Friday, 2019-12-13

*** zhanglong has joined #openstack-kolla00:04
*** k_mouza has quit IRC00:12
*** igordc has quit IRC00:12
*** wuchunyang has joined #openstack-kolla00:22
*** dosaboy has joined #openstack-kolla00:23
*** wuchunyang has quit IRC00:27
*** ricolin has joined #openstack-kolla00:43
*** xinliang has joined #openstack-kolla00:45
*** zhanglong has quit IRC00:49
*** zhanglong has joined #openstack-kolla00:56
*** ricolin has quit IRC01:09
*** ricolin has joined #openstack-kolla01:09
*** zhanglong has quit IRC01:34
*** zhanglong has joined #openstack-kolla01:37
*** gregoryo has joined #openstack-kolla02:17
*** schwicht has joined #openstack-kolla02:30
*** schwicht has quit IRC02:50
*** wuchunyang has joined #openstack-kolla02:52
*** wuchunyang has quit IRC02:56
*** skramaja has quit IRC03:44
*** zhanglong has quit IRC03:55
*** zhanglong has joined #openstack-kolla04:12
*** xinliang has quit IRC04:19
*** zhanglong has quit IRC04:22
*** zhanglong has joined #openstack-kolla04:24
*** zhanglong has quit IRC04:34
*** zhanglong has joined #openstack-kolla04:36
*** raukadah is now known as chkumar|rover05:08
*** zainub_wahid has joined #openstack-kolla05:12
*** dave-mccowan has quit IRC05:30
*** skramaja has joined #openstack-kolla06:00
mnasiadkayoctozepto: replied :)06:01
mnasiadkaso people are still using kolla-ceph even though it is deprecated? nice06:02
*** pcaruana has joined #openstack-kolla06:02
*** ricolin has quit IRC06:10
*** sri_ has joined #openstack-kolla06:16
mnasiadkayoctozepto: I'll be back around 11, so we can discuss :)06:16
sri_Morning06:20
*** gregoryo has quit IRC06:24
*** xinliang has joined #openstack-kolla06:30
*** jbadiapa has joined #openstack-kolla06:43
*** lemko has joined #openstack-kolla06:59
*** zhanglong has quit IRC07:06
*** zhanglong has joined #openstack-kolla07:08
*** logan- has quit IRC07:09
*** logan_ has joined #openstack-kolla07:10
*** logan_ is now known as logan-07:10
*** wuchunyang has joined #openstack-kolla07:46
*** wuchunyang has quit IRC07:51
*** rgogunskiy has quit IRC07:54
*** rgogunskiy has joined #openstack-kolla07:56
*** zhanglong has quit IRC07:56
*** zhanglong has joined #openstack-kolla07:58
*** wuchunyang has joined #openstack-kolla07:58
*** rgogunskiy has quit IRC07:59
*** rgogunskiy has joined #openstack-kolla07:59
*** rgogunskiy has quit IRC08:00
*** zhanglong has quit IRC08:02
*** wuchunyang has quit IRC08:02
*** zhanglong has joined #openstack-kolla08:04
*** ricolin has joined #openstack-kolla08:06
*** ricolin has quit IRC08:06
*** ricolin has joined #openstack-kolla08:07
*** wuchunyang has joined #openstack-kolla08:09
*** wuchunyang has quit IRC08:13
yoctozeptomorning08:15
yoctozeptomnasiadka: lovely indeed08:15
*** zhanglong has quit IRC08:16
*** ricolin has quit IRC08:47
*** FlorianFa has quit IRC08:49
yoctozeptomnasiadka: well, I learnt our approach is less problematic than using ceph-ansible08:52
yoctozeptoso might be the "why"08:52
hrwmorning08:54
hrwinfra patch fails terribly on CI but I left some notes there. please take a look once you have a time08:55
yoctozeptohrw: agreed08:57
hrwyoctozepto: to?08:59
yoctozeptohrw: you rebase idea, replied in review08:59
yoctozeptoyour*08:59
hrwk09:00
*** FlorianFa has joined #openstack-kolla09:01
yoctozeptomgoddard: please look at https://review.opendev.org/676128 (specifically my comment re mariadb - I suppose you are the most well-versed in it)09:05
patchbotpatch 676128 - kolla - Build images with healthcheck scripts for core ser... - 19 patch sets09:05
yoctozeptobad news09:08
yoctozeptostein->train upgrade permabroken09:08
yoctozeptohttps://storage.bhs.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_bee/698718/2/check/kolla-ansible-ubuntu-source-upgrade-ceph/bee3111/primary/logs/ansible/upgrade09:08
yoctozeptomariadb not accepting some migration09:08
*** FlorianFa has quit IRC09:09
hrwhm. I see that infra images can make some poeple angry as some overrides will be killed09:13
hrwbase -> noninfra-base -> openstack-base -> nova-base -> nova-compute09:13
hrwbase -> mariadb09:13
*** dougsz has joined #openstack-kolla09:15
*** gfidente has joined #openstack-kolla09:15
*** k_mouza has joined #openstack-kolla09:15
yoctozeptoRow size too large. The maximum row size for the used table type, not counting BLOBs, is 8126. This includes storage overhead, check the manual09:19
openstackgerritMark Goddard proposed openstack/kolla-ansible stable/train: CI: Set zuul's ansible_python_interpreter to python2  https://review.opendev.org/69871809:20
yoctozeptohttps://mariadb.com/kb/en/library/troubleshooting-row-size-too-large-errors-with-innodb/#solving-the-problem09:21
hrwmgoddard: I unified base_centos_packages similar to your change in centos8 patch. will send for review in few minutes09:21
*** FlorianFa has joined #openstack-kolla09:22
yoctozeptoIn MariaDB 10.1 and before, and in MySQL 5.6 and before, the COMPACT row format was the default row format.09:22
yoctozeptoanyone brave enough to write a script to convert table types for ubuntu upgrade? :D09:23
yoctozepto(otherwise we have b0rken CI, just so you know ;p)09:23
mnasiadkayoctozepto: well, maybe in V we could write something simple to use ceph orchestrator09:24
mnasiadkayoctozepto: I think you are brave enough :)09:25
yoctozeptomnasiadka: no, already deep in some s**t09:26
yoctozeptoin fact we might want to convert everyone09:29
openstackgerritMarcin Juszkiewicz proposed openstack/kolla master: base: unify CentOS packages list  https://review.opendev.org/69885509:29
yoctozeptobecasue if you are coming from like queens or rocky then you might also have these old format09:29
hrwmgoddard: ^^09:30
yoctozeptoin centos too09:30
yoctozeptohrw: nice09:31
hrwyoctozepto: mgoddard did that in centos8 patch and it is needed for infra images too09:31
hrwhave to also find a way to not give install_type for infra images to make sure that dockerfile will be same for bin/src09:32
hrwINFO:kolla.common.utils.base:Step 6/34 : ENV KOLLA_BASE_DISTRO=centos     KOLLA_INSTALL_TYPE=source     KOLLA_INSTALL_METATYPE=mixed     KOLLA_DISTRO_PYTHON_VERSION=2.7     KOLLA_BASE_ARCH=x86_6409:36
hrwthis block09:36
hrwfound09:37
yoctozeptomgoddard: https://bugs.launchpad.net/nova/+bug/185629609:37
openstackLaunchpad bug 1856296 in kolla-ansible "upgrade to Train might fail due to mariadb row format" [Critical,Triaged]09:37
hrwyoctozepto: how goes epel removal?09:37
yoctozeptohrw: 0 progress since then09:38
yoctozeptomostly due to all issues that pop up :-)09:38
yoctozeptowith mnasiadka we decided that we "get rid" of the tgt issue09:38
yoctozeptobecause with c8 users are forced to reinstall09:38
hrwok09:39
yoctozeptoso we could just drop the tgt and not care09:39
yoctozeptoI guess other uses were not that important09:39
yoctozeptobut still to check09:39
hrwI had to remove epel to have unified base image09:39
yoctozeptoand glad you did09:39
yoctozeptothat's the goal09:39
hrwnow base image should be install_type agnostic09:40
hrwINFO:kolla.common.utils.base:Successfully tagged infra/centos-infra-base:9.1.009:41
hrwINFO:kolla.common.utils.base:Built at 2019-12-13 10:41:07.566909 (took 0:01:05.919345)09:41
hrwand then next build with different install_type:09:41
hrwINFO:kolla.common.utils.base:Successfully tagged infra/centos-infra-base:9.1.009:41
hrwINFO:kolla.common.utils.base:Built at 2019-12-13 10:41:21.249479 (took 0:00:07.551533)09:41
hrwall steps from cache09:42
hrwso noninfra-base image will not be needed09:42
hrwlet now build all centos binary/source images09:42
hrwand then go infra image by infra image and get rid of install_type checks09:43
mgoddardyoctozepto: any idea why we have started to see this DB row isse?09:44
hrwyoctozepto: ops. epel is still there. we just had two different places to install it09:45
mgoddardshould nova migrations include a step to convert tables to dynamic row format?09:45
yoctozeptomgoddard: not sure which 'why' you are asking about :-)09:45
mgoddardyoctozepto: what changed to make it fail?09:46
hrwmgoddard: morning09:46
mgoddardmorning hrw09:46
yoctozeptomgoddard: ah, nope09:46
yoctozeptoI skimmed over recent changes but found nothing relevant09:46
yoctozeptoindeed it worked before for us09:46
mgoddardsame09:47
mgoddardcould be a mariadb change?09:47
yoctozeptothough the issue is legit, might be just mariadb started being serious about reporting09:47
openstackgerritMarcin Juszkiewicz proposed openstack/kolla master: base: unify CentOS packages list  https://review.opendev.org/69885509:47
hrwyoctozepto: indent fixed09:47
yoctozeptohttps://downloads.mariadb.org/mariadb/+releases/09:48
yoctozeptoyup09:48
hrwhm. I may see a new issue with infra images... We do 'binary' build and get infra+binary images. then we build 'source' on same machine and get infra+source. all is fine as newly tagged 'infra' images are just using cache layers from previous build. we can push infra+binary+source images to registry.09:54
*** wuchunyang has joined #openstack-kolla09:54
hrwbut if we do 'binary' on one and 'source' on other machine (or with cleaning image between) then we have separate set of layers making 'infra' images09:55
hrwso kolla-publish jobs will have to do infra+binary+source per distro09:55
hrwinstead of current binary-distro, source-distro09:55
mnasiadkaback09:56
*** wuchunyang has quit IRC09:59
yoctozeptomnasiadka: mariadb heat09:59
yoctozepto(not openstack heat I mean)09:59
mnasiadkadeprecate mariadb09:59
hrwmgoddard: thanks for review. will go on10:00
mnasiadkayoctozepto: isn't there a bloody tool for mariadb upgrades that converts everything?10:00
mnasiadkayoctozepto: like mysql_upgrade?10:00
yoctozeptomnasiadka: well, you might not want to convert :-)10:01
yoctozeptosuch long rows are really not so performant anyway ;p10:01
mnasiadkayoctozepto: I don't care about performance for now, just care that it works - shouldn't we run mysql_upgrade/mysqlcheck to find out if we need to do anything?10:02
mnasiadkaoh geez, I think mysql_upgrade doesn't even check it10:04
mnasiadkaso now we need to update to DYNAMIC? uhh10:04
yoctozeptomnasiadka: yeah10:04
yoctozeptobest to do it on all tables - simplest and safest for the future10:04
mnasiadkawell, just to be sure - it would be nice to run mysqlcheck --verbose to understand if mysql_upgrade would help us here10:05
*** wuchunyang has joined #openstack-kolla10:06
yoctozeptomnasiadka: it wouldn't - it's just the default that changed10:06
mnasiadkawell then we need to alter all tables, and we're done10:06
mnasiadkashould we do some backup before? :D10:06
yoctozeptomnasiadka: we leave that exercise to users10:07
yoctozeptoI do wonder why it suddenly started appearing10:07
yoctozeptonewer mariadb got released10:07
yoctozeptoand that's it10:07
openstackgerritMarcin Juszkiewicz proposed openstack/kolla master: base: unify CentOS packages list  https://review.opendev.org/69885510:08
yoctozeptoas if it started checking this more rigorously10:08
hrwmgoddard: commit message with explanation + note + overrides compatibility10:08
mnasiadkayoctozepto: well, then we just need to check for mariadb version after upgrade and generate an sql statement to update all tables to row_format=DYNAMIC10:09
yoctozeptomnasiadka: yeah, we do :-) the question was who was to do it10:09
*** wuchunyang has quit IRC10:10
osmanlicilegimorning10:11
hrwosmanlicilegi: hi10:11
hrwosmanlicilegi: https://review.opendev.org/698789 will make rebase to 3.8 a bit easier10:12
patchbotpatch 698789 - kolla - Use APT pinning to set preferred version of RabbitMQ - 1 patch set10:12
osmanlicilegihrw: I just saw it, thank you10:12
hrwosmanlicilegi: side effect of looking at too many files while working on infra images10:12
osmanlicilegihrw: I agree with you. pinning something everywhere is totaly a mess10:13
hrwosmanlicilegi: and complicates dockerfiles10:14
osmanlicilegimgoddard: with erlang 22, hipe support is dropped.10:15
osmanlicilegihttps://www.rabbitmq.com/which-erlang.html10:16
*** wuchunyang has joined #openstack-kolla10:17
*** wuchunyang has quit IRC10:22
yoctozeptoyeah, no hype :D10:23
hrwdid infra+binary build. started infra+source and first 78 images were built in 0:00.00.* time - pure cache use10:33
hrwto compare: first source one was 3:3410:34
openstackgerritMerged openstack/kolla master: Build images with healthcheck scripts for core services  https://review.opendev.org/67612810:38
mnasiadkawhoa, I haven't noticed +2s for that10:39
hrw;D10:39
openstackgerritMark Goddard proposed openstack/kayobe master: WIP: Add seed VM provisioning CI job  https://review.opendev.org/55970610:40
openstackgerritMarcin Juszkiewicz proposed openstack/kolla master: base: unify CentOS packages list  https://review.opendev.org/69885510:42
hrwpep8 fixed10:42
openstackgerritMerged openstack/kolla master: Add Networking-MLNX to Neutron-Base packages  https://review.opendev.org/68911010:47
openstackgerritMark Goddard proposed openstack/kayobe master: WIP: Test control plane deployment using tenks  https://review.opendev.org/62007410:48
yoctozeptomnasiadka: yea, I had a comment about different treatment of mariadb wsrep status10:56
yoctozeptomgoddard: ^10:56
mnasiadkayoctozepto: yeah, found it - I based it on clustercheck script and mariadb docs10:56
openstackgerritMerged openstack/kolla master: Use APT pinning to set preferred version of RabbitMQ  https://review.opendev.org/69878910:59
*** ricolin has joined #openstack-kolla10:59
mgoddardmnasiadka: we do mysql_upgrade during upgrade11:01
mgoddardno mention of this in release notes that I can see: https://mariadb.com/kb/en/library/mariadb-10321-release-notes/11:01
mnasiadkamgoddard: good, then we need to alter the damn tables11:01
mgoddardyep11:01
*** ricolin_ has joined #openstack-kolla11:02
*** ricolin has quit IRC11:04
openstackgerritMarcin Juszkiewicz proposed openstack/kolla master: docker: fix 'Empty continuation line found' warnings  https://review.opendev.org/69887511:17
*** ricolin_ has quit IRC11:18
openstackgerritMark Goddard proposed openstack/kayobe master: WIP: Add seed VM provisioning CI job  https://review.opendev.org/55970611:26
openstackgerritDoug Szumski proposed openstack/kayobe master: Support custom Kolla group_vars  https://review.opendev.org/69888211:28
openstackgerritMark Goddard proposed openstack/kayobe master: WIP: Test control plane deployment using tenks  https://review.opendev.org/62007411:29
*** zainub_wahid has quit IRC11:51
openstackgerritMarcin Juszkiewicz proposed openstack/kolla master: WIP: Create 'infra' type of images  https://review.opendev.org/69875312:00
hrwCI will probably still fail but that's update of work12:01
*** EmilienM has quit IRC12:12
*** EmilienM has joined #openstack-kolla12:12
openstackgerritMarcin Juszkiewicz proposed openstack/kolla master: Add Debian/source aarch64 experimental job  https://review.opendev.org/55765912:13
hrwCentOS job got dropped12:17
*** EmilienM is now known as EvilienM12:17
*** kplant has joined #openstack-kolla12:19
*** wuchunyang has joined #openstack-kolla12:22
*** mgoddard has quit IRC12:30
*** kozhukalov has joined #openstack-kolla12:30
openstackgerritMark Goddard proposed openstack/kayobe master: WIP: Test control plane deployment using tenks  https://review.opendev.org/62007412:33
openstackgerritMark Goddard proposed openstack/kayobe master: WIP: Add seed VM provisioning CI job  https://review.opendev.org/55970612:37
*** xinliang has quit IRC12:37
*** sm806 has quit IRC12:38
*** sm806 has joined #openstack-kolla12:39
*** dave-mccowan has joined #openstack-kolla12:40
yoctozeptomgoddard (hopefully around), mnasiadka: I might be able to spare some cycles to fix upgrade this evening if you don't do before12:40
yoctozeptoper the agreed actions ;-)12:40
yoctozeptobtw, https://ethercalc.openstack.org/kolla-infra-service-matrix12:41
yoctozeptoI think this might be helpful to have an overview over why something breaks in this or that release/flavor12:41
yoctozeptohrw: is that you?12:44
osmanlicilegime12:45
yoctozeptoosmanlicilegi: ack12:47
yoctozeptothanks12:47
osmanlicilegiyw12:47
*** kozhukalov has quit IRC12:48
*** kozhukalov has joined #openstack-kolla12:48
openstackgerritMerged openstack/kayobe master: Adds support for Ansible 2.8, bump mimimum to 2.6  https://review.opendev.org/69770212:58
*** wuchunyang has quit IRC12:59
*** wuchunyang has joined #openstack-kolla12:59
*** goldyfruit has joined #openstack-kolla13:07
osmanlicilegiyoctozepto: rmq and erlang are ok13:11
hrwnice sheet13:17
hrwlet me add Debian13:18
*** negronjl has quit IRC13:18
*** negronjl has joined #openstack-kolla13:20
*** mgoddard has joined #openstack-kolla13:21
openstackgerritMichal Nasiadka proposed openstack/kolla-ansible master: Initialize fernet keys before starting keystone container  https://review.opendev.org/69871013:33
*** ktibi has joined #openstack-kolla13:39
*** negronjl has quit IRC13:49
*** negronjl has joined #openstack-kolla13:50
openstackgerritPierre Riteau proposed openstack/kayobe master: Bump stack.libvirt-vm role to 1.8.0  https://review.opendev.org/69890613:55
*** negronjl has quit IRC13:59
hrwcan we kill opendaylight images?14:00
hrwit failed again14:00
*** negronjl has joined #openstack-kolla14:03
openstackgerritMark Goddard proposed openstack/kayobe master: WIP: Add seed VM provisioning CI job  https://review.opendev.org/55970614:03
mnasiadkahrw: no, but mark them as unbuildable and wait for tripleo to come :)14:06
hrwmnasiadka: ;D14:07
hrwhm. why infra images patch works locally and not on CI...14:07
hrw(venv3) 15:10 (0s) hrw@puchatek:kolla$ tox -e build-debian-source14:11
hrwalso works14:11
openstackgerritDoug Szumski proposed openstack/kayobe master: Support custom Kolla group_vars  https://review.opendev.org/69888214:12
*** skramaja has quit IRC14:13
yoctozeptohrw: where does it fail?14:23
ktibiHi guys, anyone had use kolla on RHEL os ?14:23
hrwyoctozepto: https://review.opendev.org/#/c/698753/14:24
patchbotpatch 698753 - kolla - WIP: Create 'infra' type of images - 3 patch sets14:24
yoctozeptoktibi: tripleo kinda do14:24
ktibiyoctozepto, yeah but I don't like tripleo :p14:24
yoctozeptoktibi: you want k-a?14:25
hrwktibi: use kolla as 'do builds with rhel as host' or 'do builds with rhel as container os' or 'deploys on rhel'?14:25
yoctozeptothen follow for centos and let us know ;D14:25
ktibiyeah sorry , k-a14:25
mgoddardktibi: it is 'supported' but untested14:26
ktibihrw, yep the question is : anyone had deploy kolla with k-a on a rhel system ?14:26
*** rgogunskiy has joined #openstack-kolla14:26
ktibimgoddard, ok thx ;)14:26
*** rgogunskiy has quit IRC14:31
*** vmixor has joined #openstack-kolla14:32
*** dmellado has quit IRC14:36
*** dmellado has joined #openstack-kolla14:39
*** igordc has joined #openstack-kolla14:40
*** negronjl has quit IRC14:41
*** sri_ has quit IRC14:43
*** negronjl has joined #openstack-kolla14:44
*** graysonh has joined #openstack-kolla14:45
yoctozeptoINFO:kolla.common.utils.ceph-base:Step 1/7 : FROM kolla/debian-infra-base:9.1.014:48
yoctozeptoINFO:kolla.common.utils.ceph-base:Step 2/7 : FROM kolla/debian-source-base:9.1.014:48
yoctozeptolooking wrong14:48
yoctozeptohrw: ^14:48
yoctozeptoI guess you locally have this other image14:48
mgoddardshould we call it debian-base?14:50
mgoddarddebian-mariadb etc14:50
openstackgerritWill Szumski proposed openstack/kolla-ansible master: Add extras directory to prometheus config  https://review.opendev.org/68837414:56
*** kozhukalov has quit IRC14:56
vmixorHi14:57
mgoddardhi vmixor14:59
openstackgerritMerged openstack/kayobe master: Improve detection of base path  https://review.opendev.org/65586415:01
openstackgerritWill Szumski proposed openstack/kayobe-config master: Synchronise with kayobe for Train  https://review.opendev.org/69892715:01
vmixorI am trying to deploy Train and I am getting error: "The conditional check 'api_interface_address is not defined' failed. " What should I set to this variable? It had been initiated automaticaly up to Stein.15:02
vmixorhi mgoddard15:03
yoctozeptovmixor: does it mention kolla_address in there?15:03
yoctozeptoif so, please ensure ansible and kolla_ansible are both installed in the same environment15:03
yoctozeptoverify with pip list15:03
yoctozepto(or pip freeze)15:04
graysonhvmixor: In response to your reply Thursday, I am setting the bootstrap_bs label on the devices. Each host has 16 spinny disks and 8 SSDs and I'm trying to move the block.db to the latter while keeping block on the former. Here is what they look like right now https://pastebin.com/BT6tawcj, which is what my interpretation of the docs make me think they should be, but kolla never detects them15:04
graysonhAnd yeah, I know ceph kolla support is going away, but we've had no luck with ceph-ansible yet, and we already have a cluster in prod deployed this way that needs to be expanded :(15:04
vmixoryoctozepto: ah, you are right, Ansible is not present in that virtual environment15:06
hrwyoctozepto: infra-base got built just before infra-ceph-base etc15:07
hrwyoctozepto: have to find why it goes that way15:08
hrwyoctozepto: locally it works with clean docker15:08
yoctozeptohrw: I mean source-base15:08
yoctozeptolook more carefully15:08
hrwyoctozepto: locally it works with clean docker15:08
yoctozeptomgoddard: looks like we should guard against the case that ansible is called from outside the virtualenv15:08
yoctozeptohrw: ok15:09
hrwyoctozepto: I use new namespace per build even15:09
openstackgerritDoug Szumski proposed openstack/kayobe master: Support custom Kolla group_vars  https://review.opendev.org/69888215:13
vmixoryoctozepto: now it works, thanks15:17
vmixorgraysonh: tbh, I have no experience with such setup in production, maybe someone else there?15:19
hrw1444 images built with infra patch.15:20
hrwhttps://paste.centos.org/view/c3f6afad15:21
yoctozeptovmixor: yw15:22
mgoddardyoctozepto: hmm. a better check would be to run a playbook that uses ansible_playbook_python to import kolla_ansible15:24
*** wuchunyang has quit IRC15:28
yoctozeptomgoddard: mhm15:31
yoctozeptomgoddard: I guess it's the mariadb upgrade fix for this for rc3 and final release15:32
mgoddardhmm, yes I suppose we need to block the release on it15:34
hrwhttp://zuul.openstack.org/status/change/557659,58 runs on new arm64 pipeline ;D15:43
yoctozeptohrw: at least one good info today ;D15:43
hrwyoctozepto: I have few more good infos today. 4th season of The Expanse is on Amazon Prime. My infra images work (locally). Few more new patches done. Jira card for infra images done. Got info about new arm64 nodes work.15:45
yoctozepto:D15:45
*** kozhukalov has joined #openstack-kolla15:46
hrwweek late birthday gifts arrived at their destination, xmas gifts arrived in paczkomat...15:47
hrwand few more15:47
*** jbadiapa has quit IRC15:48
openstackgerritMarcin Juszkiewicz proposed openstack/kolla master: Add Debian/source aarch64 CI job  https://review.opendev.org/55765915:52
hrwno longer experimental15:53
*** chkumar|rover is now known as raukadah15:53
*** ricolin_ has joined #openstack-kolla15:53
*** ktibi has quit IRC15:54
*** kozhukalov has quit IRC15:56
vmixorMay I ask for review of https://review.opendev.org/#/c/695054/ and https://review.opendev.org/#/c/674579/15:58
patchbotpatch 695054 - kolla-ansible - Support for custom api-paste files - 6 patch sets15:58
patchbotpatch 674579 - kolla-ansible - Adds support of CADF notifying for other APIs next... - 10 patch sets15:58
yoctozeptoasking is permitted :-)16:01
vmixor:)16:02
*** schwicht has joined #openstack-kolla16:07
*** schwicht has quit IRC16:07
mgoddardis anyone looking at https://bugs.launchpad.net/kolla-ansible/+bug/1856296?16:08
openstackLaunchpad bug 1856296 in kolla-ansible "upgrade to Train might fail due to mariadb row format" [Critical,Triaged]16:08
openstackgerritRadosław Piliszek proposed openstack/kolla-ansible stable/stein: CI: Set zuul's ansible_python_interpreter to python2  https://review.opendev.org/69894516:08
openstackgerritWill Szumski proposed openstack/kayobe-config master: Synchronise with kayobe for Train  https://review.opendev.org/69892716:09
openstackgerritRadosław Piliszek proposed openstack/kolla-ansible stable/rocky: CI: Set zuul's ansible_python_interpreter to python2  https://review.opendev.org/69894616:09
openstackgerritWill Szumski proposed openstack/kayobe-config master: Synchronise with kayobe for Train  https://review.opendev.org/69892716:11
*** igordc has quit IRC16:15
yoctozeptomgoddard: /me as promised16:15
mgoddardyoctozepto: great. I missed the promise16:15
yoctozeptomgoddard: already split off backport of your change16:15
mgoddardyoctozepto: will need to combine with ansible_python_interpreter16:15
yoctozeptoyeah, yeah16:15
mgoddarddoesn't the change need to go to train?16:16
yoctozeptomgoddard: to train16:16
mgoddardk16:16
openstackgerritWill Szumski proposed openstack/kayobe-config master: Synchronise with kayobe for Train  https://review.opendev.org/69892716:32
*** ricolin_ has quit IRC16:36
openstackgerritRadosław Piliszek proposed openstack/kolla-ansible stable/train: Fix upgrades from older (<=10.1) MariaDB releases  https://review.opendev.org/69895716:39
yoctozeptomgoddard, mnasiadka: a bit dirty ^16:39
yoctozeptoW-1 mostly due to missing comments, no_log and CI trimmed to just one job16:40
yoctozeptotesting otherwise16:40
yoctozeptoI checked my deploy on c7 and all tables are dynamic16:40
yoctozeptoso we should be fine going for all in Ubuntu16:40
*** kozhukalov has joined #openstack-kolla16:41
yoctozeptoand c7 if need be (for example for some old deployments)16:41
yoctozeptoreviews welcome still ;-)16:41
mgoddardlooks ok to me16:42
mgoddardI have left a very helpful comment16:42
yoctozeptomgoddard: game changer!16:43
mgoddardwon't we have the same problem in stein for centos?16:44
mgoddardstein upgraded to 10.316:44
yoctozeptomgoddard: CI is checking16:44
yoctozeptowe can always backport it16:44
mgoddardwho needs CI when you have me16:44
yoctozeptobut maybe relevant migrations are only in Train16:44
mgoddardtrue, could be16:44
yoctozeptook, proposing to drop CI16:44
mgoddard:)16:44
openstackgerritMerged openstack/kayobe-config master: Synchronise with kayobe for Train  https://review.opendev.org/69892716:44
yoctozeptoall blame on Mark16:45
yoctozeptofor now let Train amend deploys16:45
yoctozeptoif ppl come complaining we can backport16:45
yoctozeptoStein already has mariadb issues :-)16:45
yoctozeptowonder how we avoided the circus with Ubuntu16:46
yoctozeptobut somehow we did16:46
yoctozepto:D16:46
openstackgerritRadosław Piliszek proposed openstack/kolla-ansible stable/train: Fix upgrades from older (<=10.1) MariaDB releases  https://review.opendev.org/69895716:47
yoctozeptoit did not start the job so I fixed the whitespace, eh ;D16:47
yoctozeptosome reply from Matt here: https://bugs.launchpad.net/kolla-ansible/+bug/185629616:47
openstackLaunchpad bug 1856296 in kolla-ansible "upgrade to Train might fail due to mariadb row format" [Critical,Triaged]16:47
yoctozeptomgoddard: ^16:47
mgoddardI saw. Fair enough I suppose16:48
yoctozeptonotified, closed16:49
yoctozeptoall happy16:49
*** schwicht has joined #openstack-kolla16:55
mnasiadkaback again - crazy day16:57
yoctozeptomnasiadka: you don't say16:58
mnasiadkait's good that it's Friday...16:58
mnasiadkaok, let's get back to this bloody fernet16:59
*** graysonh has quit IRC17:01
*** hjensas has joined #openstack-kolla17:03
hrwhttps://review.opendev.org/#/c/557659/ - please review and accept. it works17:12
patchbotpatch 557659 - kolla - Add Debian/source aarch64 CI job - 59 patch sets17:12
hrwand will allow me to close oldest Jira card ;D17:13
mnasiadkathat's interesting, keystone bootstrap task returns always changed=true17:20
hrwand it is now green on all CI jobs17:25
hrwyoctozepto: if you want another trivial change - go for it.17:27
* hrw -> weekend ;D17:27
hrwhave a nice weekend folks17:27
*** wuchunyang has joined #openstack-kolla17:29
*** openstackgerrit has quit IRC17:29
*** openstackgerrit has joined #openstack-kolla17:31
openstackgerritDoug Szumski proposed openstack/kayobe master: Support customising nova-conductor group for cells  https://review.opendev.org/69897017:31
openstackgerritDoug Szumski proposed openstack/kayobe master: WIP: Cells documentation  https://review.opendev.org/69897117:31
*** wuchunyang has quit IRC17:34
*** kozhukalov has quit IRC17:35
yoctozeptomgoddard, mnasiadka: https://review.opendev.org/698957 it worked :D17:39
patchbotpatch 698957 - kolla-ansible (stable/train) - Fix upgrades from older (<=10.1) MariaDB releases - 2 patch sets17:39
*** graysonh has joined #openstack-kolla17:44
mnasiadkamgoddard: cool, but I see a module mysql_command would be useful (if it existed) ;)17:49
openstackgerritRadosław Piliszek proposed openstack/kolla-ansible stable/train: Fix upgrades from older (<=10.1) MariaDB releases  https://review.opendev.org/69895717:52
yoctozeptomnasiadka: we rarely have to resort to doing this17:52
mnasiadkayeah, s/mgoddard/yoctozepto/ ;)17:53
yoctozeptothe only pending issue is the check for ansible+kolla_ansible in the same venv17:54
yoctozeptocomplaints about the kolla_address missing are too numerous17:54
*** dougsz has quit IRC17:57
*** k_mouza has quit IRC17:57
*** wuchunyang has joined #openstack-kolla18:06
*** wuchunyang has quit IRC18:11
yoctozeptohttps://bugs.launchpad.net/kolla-ansible/+bug/185634618:16
openstackLaunchpad bug 1856346 in kolla-ansible "check for kolla_ansible being installed does not take into account it's ansible that has to see it" [Medium,Triaged]18:16
openstackgerritMichal Nasiadka proposed openstack/kolla-ansible master: Initialize fernet keys before starting keystone container  https://review.opendev.org/69871018:25
openstackgerritMichal Nasiadka proposed openstack/kolla-ansible master: Initialize fernet keys before starting keystone container  https://review.opendev.org/69871018:26
openstackgerritMichal Nasiadka proposed openstack/kolla-ansible master: Initialize fernet keys before starting keystone container  https://review.opendev.org/69871018:30
*** wuchunyang has joined #openstack-kolla18:30
*** wuchunyang has quit IRC18:35
*** gfidente has quit IRC18:40
*** wuchunyang has joined #openstack-kolla18:42
openstackgerritMichal Nasiadka proposed openstack/kolla-ansible master: Initialize fernet keys before starting keystone container  https://review.opendev.org/69871018:43
*** wuchunyang has quit IRC18:47
*** lemko has quit IRC18:48
*** wuchunyang has joined #openstack-kolla18:51
*** vkmc has joined #openstack-kolla18:53
vkmchi folks18:53
vkmcmaybe someone can help me out18:53
vkmcI'm building an image with kolla, it never finishes kolla_start18:54
vkmclooking at the error code, the container runs out of memory18:54
vkmcthe host has memory18:54
vkmcbut apparently there is something crazy going inside the container18:54
vkmcsome suggestions on how to debug this?18:54
vkmcattaching a shell to the container is not possible because it respawns every 1min or less18:54
*** wuchunyang has quit IRC18:59
*** igordc has joined #openstack-kolla19:05
klindgren_change the start command in /etc/kolla/<whatever/config.json to something like sleep 100000; <normal command>19:07
klindgren_then you should be able to exec into the container and trying running the command yourself to see what happens19:08
vkmcklindgren_++19:13
vkmcthanks, I will try that19:14
*** rgogunskiy has joined #openstack-kolla19:31
*** rgogunskiy has quit IRC19:36
yoctozeptomgoddard, mnasiadka: if you were so kind to approve patches to 3 stable branches to unbreak them, thanks :-)19:46
*** kplant has quit IRC20:02
*** kozhukalov has joined #openstack-kolla20:10
*** vmixor has quit IRC20:12
*** gmann is now known as gmann_afk20:13
*** wuchunyang has joined #openstack-kolla20:14
*** wuchunyang has quit IRC20:19
*** wuchunyang has joined #openstack-kolla20:25
openstackgerritMichal Nasiadka proposed openstack/kolla-ansible master: Initialize fernet keys before starting keystone container  https://review.opendev.org/69871020:26
openstackgerritMichal Nasiadka proposed openstack/kolla-ansible master: Initialize fernet keys before starting keystone container  https://review.opendev.org/69871020:27
openstackgerritMichal Nasiadka proposed openstack/kolla-ansible master: Initialize fernet keys before starting keystone container  https://review.opendev.org/69871020:28
*** wuchunyang has quit IRC20:30
*** kozhukalov has quit IRC20:31
*** graysonh has quit IRC20:35
*** kozhukalov has joined #openstack-kolla20:37
*** oncall-pokemon has joined #openstack-kolla20:56
*** EvilienM is now known as EmilienM20:57
*** wuchunyang has joined #openstack-kolla21:16
*** vmixor has joined #openstack-kolla21:17
*** wuchunyang has quit IRC21:20
*** wuchunyang has joined #openstack-kolla21:27
*** igordc has quit IRC21:30
*** wuchunyang has quit IRC21:32
*** wuchunyang has joined #openstack-kolla21:42
*** schwicht has quit IRC21:47
*** wuchunyang has quit IRC21:48
*** wuchunyang has joined #openstack-kolla21:49
*** vmixor has quit IRC21:53
*** wuchunyang has quit IRC21:54
*** kozhukalov has quit IRC21:54
*** logan- has quit IRC22:04
*** logan- has joined #openstack-kolla22:06
*** schwicht has joined #openstack-kolla22:10
*** KeithMnemonic1 has joined #openstack-kolla22:15
*** schwicht has quit IRC22:18
*** KeithMnemonic has quit IRC22:19
*** schwicht has joined #openstack-kolla22:20
*** wuchunyang has joined #openstack-kolla22:42
*** schwicht has quit IRC22:44
*** hjensas has quit IRC22:45
*** wuchunyang has quit IRC22:47
*** pcaruana has quit IRC22:48
*** wuchunyang has joined #openstack-kolla22:50
*** wuchunyang has quit IRC22:54
*** wuchunyang has joined #openstack-kolla22:56
openstackgerritWill Szumski proposed openstack/kayobe master: Switch default IPA images to centos  https://review.opendev.org/69904622:56
*** wuchunyang has quit IRC23:06
*** schwicht has joined #openstack-kolla23:11
*** KeithMnemonic1 has quit IRC23:19
*** wuchunyang has joined #openstack-kolla23:33
*** wuchunyang has quit IRC23:42
*** wuchunyang has joined #openstack-kolla23:48
*** wuchunyang has quit IRC23:53

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