolla-ansible-dsvm-deploy-centos-binary-centos-7-nv/09f74b7/console.html13:56
sdakejinja2.8 should be in the pip archive - is this an infra problem or  a gate regressoin?13:57
*** breitz has quit IRC13:59
Jeffrey4l_sdake, yep.14:08
Jeffrey4l_i have a patch fixing this.14:08
sdakelink pls so I can ack14:09
sdakeblocking my demo wednesday - and running out of time14:09
sdakei know that isn't your problem or a community problem14:09
sdakehowever i am representing the community here a tthis meetup, and and i can't get a  demo of kolla-ansible and kolla-kubernetes rolling in time, our community will look kind of flakey :(14:09
sdakei had believed the code base was in good shape and ready to roll for a demo14:10
sdakeappears we aren't quite there yet- so will have to cobble together something usable14:10
openstackgerritJeffrey Zhang proposed openstack/kolla-ansible master: Install ansible without upper constraints  https://review.openstack.org/43273914:10
sdakeI may do a demo of stable branches of kolla-ansible14:10
*** eaguilar has joined #openstack-kolla14:10
Jeffrey4l_sdake, https://review.openstack.org/43273914:10
sdakeJeffrey4l_ may I ask - is this because of a requirements.txt change ?14:11
openstackgerritJeffrey Zhang proposed openstack/kolla-ansible master: Fix horizon pull error  https://review.openstack.org/43273814:11
sdakeI saw on the [ptl] tag that requirements would be unfrozen14:11
sdakeso we will have to deal with that problem as a trailing release cycle project by -2'ing all requirements changes until the last minute and doing a manual sync14:11
sdakewhy are upper constraints for jinja2 -=2.9.514:12
sdake==2.9.5?14:12
sdakeconstraints project shouldn't hard pin omo14:12
Jeffrey4l_sdake, requirements creates ocata branch.14:12
Jeffrey4l_so constraints is changed.14:12
Jeffrey4l_i have say we should create ocata branch now.14:13
sdakeyup makes sense14:13
Jeffrey4l_have to say.14:13
sdakethat will sovle the problem with the requirements sync14:13
sdakehowever create a backport problem14:13
Jeffrey4l_yep.14:13
sdakealthough I feel those problems are solveable given the experiences we had with newton14:13
sdakeas long as new fetaures aren't going in14:13
Jeffrey4l_no perfect solution ;)14:13
sdakethis is why i sugggested no new features after 4.0.0.b3 when you asked14:14
sdakeare new features going into ocata?14:14
Jeffrey4l_almost all projects creates ocata branch.14:14
Jeffrey4l_i think it is really time to do so in kolla.14:14
Jeffrey4l_new.14:14
sdakeJeffrey4l_ ok - well you are the release laison - you have ultimate authority over when that is done14:14
Jeffrey4l_no.14:14
sdakeno you don't?14:15
sdakeor no new features14:15
Jeffrey4l_no new features.14:15
sdakecool - so just bug fixes14:15
Jeffrey4l_yep.14:15
sdakewe should be backporting those as best we can14:15
Jeffrey4l_sure.14:15
sdakeso to branch you need a tag first14:15
sdaketo branch from the tag14:15
sdakei think is how it works ;-)14:15
sdakewhich means rc1 should be tagged with current state of a  (working) master14:16
Jeffrey4l_yep. i will create rc1 tag at the same time create ocata branch.14:16
sdakenice14:16
sdakeJeffrey4l_ all this stuff is harder then it looks huh :)14:16
sdakemanaging the release process, managing backports, managing features, etc14:16
Jeffrey4l_sdake, after this https://review.openstack.org/432738 is merged. i will start the branching.14:17
Jeffrey4l_yep. it is.14:17
sdakeif we can get my patch in for genconfig i'd appreciate that too14:17
sdakeif not, i can backport it14:17
Jeffrey4l_no problem.14:17
Jeffrey4l_is there a patch?14:17
sdakehttps://review.openstack.org/#/c/432738/5/ansible/roles/horizon/tasks/pull.yml14:17
sdakewith dict does what?14:17
Jeffrey4l_got it.14:18
Jeffrey4l_with_items works with list, with_dict works with dict.14:18
Jeffrey4l_with_dict just like items()14:18
sdakehttps://review.openstack.org/#/c/432725/14:18
sdakewhere is the dictionariy in that patch?14:19
sdake(just tryign to learn not suggesting yoru patch is wrong)14:19
Jeffrey4l_would u mind to add some NOTE to your patch? which is helpful to be removed in the feature. because this is a dirty workaround atm.14:19
sdakewhere is horizon_services defined?14:19
sdakeJeffrey4l_ leave review comments and I'll fix as necessary14:19
Jeffrey4l_cool14:20
sdakeJeffrey4l_ i a mwaiting on a functional gate, so we need your patch merged14:20
Jeffrey4l_sdake, re horizon_services  not suggesting yoru patch is wrong)14:20
Jeffrey4l_https://github.com/openstack/kolla-ansible/blob/master/ansible/roles/horizon/defaults/main.yml#L414:20
Jeffrey4l_roger.14:20
sdakecool thanks for teaching me that Jeffrey4l_ :)14:20
Jeffrey4l_np ;)14:20
openstackgerritJeffrey Zhang proposed openstack/kolla master: [TEST][MASTER][KOLLA]  https://review.openstack.org/42662514:21
sdakeJeffrey4l_ if you can lave coments in https://review.openstack.org/#/c/432725/1/ansible/roles/nova/templates/nova.conf.j2 now - i'll address them as I have a slew of prep work to do for this seminar14:22
sdakeJeffrey4l_ please read my latest comment on https://review.openstack.org/#/c/432721/714:24
sdakeoh nm, that was sbezverk that made that statement14:25
*** sdake has quit IRC14:26
*** Pavo has quit IRC14:27
Jeffrey4l_sdake, i left just now.14:28
*** ipsecguy_ has joined #openstack-kolla14:31
*** ipsecguy has quit IRC14:31
*** sdake has joined #openstack-kolla14:32
Jeffrey4l_sdake, could u ack this one ? https://review.openstack.org/42935814:33
sdakeJeffrey4l_ looking14:34
Jeffrey4l_and this one if possible https://review.openstack.org/424504 ;)14:34
sdakeJeffrey4l_ give me a moment - rebasing my patches so we can get rc1 rolling14:34
*** mannamne has quit IRC14:35
*** ipsecguy_ has quit IRC14:36
*** ipsecguy has joined #openstack-kolla14:36
Jeffrey4l_roger. work for me.14:39
openstackgerritSerguei Bezverkhi proposed openstack/kolla-kubernetes master: Stabilizing the gate, fixing  resolv.conf issue  https://review.openstack.org/43276414:43
openstackgerritSteven Dake proposed openstack/kolla-ansible master: Fixes problem with kolla-kubernetes 3.0.2 images  https://review.openstack.org/43272514:44
openstackgerritSteven Dake proposed openstack/kolla-ansible master: Install ansible without upper constraints  https://review.openstack.org/43273914:44
openstackgerritSteven Dake proposed openstack/kolla-ansible master: Fix horizon pull error  https://review.openstack.org/43273814:44
openstackgerritSteven Dake proposed openstack/kolla-ansible master: Fixes problem with kolla-kubernetes 3.0.2 images  https://review.openstack.org/43272514:44
openstackgerritSteven Dake proposed openstack/kolla-ansible master: Install ansible without upper constraints  https://review.openstack.org/43273914:44
openstackgerritSteven Dake proposed openstack/kolla-ansible master: Fix horizon pull error  https://review.openstack.org/43273814:44
sdakeJeffrey4l_ I don't understand htis statement could you rephrase in irc: hrm.  we haven't mark this as deprecated and use config.json instead.14:44
sdakeJeffrey4l_ zomg that last one will take a little bit to process :)14:45
*** ipsecguy has quit IRC14:46
*** ipsecguy has joined #openstack-kolla14:47
*** mannamne has joined #openstack-kolla14:48
Jeffrey4l_thanks.14:50
sdakeJeffrey4l_ I see a slew of conflicting patches for that optimize patch14:50
sdakeJeffrey4l_ that implies that making reconfigure work well is sitll an outstanding feature14:51
sdakeJeffrey4l_ I think neutron is clearly a priority14:51
sdakeJeffrey4l_ looking at that patch - as I haven't been following the reconfigure work, I don't feel very comfortable adding a +2/+W to it14:52
sdakeJeffrey4l_ although the patch looks good to me, and happy to +1 it14:52
sdake(I don't see any glaring errors)14:52
sdakeJeffrey4l_ given that the kolla-ansible gate doesn't test reconfigure (yet), I'd have to either a) have context for the patch or b) test it personally14:53
sdakeI lack A, and don't have time for b given the time pressure of the seminar in 3 days14:53
sdakeJeffrey4l_ any chance you can hunt down a different reviewer for it?14:53
sbezverkgood morning, what is new and exciting ;)?14:55
Jeffrey4l_sdake, yep. we still use extend_start.sh14:56
sdakeJeffrey4l_ right i am aware of that :)14:56
sdakeJeffrey4l_ that last set of messages was for better-reconfigure blueprint14:57
sdakeJeffrey4l_ lets talk about this review: https://review.openstack.org/42935814:57
Jeffrey4l_i left some comments. but seem egonzalez missed it.14:58
sdakeJeffrey4l_ i don't understand the comment you left -  could you rephrase14:59
*** rwallner has joined #openstack-kolla14:59
sdakeJeffrey4l_ If I read the review right, an dfrom my understanding of config.json, config.json should be doing the chowns15:00
sdakeJeffrey4l_ maybe that has changed in the last month or so15:00
sdakeJeffrey4l_ did we deprecate config.json doing the chowns?15:00
*** rwallner has quit IRC15:00
*** rwallner has joined #openstack-kolla15:01
*** tonanhngo has joined #openstack-kolla15:01
*** david-lyle has joined #openstack-kolla15:02
*** tonanhngo has quit IRC15:02
*** huzhengchuan has joined #openstack-kolla15:05
*** david-lyle has quit IRC15:10
Jeffrey4l_sdake, gate tests reconfigre.15:10
Jeffrey4l_sdake, check https://github.com/openstack/kolla-ansible/blob/master/tools/deploy_aio.sh#L11215:10
Jeffrey4l_ok. i will.15:11
Jeffrey4l_thanks.15:11
openstackgerritSerguei Bezverkhi proposed openstack/kolla-kubernetes master: Stabilizing the gate, fixing  resolv.conf issue  https://review.openstack.org/43276415:15
sdakeJeffrey4l_ cool - re reconfigure I'd be happy to change my vote to +2 then15:15
Jeffrey4l_sdake, do u get what egonzalez said?15:18
Jeffrey4l_now, for log folder, we have two solutions. one is in extend_start.sh ( what that patch does). another is config.json, which is a workaround for uid and gid change.15:18
Jeffrey4l_right?15:18
Jeffrey4l_we never deprecated either one.15:18
Jeffrey4l_on the other hand i think the former one is better, since we have multi deliverables now.15:18
Jeffrey4l_sdake, we didn't deprecated either extend_start or config.json.15:18
*** l4yerffeJ has joined #openstack-kolla15:18
*** salv-orlando has joined #openstack-kolla15:18
*** salv-orlando has quit IRC15:23
*** huzhengchuan has quit IRC15:23
sdakeJeffrey4l_ got it15:24
sdakeJeffrey4l_ let me hunt down the review again15:24
Jeffrey4l_thanks.15:28
*** Jeffrey4l_ has quit IRC15:30
*** Jeffrey4l__ has joined #openstack-kolla15:30
Jeffrey4l__sdake, gate is green again. https://review.openstack.org/43273915:31
openstackgerritzhubingbing proposed openstack/kolla-ansible master: Fix keystone revocation error  https://review.openstack.org/43276815:31
*** breitz has joined #openstack-kolla15:31
*** Jeffrey4l__ has quit IRC15:32
*** Jeffrey4l__ has joined #openstack-kolla15:33
*** Jeffrey4l__ is now known as Jeffrey4l15:33
sdakeJeffrey4l check out https://review.openstack.org/#/c/429358/115:34
Jeffrey4lthanks.15:35
sdakezhubingbing where is the workflow dude? :)15:35
sdakezhubingbing  on https://review.openstack.org/43273915:36
zhubingbingyes15:36
zhubingbingwait gate15:36
*** leseb has quit IRC15:36
sdakegate is green isn't it?15:36
sdakeok - wfm :)15:36
Jeffrey4lhe -w before the gate.15:36
sdakeya i rebased those patches15:37
zhubingbing;)15:37
sdakeso I could stack mine on top15:37
sdakeand forgot that needs a new gat erun15:37
sdake8am here ;)15:37
zhubingbingdone15:37
zhubingbing;)15:37
sdakecan folks have a look at https://review.openstack.org/#/c/432725/315:38
sdakethen I htink we can branch15:38
sdakenote i tested this patch in the gate (see very first comment) of kolla-kubernetes and the gate is gloriously green15:39
openstackgerritSerguei Bezverkhi proposed openstack/kolla-kubernetes master: Stabilizing the gate, fixing  resolv.conf issue  https://review.openstack.org/43276415:39
Jeffrey4l+2ed15:39
openstackgerritJeffrey Zhang proposed openstack/kolla-ansible master: Optimize reconfigure action for neutron  https://review.openstack.org/42450415:39
sdakesbezverk shortly genconfg problem will be gtg - nice job sorting out the patch and the need for the placement api15:40
openstackgerritMerged openstack/kolla-ansible master: Install ansible without upper constraints  https://review.openstack.org/43273915:40
*** rwallner has quit IRC15:40
Jeffrey4lgare is green too. please ack https://review.openstack.org/43273815:41
sdakeJeffrey4l why did you rebase 42450415:41
Jeffrey4lsdake, run the gate again ;)15:42
sdakerecheck?15:42
sdakerecheck runs the gate again against the latest master15:42
Jeffrey4l recheck do not rebase to the latest code.15:42
sdakei have had this conversatio nabout 20 times with core reviewers and everyone argues with me15:42
sdakeyet i proved it to sbezverk and kfox a few weeks ago15:42
sdakeso they finally stopped arguing15:42
Jeffrey4lok.15:42
sdakeJeffrey4l fwiw i originally thought the same as you did15:43
sdakehowever, openstack-infra corrected me on the mailing list15:43
Jeffrey4lbut how the magic happened?15:43
sdakesam and I both verified  their statements15:43
Jeffrey4lon the other hand, recheck and rebase is almost the same ;)15:43
sdakerebase is slightly different wrt how stackalytics processses content15:44
sdakei'd far prefer rechecks to rebases, and i think all contributors would agree15:44
openstackgerritJeffrey Zhang proposed openstack/kolla-ansible master: [TEST] test recheck  https://review.openstack.org/43276915:50
sdakeJeffrey4l here is the thread: https://openstack.nimeyo.com/75611/openstack-dev-kolla-unblocking-the-gate15:51
openstackgerritzhubingbing proposed openstack/kolla-ansible master: Fix keystone revocation error  https://review.openstack.org/43276815:51
*** zigo has quit IRC15:51
Jeffrey4lroger. thanks.15:52
sdakeJeffrey4l recommend reading throug hthat to understand the how it works part :)15:52
sdakethere is a good description in the thread after I made the mistaken request15:52
*** zigo has joined #openstack-kolla15:55
openstackgerritJeffrey Zhang proposed openstack/kolla-ansible master: [TEST] test recheck  https://review.openstack.org/43276915:56
sdakezhubingbing can you ack https://review.openstack.org/#/c/432725/ pls16:00
zhubingbingok16:00
Jeffrey4lsdake, verified. recheck works as expect. thanks.16:01
sdakeJeffrey4l cool :)16:01
zhubingbingi need test it16:01
zhubingbingLGFM16:02
sdakeJeffrey4l your test approach is pretty cool :)16:02
sdakemuch smipler then mine was16:02
sdakegit log - didn't think of that16:02
Jeffrey4l:)16:03
Jeffrey4lzuul-cloner always rebase the patch on the master the project code.16:03
*** goldyfruit has joined #openstack-kolla16:03
sdakeright16:03
Jeffrey4lso the commit log/hash changed.16:03
zhubingbingsdake ;) next week i have began start kolla-k8s16:04
sdakezhubingbing hotness16:04
sdakezhubingbing how do you feel about implementing ironic :)16:04
sdakezhubingbing the  dev env is about ready to merge16:04
zhubingbingironic?  i like cobbler16:04
zhubingbingi dont‘t like it16:04
sdakezhubingbing i see - fair enough - how about magnum then?16:05
sdakezhubingbing if you can't do any implementation work - reviews are also helpful :)16:06
zhubingbingroger16:07
sdakeroger on magnum or roger on reviews16:07
zhubingbingfor reviews16:08
sdakethanks :)16:08
zhubingbingmagnum i think it16:08
zhubingbingit is too complicated for a container solution16:09
sdaketend to agree16:09
zhubingbingBut recently I did not continue study  this project16:09
openstackgerritMerged openstack/kolla master: Fix panko log folder owner  https://review.openstack.org/42935816:12
openstackgerritMarcin Juszkiewicz proposed openstack/kolla master: build: sort list of built/failed images before printing  https://review.openstack.org/43277416:16
sbezverksdake: please check this one, it will eliminate one common types of failure at the kube gate. https://review.openstack.org/43276416:17
hrwcan someone look at https://review.openstack.org/432774 and tell does it make sense for you?16:17
sdakehrw note i responded on the ml re your review16:17
sdakehrw did you happen to read it yet this fine sunday morning?16:18
hrwsdake: thx, not checked mails yet16:18
sdakeplease read that email16:18
sdakerathe rthat mailing list message16:18
hrwsdake: thanks16:18
hrwsdake: makes sense.16:19
sdakehrw cool - so we arn't ignoring you - we are overloaded :)16:20
hrwsdake: https://blueprints.launchpad.net/kolla/+spec/multiarch-and-arm64-containers is my BP - have to link it somehow to review/branch16:20
hrwalso have to update that bp a bit16:20
sdakehrw the place to udpat eis the work items field16:21
hrwyep, doing that now.16:21
hrwlong time since I played with LP16:21
sdakehrw ya launchpad is not ideal - its what we have :(16:22
hrw;)16:22
hrwsdake: Linaro moved to JIRA16:22
sdakehrw opentack is moving to a free tool called storyboard16:22
sdakejira costs way too much for the 3000 individuals contributing to a release16:22
sdake(in terms of commits/bug triage/etc)16:23
hrwsdake: I have no idea is it free or not - that's what I have to work with16:23
*** mattmceuen has quit IRC16:23
hrwsdake: So, while waiting for Ocata release, I will do some tweaks here and there and send them for review. Then will wait patiently for someone to review them.16:24
sdakehrw right - i'd expect more action around march 10th when we cut ocata iirc16:25
hrwfine for me16:25
hrwsdake: from 21th Feb to end of week I am on PTO, then week of work, then conference week16:26
sdakecool - i know it sucks being  right in the midst of a release rc1 and trying to get work done for pike16:26
*** adrian_otto has joined #openstack-kolla16:27
hrwsdake: BP updated16:27
hrwsdake: Kolla uses packages so for some time it will still build Ocata, right?16:28
*** FiveBroDeepBook has joined #openstack-kolla16:28
*** FiveBroDeepBook has left #openstack-kolla16:28
sdakehrw i have provided a partial review of the review you linked earlier16:29
hrwsdake: sorting is not related to any BP16:30
hrwit is just a small usability tweak16:30
hrwadded note16:31
sdakehrw - i see - well i thought that was related to multiarch16:31
hrwsdake: https://review.openstack.org/#/c/430940 ones is16:31
hrwwill add link there16:31
sdakehrw that is how you do the linkage, you can apply it to your other review :)16:31
sdakesounds good16:31
hrwI have V7 of 430940 to send - just not enough changes to make an update16:32
sdakehrw you can always update the commit message without having other changes attached16:34
sdakegit rebase -i16:34
sdakeuse the reword keyword on the patch in question16:35
sdakegit review16:35
hrwsdake: is there a way to mark image as not-buildable? so "kolla failingimage" would end immediately without doing failed build?16:35
hrwsdake: I use 'git commit --amend' - effect same16:35
hrwbranch has only this one patch ;d16:35
*** mannamne has quit IRC16:35
sdakehrw yup that will work too :)16:36
hrwand yes, git rebase is better than it was years ago16:36
hrwsdake: is your -1 on sorting patch related to lack of link to BP or to code? if first then please drop it16:41
sdakehrw lack of link to bp - use the reply button in the review to tell me i'm wrong and click post16:41
hrwok16:41
hrwsdake: I wrote ~that in my reply16:41
Jeffrey4lsdake, could u ack https://review.openstack.org/43273816:42
sdakehrw click this link: https://review.openstack.org/#/c/432774/1//COMMIT_MSG16:43
hrwdone. hope that properly. gerrit is not my strongest part16:44
*** sdake has quit IRC16:50
openstackgerritMerged openstack/kolla-ansible master: Fix horizon pull error  https://review.openstack.org/43273816:51
openstackgerritMerged openstack/kolla-ansible master: Fixes problem with kolla-kubernetes 3.0.2 images  https://review.openstack.org/43272516:51
*** adrian_otto has quit IRC16:52
*** zhubingbing has quit IRC16:56
*** tonanhngo has joined #openstack-kolla16:58
sbezverkJeffrey4l: Do you have more info about placement? and if its binaries are already in one of kolla's images17:00
Jeffrey4lsbezverk, yep. check https://github.com/openstack/kolla/blob/master/docker/nova/nova-placement-api/Dockerfile.j217:00
Jeffrey4lit is part of nova now. and will be split into a stand alone project in the future. sbezverk17:01
*** mgoddard has joined #openstack-kolla17:01
sbezverkJeffrey4l: thanks, what about resources it uses, like ports etc17:01
Jeffrey4lit use 8780.17:02
Jeffrey4land require registering a new endpoint.17:03
Jeffrey4lsbezverk, check https://github.com/openstack-dev/devstack/blob/master/lib/placement17:03
sbezverkJeffrey4l: what about db?17:07
sbezverkdoes it use different db, or nova-api?17:07
Jeffrey4lsbezverk, now it share the nova_api database.17:07
Jeffrey4lit will own its own data in the future.17:07
Jeffrey4li saw one nova bp for this.17:08
sbezverkJeffrey4l: cool, so we just need to bootstrap endpoints and create a service for it17:08
Jeffrey4lyep.17:08
sbezverkJeffrey4l: thanks..17:09
Jeffrey4lnp.17:09
Jeffrey4lgonna to bed. bye.17:09
hrwbye Jeffrey4l17:09
*** rdotcom has joined #openstack-kolla17:15
*** salv-orlando has joined #openstack-kolla17:19
*** jrich523 has quit IRC17:24
*** salv-orlando has quit IRC17:24
*** leseb has joined #openstack-kolla17:29
openstackgerritMarcin Juszkiewicz proposed openstack/kolla master: debian: import key for download.ceph.com repository  https://review.openstack.org/43277817:31
*** vcn[m] has quit IRC17:34
*** bachp has quit IRC17:36
*** retr0h has quit IRC17:36
openstackgerritMarcin Juszkiewicz proposed openstack/kolla master: keystone-base: enable Debian support  https://review.openstack.org/43277917:37
hrwhm. debian/jessie is complicated stuff when jessie-backports is used as a tag17:43
*** sdake has joined #openstack-kolla17:53
sdakehrw after clicking that link click my review - and click reply17:53
openstackgerritMarcin Juszkiewicz proposed openstack/kolla master: debian: default to jessie-backports if it is used as base_tag  https://review.openstack.org/43278017:54
sdakehrw and type in what yo uthink is appropriate17:56
hrwsdake: done17:56
sdakehrw now type post to the review17:56
sdakeyou can use the up arrow to get back to the oriignal review17:56
*** eaguilar has quit IRC17:57
sdakethe green one in gerrit17:57
openstackgerritMarcin Juszkiewicz proposed openstack/kolla master: keystone-base: enable Debian support  https://review.openstack.org/43277917:57
*** eaguilar_ has joined #openstack-kolla17:57
sdakehrw i se you figured it out :)17:57
sdakenice job :)17:57
openstackgerritSerguei Bezverkhi proposed openstack/kolla-kubernetes master: DO NOT MERGE: Hunting for canal plugin failure  https://review.openstack.org/43278118:00
hrwsdake: I will do some work on x86-64 to get Debian stuff enabled and then check on aarch64;D18:00
sdakesound sgood18:00
sdakeI think enabling x86_64 for debian will be easier as my understanding is the package names are different from ubuntu18:01
sdakeand aarch64 has no gating mechanism in infra18:01
sdakeso sounds like a good place to start18:01
hrwsdake: it is cleanly visible that no one built Debian target for a while18:01
sdakehrw for awhile/ever18:01
sdakeit was almsot deprecated this cycle18:01
sdakeyou came just at the right time18:01
hrw;)18:02
sdakeour general rule of thumb is if someone maintains it, it can stay in tree18:02
sdakeif nobody maintains it we deprecate it18:02
*** rwallner has joined #openstack-kolla18:03
sdakethen there ia a deprecation policy that applies to openstack which we follow18:03
*** flaper87 has quit IRC18:05
*** rwallner has quit IRC18:08
*** flaper87 has joined #openstack-kolla18:11
*** sp__ has quit IRC18:12
*** goldyfruit has quit IRC18:14
hrw 143 files changed, 176 insertions(+), 176 deletions(-)18:27
hrwenabling debian in docker/ whereever ubuntu is present18:27
hrwsdake: changeset which has +2 (like that sorting one) needs to be then merged by someone with proper permissions, right? so >10 March too probably, right?18:33
sdakehrw it might merge prior to march 10 - depends on the kindness of reviewers18:33
hrwsure18:33
sdakehrw an easy to review patch like that is easy to handle18:34
hrwmy laptop will have lot of things to do during night ;D building all images for x86-64/debian ;D18:34
hrwsdake: yep18:34
sdakehrw it takes two +2s to approve a change18:34
hrwok18:34
hrwgood to know18:34
sdakehrw the second person to +2 it applies a workflow +w to the patch18:34
sdakethis merges the change into master18:35
hrwok18:35
sdakeanyone can vote +1 on a patch if they like, however, a +1 does not trigger a merge18:35
sdakeanyone can also vote -1 on a patch if they like18:35
hrwsure18:35
sdakehrw will you be at the ptg?18:36
hrwsdake: no. make next one in Europe so I will attend18:36
sdakehrw i dont choose where they are located18:37
sdakethe openstack foundation determines the location of the ptgs18:37
sdakealthough I did see a tweet from the foundation that as a result of the immigration chaos in the US, 2018 events will be located outside the US18:37
sdakeit was unclera if they were speaking of summit only or also PTGs18:38
hrwsdake: team's travel budget is limited. I already have few trips during year where my manager has to approve.18:38
sdakehrw: http://lists.openstack.org/pipermail/foundation/2017-January/002465.html18:39
sdakehrw ya everyone has travel budget constraints18:39
*** Pavo has joined #openstack-kolla18:39
sdakehrw there is always the travel support program as well18:39
sdakeas kolla is global and mostly not US based, we struggle to gain quorum in any location f2f18:40
hrwsdake: I prefer to leave travel support program to those who are fresh in community, freelancers etc.18:41
sdakehrw sounds good18:41
*** mdorman has quit IRC18:42
hrwsdake: I remember 2007 when my FOSDEM trip was so complicated that one company just bought me flight tickets so I could attend it as whole.18:42
sdakencie i have never been to fosdem18:42
sdakei don't have a desire to attend though as I'm not super keen on travel18:43
hrwfew years later I took one of local friends on similar conditions to FOSDEM. He got travel and hotel covered and only had to take money for food/living.18:43
hrwon second day we gave him also some € for that.18:43
*** mgoddard has quit IRC18:43
hrwhe was18:44
hrwhe was 'contributing developer' in arch or alpine at that moment and spent most of time with distro developers. lot of f2f meetings etc18:44
hrwfrom my perspective it was worth spending my money ;D18:45
sdakeyup i often help people out in the community with room sharing although I dont prefer it18:45
hrwsdake: fosdem is quite unique. this year it had 645 talks during two days18:45
sdakealthough my wife travels with me to most events I go to as she has her own set of things to work on in openstack18:45
*** tonanhngo has quit IRC18:46
hrw;)18:46
sdakeanother way of putting that is i travel with her as she is platinum airliner :)18:46
sdakethis year I may sadly make platinum...18:46
* sdake groans18:47
hrw;DD18:47
sdakei am gold this year18:47
sdakewith only 4 travel events in 201618:47
sdakeor 3 - i dont recall18:47
hrwIn europe I usually travel with cheap airlines18:48
sdakethis year I've got about 6-8 weeks of travel18:48
hrwas all I want is to get from A to B18:48
sdakecisco requires us to use specific carriers - and they handle all the travel parts18:48
hrwsure18:49
hrwRed Hat uses some travel agency but it does not cover Poland so I can choose whatever18:49
sdakeand I can't even book it to my amex to earn points18:49
sdakeya I spent some time at red hat18:49
hrwamex... do not take it with you to europe18:49
hrwsdake: I know ;D18:50
sdakered hat went through a few trave l agendcies18:50
sdakehrw I have an amex and a visa18:50
sdakevisa is for europe18:50
sdakeamex is for everywhre visa is not accepted18:50
sdakei centralize everything on my amex - one bill to pay every month plus lots o points18:50
hrwsdake: ;)18:50
sdakethe best part about amex is its not  a credit card18:51
sdakeso you really can't  go into debt ;-)18:51
hrw;)18:51
sdakeif yo udont pay the bill, they shut it off (I assume, I have always paid my bill in full every month)18:51
hrwkolla is perfect tool to make machine busy... lot of i/o and cpu used18:51
sdakehrw kolla-build is very busy18:52
hrwyep18:52
hrwespecially without -T118:52
sdakeyar - I have gige, and arizona is the central peering location for netwroking in the united states18:52
sdakeso my internet is *really* fast18:52
hrw146 containers so far18:52
hrwI have 250/20Mbps18:53
hrwbut still a way to inject proxy into kolla would be lovely18:53
sdakethats pretty good18:53
sdakethere is a way to do proxy with kolla18:53
sdakethat problem has long since been solved18:53
hrwhttp_proxy is one thing. apt has own proxy settings18:58
hrw+{% block base_ubuntu_package_apt_conf %}18:59
hrw+COPY apt_conf.{{ base_distro }} /etc/apt/apt.conf.d/18:59
hrw+{% endblock %}18:59
hrwin docker/base/Dockerfile.j2 handles it18:59
hrwI have 10GB squid proxy in lan18:59
*** openstackgerrit has quit IRC19:03
*** Pavo has quit IRC19:03
hrwbye19:05
*** pcaruana has joined #openstack-kolla19:12
*** vcn[m] has joined #openstack-kolla19:14
sdakehrw im off as well enjoy19:17
*** rhallisey has joined #openstack-kolla19:19
*** salv-orlando has joined #openstack-kolla19:20
*** salv-orlando has quit IRC19:24
*** mattmceuen has joined #openstack-kolla19:27
*** mgoddard has joined #openstack-kolla19:27
*** fragatina has joined #openstack-kolla19:33
*** fragatina has quit IRC19:37
*** fragatina has joined #openstack-kolla19:38
*** sdake_ has joined #openstack-kolla19:46
*** sdake has quit IRC19:48
*** openstackgerrit has joined #openstack-kolla19:50
openstackgerritMarcin Juszkiewicz proposed openstack/kolla master: debian: enable all images enabled for Ubuntu  https://review.openstack.org/43278719:50
*** eaguilar_ has quit IRC19:55
*** matrohon has joined #openstack-kolla19:55
*** mannamne has joined #openstack-kolla19:56
*** bachp has joined #openstack-kolla19:58
*** retr0h has joined #openstack-kolla19:58
*** sdake_ has quit IRC20:02
*** sdake has joined #openstack-kolla20:06
*** sdake has quit IRC20:14
*** sdake has joined #openstack-kolla20:14
*** schwicht_at_work has quit IRC20:24
sbezverksdake: if you want to chat, I will be back in 10 minutes with my coffee and ready20:25
sdakeok20:25
*** fragatina has quit IRC20:26
*** sdake has quit IRC20:34
*** sdake has joined #openstack-kolla20:37
*** tonanhngo has joined #openstack-kolla20:37
*** tonanhngo has quit IRC20:38
*** mdorman has joined #openstack-kolla20:39
*** bmace has quit IRC20:40
*** bmace has joined #openstack-kolla20:40
*** tvignaud has joined #openstack-kolla21:16
*** matrohon has quit IRC21:17
*** eaguilar has joined #openstack-kolla21:17
*** salv-orlando has joined #openstack-kolla21:21
*** eaguilar has quit IRC21:22
*** salv-orlando has quit IRC21:26
*** l4yerffeJ_ has joined #openstack-kolla21:34
*** Jeffrey4l_ has joined #openstack-kolla21:35
*** Jeffrey4l has quit IRC21:35
*** l4yerffeJ has quit IRC21:35
*** pcaruana has quit IRC21:42
*** sdake_ has joined #openstack-kolla21:44
*** sdake has quit IRC21:48
*** l4yerffeJ_ has quit IRC21:52
*** l4yerffeJ_ has joined #openstack-kolla21:52
*** tvignaud has quit IRC21:57
*** n0isyn0ise has quit IRC21:58
*** n0isyn0ise has joined #openstack-kolla22:00
*** l4yerffeJ_ has quit IRC22:01
*** l4yerffeJ_ has joined #openstack-kolla22:01
*** sdake_ has quit IRC22:02
*** sdake has joined #openstack-kolla22:02
*** mannamne has quit IRC22:03
*** mattmceuen has quit IRC22:28
*** rwallner has joined #openstack-kolla22:42
*** financeMan has joined #openstack-kolla22:46
*** mgoddard has quit IRC22:50
financeManhowdy, how do I pull down a stable branch of kolla for deploy?22:50
financeMankolla-ansible pull -branch branchname?22:51
*** sdake has quit IRC22:52
*** mattmceuen has joined #openstack-kolla22:52
*** mannamne has joined #openstack-kolla23:01
sbezverkfinanceMan: git clone kolla-ansible /cd kolla-ansible / git checkout {branch you want}23:03
financeManam I able to do a apt-get install openstack-kolla yet :)?23:05
financeMani totally blew my kolla instance23:07
financeManand I think its cause i pulled master and it's not stable sbezverk23:08
sbezverkfinanceMan: I really doubt that master is that unstable..23:08
*** goldyfruit has joined #openstack-kolla23:15
*** goldyfruit has quit IRC23:15
financeMansbezverk I ran "kolla-ansible deploy" and got "Starting Heka Container ->  Internal Server Error (\"{\"message\":\"maximum retry count cannot be used with restart policy \\'unless-stopped\"23:17
sbezverkfinanceMan: there might be 1024 reason for this error, in 99.99% cases is misconfiguration..23:22
*** salv-orlando has joined #openstack-kolla23:22
sbezverkI suggest you carfully examin your configuration and the destination system.23:22
financeMansbezverk http://pastebin.com/hvPbeXhC23:23
financeMansingle node machine23:23
sbezverkfinanceMan: kolla-ansible docs are in good shape, you need to make sure that the host where you install complies with all requirements..23:25
sbezverkfinanceMan: one of the reason why I think it is your local machine problem is master runs at the gate now and it does not show failures23:25
*** salv-orlando has quit IRC23:27
financeMansbezverk I tried to destroy my instances through kolla-ansible destroy and the now deploy don't work23:32
financeManit used to be fine23:32
financeManis there any commands I can from to just start from scratch?23:32
financeMandelete the kolla folder?23:32
*** dtturner has quit IRC23:33
*** dtturner has joined #openstack-kolla23:34
sbezverkfinanceMan: no, there is a tool that helps you to really clean up, 1 sec, let me check23:36
*** jmccarthy has quit IRC23:36
sbezverkfinanceMan: in kolla-ansible/tools folder23:37
sbezverkthere are some cleanup_* script23:37
sbezverkthey will help you to make sure that all leftovers from previous installation are gone23:37
*** rwallner_ has joined #openstack-kolla23:47
*** rwallner has quit IRC23:47
*** rwallner_ has quit IRC23:52
openstackgerritOpenStack Proposal Bot proposed openstack/kolla master: Updated from global requirements  https://review.openstack.org/43196123:57
openstackgerritOpenStack Proposal Bot proposed openstack/kolla-ansible master: Updated from global requirements  https://review.openstack.org/43196223:57
*** financeMan has quit IRC23:58
*** jrobinson has joined #openstack-kolla23:59

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