*** schwicht has joined #openstack-kolla | 00:01 | |
*** dciabrin_ has quit IRC | 00:06 | |
*** dciabrin has joined #openstack-kolla | 00:16 | |
*** schwicht has quit IRC | 00:56 | |
*** spsurya has joined #openstack-kolla | 01:59 | |
openstackgerrit | gugug proposed openstack/kolla-ansible master: Bump Sphinx lower constraints to 1.8.0 https://review.opendev.org/690782 | 02:24 |
---|---|---|
openstackgerrit | Merged openstack/kolla-ansible master: Remove delegate_to from service-ks-register tasks https://review.opendev.org/685008 | 03:00 |
*** dannins has joined #openstack-kolla | 03:16 | |
*** shyamb has joined #openstack-kolla | 03:59 | |
*** shyamb has quit IRC | 04:19 | |
*** radeks has joined #openstack-kolla | 04:21 | |
*** shyamb has joined #openstack-kolla | 04:42 | |
*** shyamb has quit IRC | 05:39 | |
*** shyamb has joined #openstack-kolla | 05:46 | |
*** cah_link has joined #openstack-kolla | 05:51 | |
*** dpawlik has joined #openstack-kolla | 05:59 | |
*** dpawlik has quit IRC | 06:03 | |
*** radeks has quit IRC | 06:07 | |
*** dr_feelgood has joined #openstack-kolla | 06:16 | |
*** dpawlik has joined #openstack-kolla | 06:23 | |
openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible master: WIP: Ceph-Ansible CI https://review.opendev.org/676376 | 06:32 |
*** pcaruana has joined #openstack-kolla | 06:41 | |
yoctozepto | morning | 06:43 |
mnasiadka | morning | 06:43 |
hrw | kawusia++ | 06:57 |
mnasiadka | tripleo rc packages still didn't make to mirror.centos.org... grr | 07:01 |
*** jaicaa has quit IRC | 07:02 | |
*** jaicaa has joined #openstack-kolla | 07:05 | |
yoctozepto | mnasiadka: are they going to be mirrored? | 07:09 |
yoctozepto | maybe you are waiting for something that is not going to happen | 07:09 |
mnasiadka | yoctozepto: RDO guys promised they would be there yesterday in the evening, let's give them some more time. | 07:09 |
*** jbadiapa has joined #openstack-kolla | 07:13 | |
openstackgerrit | Radosław Piliszek proposed openstack/kolla-ansible master: Default to disabling haproxy for memcached https://review.opendev.org/690869 | 07:13 |
yoctozepto | mnasiadka: what mirror are we talking about? RDO's or infra's? | 07:13 |
mnasiadka | yoctozepto: first it needs to be on mirror.centos.org, to be on infra's :D | 07:14 |
yoctozepto | mnasiadka: all right, I thought you were about the infra | 07:14 |
openstackgerrit | Radosław Piliszek proposed openstack/kolla-ansible master: Bump Sphinx lower constraints to 1.8.0 https://review.opendev.org/690782 | 07:16 |
openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible master: Add proper wsgi loglevel when openstack_logging_debug https://review.opendev.org/690871 | 07:30 |
*** gfidente has joined #openstack-kolla | 07:34 | |
openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible master: WIP: Ceph-Ansible CI https://review.opendev.org/676376 | 07:42 |
*** tonythomas has joined #openstack-kolla | 07:47 | |
*** rpittau|afk is now known as rpittau | 07:52 | |
*** rgogunskiy has joined #openstack-kolla | 07:54 | |
*** shyamb has quit IRC | 08:02 | |
*** amoralej|off is now known as amoralej | 08:04 | |
*** k_mouza has joined #openstack-kolla | 08:13 | |
mgoddard | morning all | 08:18 |
mgoddard | btw yoctozepto, mnasiadka, hrw: I will be out next Monday - Wednesday (inclusive) | 08:18 |
hrw | ok | 08:18 |
hrw | mnasiadka: asked boss? | 08:18 |
mnasiadka | hrw: in the process, awaiting answer ;) | 08:19 |
mnasiadka | mgoddard: have fun | 08:19 |
hrw | to be honest I did not asked my wife yet | 08:19 |
*** skramaja has joined #openstack-kolla | 08:19 | |
hrw | but it was never an issue | 08:19 |
hrw | bb in ~30 minutes | 08:20 |
mnasiadka | my wife only complains behind back and reminds about it for the next 30 years ;) | 08:22 |
yoctozepto | so hrw needs 30 minutes to ask wifey | 08:35 |
*** k_mouza_ has joined #openstack-kolla | 08:42 | |
openstackgerrit | Merged openstack/kolla-ansible master: Zun: fix Cinder (volume) iSCSI support https://review.opendev.org/690614 | 08:42 |
*** k_mouza has quit IRC | 08:43 | |
openstackgerrit | Mark Goddard proposed openstack/kolla-ansible master: Fix swift.conf custom configuration path collision https://review.opendev.org/690881 | 08:44 |
*** k_mouza_ has quit IRC | 08:48 | |
hrw | yoctozepto: no, it was other reason to be out | 08:48 |
yoctozepto | hrw: just kidding :-) | 08:49 |
*** k_mouza has joined #openstack-kolla | 08:49 | |
hrw | manylinux2014 is coming. pypi stuff | 08:51 |
yoctozepto | hrw: I guess will support aarch64 | 08:53 |
*** chrizl has joined #openstack-kolla | 08:53 | |
hrw | aarch64, ppc64le, armv7l etc | 08:53 |
hrw | s390x | 08:53 |
hrw | https://github.com/pypa/manylinux/pull/348 | 08:54 |
yoctozepto | hrw: glad you have that under control | 09:01 |
openstackgerrit | Mark Goddard proposed openstack/kolla-ansible master: CI: Add mariadb test https://review.opendev.org/655663 | 09:10 |
yoctozepto | ERROR:kolla.common.utils.barbican-base:error creating overlay mount to /var/lib/docker/overlay2/2bb23bb31aff64351da2932a683c0b72529f6e72c31492e2298f8f14b5837fe4-init/merged: device or resource busy | 09:26 |
yoctozepto | 'tis lame | 09:26 |
openstackgerrit | Numan Siddique proposed openstack/kolla master: ovn: Create the ovn dbs only if the schema file exists https://review.opendev.org/690065 | 09:28 |
openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible master: Add proper wsgi loglevel when openstack_logging_debug https://review.opendev.org/690871 | 09:28 |
openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible master: Add proper wsgi loglevel when openstack_logging_debug https://review.opendev.org/690871 | 09:33 |
openstackgerrit | Merged openstack/kolla-ansible master: Bump Sphinx lower constraints to 1.8.0 https://review.opendev.org/690782 | 09:33 |
*** shyamb has joined #openstack-kolla | 09:35 | |
openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible master: WIP: Ceph-Ansible CI https://review.opendev.org/676376 | 09:47 |
*** skramaja has quit IRC | 09:50 | |
openstackgerrit | Mark Goddard proposed openstack/kolla-ansible master: CI: Add mariadb test https://review.opendev.org/655663 | 09:55 |
* hrw out | 10:09 | |
yoctozepto | we have three outputs from timeouts | 10:15 |
yoctozepto | 2 result in 504 | 10:15 |
yoctozepto | 1 results in 500 | 10:15 |
yoctozepto | if service knows it timeouts (due to rabbitmq/mysql/other service) - it shouts 500 | 10:15 |
yoctozepto | if wsgi timeouts we get 504 from apache (though the timeout seems to be 5 minutes, still longer than haproxy's) | 10:16 |
yoctozepto | if apache timeouts (which timeouts due to wsgi above), we get 504 from haproxy | 10:16 |
*** dpawlik has quit IRC | 10:19 | |
mgoddard | yoctozepto: is 500 actually a timeout from haproxy? or just from the backend | 10:20 |
yoctozepto | 500 only if service is quick to discover timeout from its dependency | 10:21 |
mgoddard | ok | 10:21 |
yoctozepto | then it goes back to the user | 10:21 |
yoctozepto | all other cases 504 | 10:21 |
yoctozepto | (which are actually 2) | 10:21 |
mgoddard | does wsgi not detect that the client (haproxy) has disconnected? | 10:21 |
yoctozepto | keystone-wsgi is a real hanger | 10:21 |
mgoddard | possibly not | 10:21 |
mgoddard | yeah, I've seen keystone-wsgi hang | 10:21 |
yoctozepto | nope, it continues | 10:21 |
mgoddard | guess that makes sense | 10:22 |
yoctozepto | which increase the future failover rate | 10:22 |
mgoddard | I saw a case where one of the keystone backends would always hang | 10:22 |
yoctozepto | I wonder if I fix the keepalived stability | 10:22 |
mgoddard | and because every operation requires 3 calls to keystone, it always failed | 10:22 |
yoctozepto | to let VIP addr stay on one node | 10:22 |
yoctozepto | yeah, this surely makes it hard | 10:22 |
mgoddard | can we get more info from keystone on what it's doing? | 10:23 |
yoctozepto | we already have it in debug | 10:23 |
yoctozepto | maybe mnasiadka's change give more from wsgi | 10:23 |
mgoddard | strace? | 10:23 |
mgoddard | will be a big log :) | 10:23 |
yoctozepto | yeah xD | 10:23 |
*** rgogunskiy has quit IRC | 10:23 | |
yoctozepto | for the next run I take mnasiadka's extra logging + keepalived fix if I make it happen | 10:26 |
yoctozepto | then either it fixes the issue or we get some more info about it | 10:26 |
*** shyamb has quit IRC | 10:27 | |
yoctozepto | (or nothing helps and we are bound to strace) | 10:27 |
mgoddard | what do you think is the problem with keepalived? | 10:29 |
*** rgogunskiy has joined #openstack-kolla | 10:31 | |
*** shyamb has joined #openstack-kolla | 10:37 | |
*** shyamb has quit IRC | 10:43 | |
*** shyamb has joined #openstack-kolla | 10:43 | |
*** dpawlik has joined #openstack-kolla | 10:44 | |
openstackgerrit | Radosław Piliszek proposed openstack/kolla-ansible master: DNM: increase keepalived timeout https://review.opendev.org/690920 | 10:45 |
yoctozepto | mgoddard: keepalived queries haproxy every 2 seconds | 10:45 |
yoctozepto | waiting 1 second for reply on the socket | 10:45 |
yoctozepto | normally haproxy responds in 1/10th of a sec | 10:46 |
yoctozepto | or less | 10:46 |
yoctozepto | but here it seems not to | 10:46 |
yoctozepto | I increased the timeout to 3 seconds | 10:46 |
yoctozepto | and had to increase interval to 3 seconds as well but I think it's ok for production either, it only extends discovery of failed haproxy by 2 seconds | 10:47 |
yoctozepto | due to 2 tries | 10:47 |
yoctozepto | before announcing the failure | 10:47 |
yoctozepto | haproxy socket down would result in 1, not timeout kill | 10:48 |
yoctozepto | looks like haproxy is busy | 10:48 |
yoctozepto | I seriously wonder why keystone discovered it has b0rken mysql connection 20 minutes after vip change | 10:50 |
yoctozepto | and whether that caused its problems | 10:51 |
yoctozepto | it was the first request after vip change after all | 10:52 |
*** jbadiapa has quit IRC | 10:55 | |
*** shyam89 has joined #openstack-kolla | 10:58 | |
*** shyamb has quit IRC | 10:59 | |
*** amoralej is now known as amoralej|lunch | 11:00 | |
yoctozepto | the other two keystone instances discovered the issue pretty immediately and reconnected | 11:04 |
yoctozepto | well, they just sent immediate 500 and then were working | 11:05 |
yoctozepto | also, wsrep issues pop out after vip addr change, wonder what that means | 11:07 |
*** radeks has joined #openstack-kolla | 11:07 | |
*** shyamb has joined #openstack-kolla | 11:08 | |
openstackgerrit | Mark Goddard proposed openstack/kolla-ansible master: Tidy up release notes for Train https://review.opendev.org/688973 | 11:08 |
openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible master: WIP: Ceph-Ansible CI https://review.opendev.org/676376 | 11:10 |
*** shyam89 has quit IRC | 11:11 | |
openstackgerrit | Merged openstack/kolla-ansible master: Limit open file descriptors for Neutron agent containers https://review.opendev.org/689548 | 11:16 |
*** shyam89 has joined #openstack-kolla | 11:17 | |
mnasiadka | yoctozepto: maybe it's a docker mtu issue? | 11:17 |
mnasiadka | I know we use --network=host, just speculating... | 11:18 |
*** X-dark has left #openstack-kolla | 11:19 | |
*** shyamb has quit IRC | 11:19 | |
*** k_mouza_ has joined #openstack-kolla | 11:23 | |
*** kplant has joined #openstack-kolla | 11:23 | |
*** shyam89 has quit IRC | 11:24 | |
*** radeks_ has joined #openstack-kolla | 11:24 | |
*** k_mouza has quit IRC | 11:25 | |
*** radeks has quit IRC | 11:26 | |
*** skramaja has joined #openstack-kolla | 11:27 | |
*** jbadiapa has joined #openstack-kolla | 11:29 | |
*** shyamb has joined #openstack-kolla | 11:30 | |
*** jbadiapa has quit IRC | 11:32 | |
yoctozepto | mnasiadka: unlikely, remember we already had 504 before, only adding keepalived+haproxy resulted in more of them | 11:34 |
mnasiadka | before vxlan? | 11:34 |
yoctozepto | as we have it in all jobs now | 11:34 |
yoctozepto | yeah | 11:34 |
*** k_mouza_ has quit IRC | 11:34 | |
mnasiadka | maybe I was blind for them :) | 11:34 |
yoctozepto | maybe, but they were rarer | 11:35 |
*** k_mouza has joined #openstack-kolla | 11:35 | |
yoctozepto | singlenode cannot have a failover in keepalived | 11:35 |
yoctozepto | so this failure vector was out of scope | 11:35 |
yoctozepto | mgoddard: how strong your RP-1 is for https://review.opendev.org/690869 - I strongly advise we get rid of this quirk in Train already ;p | 11:37 |
*** jbadiapa has joined #openstack-kolla | 11:41 | |
mnasiadka | yoctozepto: well, unless it breaks something and there's a bug filed... | 11:43 |
*** janki has joined #openstack-kolla | 11:44 | |
yoctozepto | mnasiadka: you are not helping :-) | 11:44 |
mnasiadka | yoctozepto: policy is policy | 11:44 |
mnasiadka | we only do bugs now | 11:44 |
mnasiadka | (and RDO) | 11:44 |
mnasiadka | :) | 11:44 |
yoctozepto | eh, whatever ;p | 11:45 |
yoctozepto | Thu Oct 24 10:25:54 2019: /check_alive.sh exited due to signal 15 | 11:45 |
yoctozepto | Thu Oct 24 10:25:54 2019: VRRP_Instance(kolla_internal_vip_98) Transition to MASTER STATE | 11:45 |
yoctozepto | Thu Oct 24 10:25:54 2019: VRRP_Instance(kolla_internal_vip_98) Entering MASTER STATE | 11:45 |
yoctozepto | Thu Oct 24 10:25:54 2019: VRRP_Instance(kolla_internal_vip_98) using locally configured advertisement interval (1000 milli-sec) | 11:45 |
yoctozepto | Thu Oct 24 10:26:03 2019: VRRP_Instance(kolla_internal_vip_98) Received advert with lower priority 2, ours 3, forcing new election | 11:45 |
yoctozepto | Thu Oct 24 10:26:03 2019: VRRP_Instance(kolla_internal_vip_98) Received advert with lower priority 2, ours 3, forcing new election | 11:46 |
yoctozepto | Thu Oct 24 10:26:03 2019: VRRP_Instance(kolla_internal_vip_98) Received advert with lower priority 2, ours 3, forcing new election | 11:46 |
yoctozepto | this bug does not help either (just like mnasiadka) | 11:46 |
hrw | I have patch waiting for centos8 support replacing centos7 | 11:46 |
yoctozepto | centos8? :O | 11:46 |
mnasiadka | hrw: there is no RDO for CentOS 8 | 11:46 |
mnasiadka | and there won't be for the next 3-4 weeks | 11:46 |
yoctozepto | don't tell me we are going to hold on 3 months | 11:46 |
hrw | that's why I did not even sent it for review ;D | 11:46 |
hrw | I have a copy of kolla which probably do not even work with py2 | 11:47 |
yoctozepto | good, py2 is to be dropped in ussuri | 11:47 |
yoctozepto | drop drop drop | 11:47 |
mnasiadka | yoctozepto: hrw wanted to tell you to wait until U to merge this change ;) | 11:47 |
hrw | ImportError: No module named queue | 11:51 |
hrw | yes! | 11:51 |
hrw | mission accomplished | 11:51 |
yoctozepto | xD | 11:54 |
yoctozepto | mnasiadka: ok, ok, I'm giving up | 11:55 |
*** dave-mccowan has joined #openstack-kolla | 11:55 | |
*** shyamb has quit IRC | 11:59 | |
*** dave-mccowan has quit IRC | 12:00 | |
openstackgerrit | Radosław Piliszek proposed openstack/kolla-ansible master: DNM: haproxy timeout - raise value for server https://review.opendev.org/690702 | 12:01 |
*** amoralej|lunch is now known as amoralej | 12:01 | |
* hrw -> meeting | 12:01 | |
yoctozepto | with wifey | 12:01 |
*** radeks_ has quit IRC | 12:13 | |
openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible master: WIP: Ceph-Ansible CI https://review.opendev.org/676376 | 12:14 |
hrw | yoctozepto: company one this time | 12:17 |
hrw | yoctozepto: but if you need psychotherapy then I can arrange you a meeting with my wife | 12:17 |
hrw | ;D | 12:19 |
yoctozepto | you mean legit psychotherapy? | 12:19 |
mnasiadka | hrw: who doesn't need it :D | 12:19 |
yoctozepto | or psychoshock | 12:19 |
yoctozepto | :D | 12:19 |
hrw | yoctozepto: legit. certified etc. | 12:20 |
mgoddard | hrw, yoctozepto: centos8 & source testing can begin without RDO | 12:22 |
mgoddard | actually that was mnasiadka | 12:22 |
mnasiadka | mgoddard: well, sure - but we'll soon get to a point we need some packages from RDO | 12:23 |
mgoddard | maybs | 12:23 |
yoctozepto | hrw: wow, nice! :D | 12:23 |
mnasiadka | so in T we will have both centos7 and centos8 images? :) | 12:23 |
mgoddard | in T centos7 only | 12:24 |
mnasiadka | ok, so we go for c8 in U | 12:24 |
mgoddard | yeah | 12:24 |
mnasiadka | makes sense | 12:24 |
yoctozepto | re centos8 - I already wanted centos8 for source but we want consistency too much :-) | 12:24 |
yoctozepto | it needs work for sure | 12:24 |
yoctozepto | but hrw can publish that patch | 12:24 |
mnasiadka | yoctozepto: even in source we rely on RDO packages, e.g. openvswitch | 12:24 |
mgoddard | we discussed backporting fixes to T after RDO release, if they are limited | 12:25 |
yoctozepto | mnasiadka: ah, yeah, that is a core component, argh | 12:25 |
hrw | c8 for U+. if someone will want to backport then feel free, will not review | 12:25 |
mnasiadka | and the centos8 build system is still not available for SIGs (like Cloud SIG) ;) | 12:25 |
yoctozepto | start with linuxbridge | 12:25 |
yoctozepto | yeah, U+ definitely | 12:26 |
hrw | 12:01 (0s) hrw@j13-qrep-04:~$ docker run -it centos:8 | 12:26 |
hrw | Unable to find image 'centos:8' locally | 12:26 |
hrw | 8: Pulling from library/centos | 12:26 |
hrw | docker: no matching manifest for linux/arm64/v8 in the manifest list entries. | 12:26 |
hrw | so c8 work for me then yet ;D | 12:27 |
mnasiadka | uh, arm64 will come... in future :D | 12:28 |
*** schwicht has joined #openstack-kolla | 12:30 | |
hrw | mnasiadka: my x86-64 boxes are 4-8 cores with 20-32GB ram. arm64 one has 46 cores and 96GB ram so I do quick builds in tmpfs | 12:30 |
mnasiadka | well, that's understandable :) | 12:31 |
mnasiadka | go fix centos on arm64 then ;) | 12:31 |
hrw | had to give away fast machines... | 12:31 |
hrw | s/away/back | 12:31 |
yoctozepto | is arm64 supported upstream? in rhel8 I mean | 12:33 |
hrw | yoctozepto: yes | 12:34 |
hrw | yoctozepto: rhel8 has four equal architectures: aarch64, ppc64le, s390x, x86-64 | 12:34 |
hrw | package has to build on each | 12:34 |
hrw | unless does not make sense like dmidecode which is aarch64 and x86-64 only etc | 12:35 |
openstackgerrit | Bharat Kunwar proposed openstack/kayobe master: Example should use hyphen instead of underscore https://review.opendev.org/690949 | 12:35 |
openstackgerrit | Bharat Kunwar proposed openstack/kayobe master: Example should use hyphen instead of underscore https://review.opendev.org/690949 | 12:36 |
yoctozepto | hrw: aye, thx | 12:36 |
openstackgerrit | Bharat Kunwar proposed openstack/kayobe master: Example should use hyphen instead of underscore https://review.opendev.org/690949 | 12:37 |
hrw | yoctozepto: centos will add armv7l and i686 probably as altarch but both can be ignored | 12:37 |
hrw | i686 means old junk people are too sentimental to throw away, armv7l means toys-like hardware (same sentimental stuff) | 12:38 |
mnasiadka | hrw: I think there are some problems with kolla builds on Debian... | 12:41 |
mnasiadka | https://zuul.opendev.org/t/openstack/build/0101b0e167a74fbb8146ae6beda27987/log/primary/logs/build/000_FAILED_placement-base.log#6 | 12:41 |
hrw | fucking docker... | 12:41 |
hrw | it is problem with docker on anything with fresh enough kernel | 12:42 |
mnasiadka | add support for podman and buildah :) | 12:43 |
hrw | mnasiadka: one condition: you do podman in k-a and changes in kolla to get k-a working | 12:44 |
hrw | mnasiadka: as podman runs on user account so GID/UID stuff can be complicated | 12:44 |
hrw | from what I heard | 12:44 |
yoctozepto | hrw: we will run it as root then ;p | 12:44 |
mnasiadka | and we would need systemd units for each container... | 12:44 |
yoctozepto | yeah, for persistency | 12:45 |
openstackgerrit | Merged openstack/kolla master: Fix mariabackup arguments https://review.opendev.org/690589 | 12:45 |
openstackgerrit | Merged openstack/kolla master: ovn: Create the ovn dbs only if the schema file exists https://review.opendev.org/690065 | 12:45 |
mnasiadka | hrw: well, we could start at least building images using buildah | 12:45 |
openstackgerrit | Merged openstack/kolla-ansible master: Tidy up release notes for Train https://review.opendev.org/688973 | 12:45 |
hrw | yoctozepto: ;D | 12:45 |
yoctozepto | we surely need to start toying with that | 12:45 |
yoctozepto | as it looks like the way forward | 12:45 |
hrw | works for tripleo guys with our templates | 12:45 |
mnasiadka | everything looks like a way forward until you use it extensively | 12:46 |
yoctozepto | fair point | 12:46 |
mnasiadka | hrw: yeah well, we would need to rework image build python code, to support both docker and podman ;) | 12:46 |
* hrw -> food | 12:47 | |
*** schwicht has quit IRC | 12:50 | |
mnasiadka | hrw: well, they thought they fixed it - but they didn't: https://github.com/docker/for-linux/issues/711 | 12:52 |
*** born2bake has joined #openstack-kolla | 12:56 | |
born2bake | Is anyone had the same issue - https://bugs.launchpad.net/kolla-ansible/+bug/1847905 ? | 12:56 |
openstack | Launchpad bug 1847905 in kolla-ansible "Octavia loadbalancer pending create state" [Undecided,New] | 12:56 |
*** goldyfruit has joined #openstack-kolla | 13:00 | |
yoctozepto | mnasiadka: commented there | 13:05 |
openstackgerrit | Mark Goddard proposed openstack/kayobe master: Use docker_custom_config variable https://review.opendev.org/690965 | 13:07 |
openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible master: WIP: Ceph-Ansible CI https://review.opendev.org/676376 | 13:13 |
*** janki has quit IRC | 13:13 | |
mnasiadka | born2bake: it seems like an Octavia code issue - have you raised a bug in Octavia on Storyboard? | 13:16 |
openstackgerrit | Merged openstack/kayobe master: Add nova-super-conductor to inventory https://review.opendev.org/688725 | 13:23 |
yoctozepto | no keepalived instability, no other issues at hand, still hang of keystone admin | 13:27 |
yoctozepto | https://a30336fa6a8fca5c6dba-fe779e5654b21fdff79727b204dfb7d6.ssl.cf1.rackcdn.com/690920/1/check/kolla-ansible-centos-source-ceph/8ef81a0/ | 13:27 |
yoctozepto | mariadb connections were broken here though as well | 13:28 |
yoctozepto | could be the results of rebootstraps happening in reconfigure | 13:28 |
*** spsurya has quit IRC | 13:29 | |
born2bake | mnasiadka no, is it octavia issue? i thought its related to kolla | 13:44 |
mnasiadka | born2bake: it is, read Dincer’s comment | 13:44 |
mnasiadka | What he is doing is a workaround | 13:45 |
born2bake | I am not confident enough to customize kolla images lol | 13:46 |
born2bake | also, I do have similar problem with magnum...cluster is just in pending create state... https://bugs.launchpad.net/kolla-ansible/+bug/1846476 | 13:46 |
openstack | Launchpad bug 1846476 in kolla-ansible "K8s cluster deployment is stuck at kube_cluster_deploy OS::Heat::SoftwareDeployment" [Undecided,New] | 13:46 |
born2bake | @mar | 13:47 |
born2bake | markmcclain in case they fix the isses with octavia code, how do I update my images so the bug will be fixed? | 13:47 |
*** born2bake has quit IRC | 13:49 | |
*** amoralej is now known as amoralej|mtg | 14:00 | |
openstackgerrit | Mark Goddard proposed openstack/kolla-ansible master: Restart nova scheduler to refresh cell cache https://review.opendev.org/690982 | 14:03 |
*** learnstack has joined #openstack-kolla | 14:05 | |
openstackgerrit | Radosław Piliszek proposed openstack/kolla-ansible master: Allow passing arguments to RabbitMQ server https://review.opendev.org/686369 | 14:06 |
*** cah_link has quit IRC | 14:07 | |
*** dpawlik has quit IRC | 14:11 | |
*** pcaruana has quit IRC | 14:19 | |
mnasiadka | yoctozepto: so now what, we need to triple check mariadb state? | 14:39 |
*** learnstack has quit IRC | 14:48 | |
*** evrardjp_ is now known as evrardjp | 14:48 | |
*** amoralej|mtg is now known as amoralej | 14:49 | |
openstackgerrit | Mark Goddard proposed openstack/kayobe master: Use docker_custom_config variable https://review.opendev.org/690965 | 14:51 |
*** learnstack has joined #openstack-kolla | 14:59 | |
*** gouthamr_ is now known as gouthamr | 15:01 | |
*** learnstack has quit IRC | 15:06 | |
*** born2bake has joined #openstack-kolla | 15:09 | |
*** priteau has joined #openstack-kolla | 15:09 | |
openstackgerrit | Will Szumski proposed openstack/kolla-ansible stable/stein: Apply Resource Constraints to Services. https://review.opendev.org/691000 | 15:10 |
*** rgogunskiy has quit IRC | 15:10 | |
*** gedcib has joined #openstack-kolla | 15:16 | |
openstackgerrit | Mark Goddard proposed openstack/kolla-ansible master: Fix kolla-bootstrap-servers with docker_storage_driver set https://review.opendev.org/691001 | 15:17 |
openstackgerrit | Mark Goddard proposed openstack/kayobe master: Use docker_custom_config variable https://review.opendev.org/690965 | 15:18 |
openstackgerrit | Mark Goddard proposed openstack/kolla-ansible master: Use mariabackup for database backups https://review.opendev.org/690598 | 15:20 |
openstackgerrit | Mark Goddard proposed openstack/kolla-ansible master: CI: Add mariadb test https://review.opendev.org/655663 | 15:21 |
gedcib | hi guys, when i add custom keystone.conf for one of the domains and run kolla-reconfigure it fails with this error: | 15:23 |
gedcib | TASK [keystone : Creating admin project, user, role, service, and endpoint] ******************************************************************************************************************* | 15:23 |
gedcib | fatal: [nested]: FAILED! => {"msg": "The conditional check '(keystone_bootstrap.stdout | from_json).changed' failed. The error was: Invalid control character at: line 1 column 134 (char 133)"} | 15:23 |
openstackgerrit | Mark Goddard proposed openstack/kolla-ansible master: Use mariabackup for database backups https://review.opendev.org/690598 | 15:28 |
openstackgerrit | Mark Goddard proposed openstack/kolla-ansible master: CI: Add mariadb test https://review.opendev.org/655663 | 15:28 |
openstackgerrit | Will Szumski proposed openstack/kolla-ansible stable/stein: Apply Resource Constraints to Services. https://review.opendev.org/691000 | 15:29 |
openstackgerrit | Mark Goddard proposed openstack/kolla-ansible master: CI: Add mariadb test https://review.opendev.org/655663 | 15:31 |
openstackgerrit | Merged openstack/kolla-ansible master: Add proper wsgi loglevel when openstack_logging_debug https://review.opendev.org/690871 | 15:34 |
openstackgerrit | Will Szumski proposed openstack/kolla-ansible stable/stein: Limit open file descriptors for Neutron agent containers https://review.opendev.org/691007 | 15:39 |
*** gedcib has quit IRC | 15:40 | |
*** goldyfruit has quit IRC | 15:54 | |
*** goldyfruit has joined #openstack-kolla | 15:56 | |
*** k_mouza_ has joined #openstack-kolla | 16:02 | |
*** rpittau is now known as rpittau|afk | 16:05 | |
*** k_mouza has quit IRC | 16:06 | |
openstackgerrit | Adeberg proposed openstack/kolla-ansible master: Update swift templates to support swift_extra_ring_files https://review.opendev.org/691014 | 16:32 |
*** igordc has joined #openstack-kolla | 16:39 | |
*** k_mouza_ has quit IRC | 16:45 | |
openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible master: WIP: Ceph-Ansible CI https://review.opendev.org/676376 | 16:47 |
openstackgerrit | Mark Goddard proposed openstack/kolla-ansible master: Use mariabackup for database backups https://review.opendev.org/690598 | 16:49 |
openstackgerrit | Mark Goddard proposed openstack/kolla-ansible master: CI: Add mariadb test https://review.opendev.org/655663 | 16:49 |
*** spiette has quit IRC | 16:52 | |
*** spiette has joined #openstack-kolla | 16:55 | |
*** priteau has quit IRC | 16:59 | |
*** igordc has quit IRC | 17:01 | |
*** igordc has joined #openstack-kolla | 17:02 | |
mnasiadka | hrw: https://review.opendev.org/#/c/689121/ - are we merging it? | 17:04 |
hrw | mnasiadka: I would not write it otherwise | 17:04 |
mnasiadka | hrw: rpm_setup_config has the same issue, or does it work? :) | 17:05 |
hrw | mnasiadka: should just work as it is handled other way | 17:05 |
mnasiadka | so +2+w and we're off | 17:05 |
hrw | thanks | 17:05 |
*** k_mouza has joined #openstack-kolla | 17:06 | |
* hrw off | 17:06 | |
mnasiadka | this was the last bug in kolla attached to Train milestone | 17:06 |
mnasiadka | (last not fixed) | 17:06 |
hrw | mnasiadka: rpm_setup_config is handled in docker/base/Dockerfile.j2 L141-143 | 17:08 |
hrw | as kolla/image/build.py builds command which is then run | 17:08 |
*** k_mouza has quit IRC | 17:10 | |
kplant | is there a recommended way to safely reboot control nodes? | 17:13 |
kplant | i need to restart the docker daemons to update trusted CAs :-| | 17:14 |
*** stingrayza has quit IRC | 17:15 | |
noxoid | idk if its recommended but i just restart the docker daemons one at a time to let any quorum services (ceph mon/mariadb/rabbitmq) catch up | 17:18 |
noxoid | assuming docker doesnt lose track of the runc processes and orphans them | 17:19 |
kplant | i've done that before and things kind of fell down | 17:19 |
kplant | rabbitmq shit itself | 17:19 |
kplant | thankfully i don't have ceph on the os controllers | 17:20 |
noxoid | are you using autoheal or pause_minority? | 17:20 |
kplant | whichever the default kolla-ansible is configuring for rmq | 17:21 |
noxoid | well it changed recently :P | 17:22 |
kplant | oh | 17:23 |
kplant | let me check with cluster_status | 17:23 |
noxoid | https://github.com/openstack/kolla-ansible/commit/cdfc1c23442f43b6d86010cbd5147802c5214934 | 17:23 |
noxoid | in stein its still autoheal | 17:23 |
kplant | yeah - it's autoheal | 17:24 |
noxoid | you may have better luck by switching it to pause_minority | 17:24 |
kplant | appreciate it | 17:24 |
noxoid | but even when i was on autoheal it could still handle rolling restarts | 17:24 |
noxoid | or at least i didnt notice any problems | 17:24 |
kplant | maybe i didn't wait long enough between booting nodes? | 17:24 |
noxoid | its possible | 17:24 |
noxoid | i usually wait 3-4 minutes between each | 17:24 |
kplant | i couldn't get rabbit to recover until i booted the entire physical server | 17:24 |
*** henriqueof has joined #openstack-kolla | 17:25 | |
noxoid | id run cluster_status as well as node_health_check before moving on to the next node | 17:25 |
noxoid | ive seen cases where node_health_check will report issues but cluster_status would not | 17:25 |
kplant | appreciate the insight | 17:25 |
noxoid | no problem. rabbitmq is the only component of openstack that im still having unknown issues with | 17:25 |
*** stingrayza has joined #openstack-kolla | 17:25 | |
kplant | i'm going to start with the machine that doesn't have the vip to be safe | 17:26 |
kplant | if it recovers and there are no partitions i'll keep rolling | 17:27 |
noxoid | sounds good | 17:27 |
kplant | it's immediately healthy - that's skeptical | 17:32 |
kplant | nothing works that well | 17:32 |
*** stingrayza has quit IRC | 17:48 | |
*** stingrayza has joined #openstack-kolla | 17:49 | |
*** amoralej is now known as amoralej|off | 17:52 | |
kplant | well all three nodes are healthy | 17:54 |
*** stingrayza has quit IRC | 17:54 | |
kplant | we'll see if anything blows up :] | 17:54 |
*** kplant has quit IRC | 17:55 | |
*** stingrayza has joined #openstack-kolla | 18:02 | |
*** schwicht has joined #openstack-kolla | 18:13 | |
openstackgerrit | Merged openstack/kolla master: get 'apt_*' options of kolla-build.conf working again https://review.opendev.org/689121 | 18:58 |
*** born2bake has quit IRC | 19:34 | |
*** goldyfruit_ has joined #openstack-kolla | 19:57 | |
*** goldyfruit has quit IRC | 19:59 | |
*** KeithMnemonic has quit IRC | 20:41 | |
henriqueof | Installed a let's encrypt using kolla-ansible, the certificate expired just now, how can I change it? | 20:41 |
*** KeithMnemonic has joined #openstack-kolla | 20:52 | |
*** goldyfruit_ has quit IRC | 21:38 | |
*** schwicht has quit IRC | 21:43 | |
*** goldyfruit has joined #openstack-kolla | 22:03 | |
*** gfidente has quit IRC | 22:06 | |
*** igordc has quit IRC | 22:08 | |
*** schwicht has joined #openstack-kolla | 22:29 | |
*** schwicht has quit IRC | 22:36 | |
*** schwicht has joined #openstack-kolla | 22:43 | |
*** k_mouza has joined #openstack-kolla | 22:46 | |
*** igordc has joined #openstack-kolla | 22:47 | |
*** schwicht has quit IRC | 22:51 | |
*** k_mouza has quit IRC | 22:51 | |
*** zbr has quit IRC | 23:13 | |
*** henriqueof has quit IRC | 23:33 | |
*** schwicht has joined #openstack-kolla | 23:34 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!