yoctozepto | frickler: 22.04 will bring us podman :-) | 06:41 |
---|---|---|
*** amoralej|off is now known as amoralej | 06:47 | |
mnasiadka | good morning | 06:53 |
mnasiadka | yoctozepto: what's wrong with the one from kubic (apart not being in distro packages)? :) | 06:54 |
yoctozepto | mnasiadka: for me? nothing; it's frickler who is asking | 07:07 |
mnasiadka | as long as it's 3.0+ it should be ok | 07:07 |
mnasiadka | 2.1/2.2 was a hell hole | 07:07 |
opendevreview | Merged openstack/kolla stable/xena: Update .gitreview for stable/xena https://review.opendev.org/c/openstack/kolla/+/813995 | 07:29 |
opendevreview | Merged openstack/kolla stable/xena: Update TOX_CONSTRAINTS_FILE for stable/xena https://review.opendev.org/c/openstack/kolla/+/813996 | 07:29 |
opendevreview | Pierre Riteau proposed openstack/kayobe master: Add support for configuring proxy settings https://review.opendev.org/c/openstack/kayobe/+/814046 | 08:08 |
opendevreview | Merged openstack/kolla-ansible master: Add Python3 yoga unit tests https://review.opendev.org/c/openstack/kolla-ansible/+/813994 | 08:30 |
opendevreview | Pierre Riteau proposed openstack/kayobe master: Add support for configuring proxy settings https://review.opendev.org/c/openstack/kayobe/+/814046 | 08:50 |
opendevreview | Michal Nasiadka proposed openstack/kayobe master: [release] Sync Kolla Ansible feature flags and inventory for Xena https://review.opendev.org/c/openstack/kayobe/+/814118 | 08:53 |
opendevreview | Michal Nasiadka proposed openstack/kayobe master: [release] Sync Kolla Ansible feature flags and inventory for Xena https://review.opendev.org/c/openstack/kayobe/+/814120 | 09:02 |
opendevreview | Michal Nasiadka proposed openstack/kayobe master: [release] Sync Kolla Ansible feature flags for Xena https://review.opendev.org/c/openstack/kayobe/+/814118 | 09:03 |
opendevreview | Michal Nasiadka proposed openstack/kayobe master: [release] Sync Kolla Ansible inventory for Xena https://review.opendev.org/c/openstack/kayobe/+/814120 | 09:03 |
opendevreview | Michal Nasiadka proposed openstack/kayobe master: Use OpenStack Xena release https://review.opendev.org/c/openstack/kayobe/+/814122 | 09:08 |
opendevreview | Maksim Malchuk proposed openstack/kayobe master: Add support for Ironic inspection through DHCP-relay https://review.opendev.org/c/openstack/kayobe/+/813266 | 09:26 |
opendevreview | Maksim Malchuk proposed openstack/kayobe master: Set defaults for openstack_cacert https://review.opendev.org/c/openstack/kayobe/+/793703 | 09:33 |
opendevreview | Maksim Malchuk proposed openstack/kayobe master: TLS certificates management sync with Kolla-Ansible https://review.opendev.org/c/openstack/kayobe/+/793697 | 09:33 |
opendevreview | Michal Nasiadka proposed openstack/kayobe master: [release] Sync Kolla Ansible inventory for Xena https://review.opendev.org/c/openstack/kayobe/+/814120 | 09:33 |
dardelean | hello all | 09:41 |
opendevreview | Michal Nasiadka proposed openstack/kayobe master: [release] Sync Kolla Ansible inventory for Xena https://review.opendev.org/c/openstack/kayobe/+/814120 | 09:42 |
dardelean | I want to deploy Wallaby release. If I want to make Glance HA I can use a file backend and use NFS, as of described here https://docs.openstack.org/kolla-ansible/wallaby/reference/shared-services/glance-guide.html#file-backend | 09:42 |
dardelean | "For better reliability and performance, glance_file_datadir_volume should be mounted under a shared filesystem such as NFS." | 09:42 |
dardelean | this will make glance-api HA, right? | 09:43 |
dardelean | kolla-ansible will autodetect it as shared NFS storage and make it highly available? | 09:44 |
mnasiadka | it should | 09:45 |
dardelean | and ofc, on each controller node, the NFS share should be mounted prior to deployment, mounted on glance_file_datadir_volume | 09:50 |
yoctozepto | I doubt it does autodetection; there was a toggle for that I think | 09:52 |
mnasiadka | yoctozepto: glance_api_hosts: "{{ [groups['glance-api']|first] if glance_backend_file | bool and glance_file_datadir_volume == 'glance' else groups['glance-api'] }}" | 10:02 |
mnasiadka | so if glance_file_datadir_volume will be different, it will spawn multiple glance-api's | 10:03 |
mnasiadka | dardelean: and yes, you need to mount NFS on your own, and just point to the directory in K-A | 10:03 |
dardelean | got it, so as long glance_file_datadir_volume is different than "glance" it will spawn multiple apis, thanks all | 10:05 |
dardelean | also, I will use OVN I think, so if anyone has any special comments on it or things that I should know about kolla+OVN, info will be appreciated | 10:06 |
dardelean | ovn + qos plugin for neutron | 10:06 |
mnasiadka | there's no support for QoS plugin in OVN IIRC | 10:18 |
mnasiadka | https://docs.openstack.org/neutron/latest/ovn/gaps.html | 10:18 |
dardelean | #mnasiadka only for minimum bandwidth, but for maximim I think it is supported | 10:25 |
mnasiadka | dardelean: just ask Neutron guys to make sure | 10:25 |
dardelean | ok, thanks | 10:25 |
kevko | hi | 11:11 |
kevko | yoctozepto: can't find what label i should add to instance to be HA in masakari doc ..could you please help me ? | 11:12 |
opendevreview | Michal Nasiadka proposed openstack/kayobe master: [release] Sync Kolla Ansible inventory for Xena https://review.opendev.org/c/openstack/kayobe/+/814120 | 11:13 |
*** amoralej is now known as amoralej|lunch | 12:20 | |
mnasiadka | wsrep-notify.sh script is deprecated and will be removed in Xena | 12:26 |
mnasiadka | yoctozepto: did we forget to remove something? | 12:26 |
yoctozepto | mayhaps? | 12:33 |
yoctozepto | is it still there? | 12:33 |
mnasiadka | yup, when clustercheck is not enabled | 12:34 |
yoctozepto | ah, we were supposed to drop support for that scenario | 12:38 |
yoctozepto | oh well | 12:38 |
yoctozepto | let's do it now ;p | 12:38 |
mnasiadka | I found that, because trying to do an aio deploy, and mariadb decided to not cooperate - but after getting rid of wsrep-notify it works fine | 12:39 |
yoctozepto | hmm, interesting | 12:50 |
yoctozepto | anyhow, something weird in the nodes held by frickler - fluentd does not collect logs to local dir | 12:53 |
yoctozepto | anyway, good news is that a restart is enough | 12:55 |
yoctozepto | the node never joined the cluster; so the other two nodes work as if it was a two-node cluster | 12:56 |
yoctozepto | and can be restarted fine | 12:56 |
yoctozepto | and the failed one can be restarted and simply joins right away | 12:56 |
yoctozepto | otherwise it is stuck in that dumb state | 12:56 |
yoctozepto | frickler, mnasiadka, mgoddard: ^ | 12:57 |
*** amoralej|lunch is now known as amoralej | 12:58 | |
frickler | yoctozepto: but no indication as to why joining the cluster failed? | 13:00 |
yoctozepto | frickler: so far not; only what we have seen already | 13:02 |
yoctozepto | but /me not giving up yet | 13:02 |
opendevreview | Radosław Piliszek proposed openstack/kolla-ansible master: [WIP] [mariadb] Start new nodes serially https://review.opendev.org/c/openstack/kolla-ansible/+/814187 | 15:27 |
opendevreview | Radosław Piliszek proposed openstack/kolla-ansible master: [DNM] Test MariaDB 10.5 https://review.opendev.org/c/openstack/kolla-ansible/+/812944 | 15:37 |
opendevreview | Radosław Piliszek proposed openstack/kolla-ansible master: [DNM] Test MariaDB 10.5 https://review.opendev.org/c/openstack/kolla-ansible/+/812944 | 15:38 |
yoctozepto | frickler, mnasiadka, mgoddard: ^ | 15:39 |
yoctozepto | that might fight that race condition | 15:39 |
yoctozepto | it looks like a bug in galera that it has somehow become dumber | 15:40 |
yoctozepto | yet it might have always been but was hiding the actual issue until later | 15:40 |
yoctozepto | who knows | 15:40 |
opendevreview | Radosław Piliszek proposed openstack/kolla-ansible master: [DNM] Test MariaDB 10.6 https://review.opendev.org/c/openstack/kolla-ansible/+/813202 | 15:44 |
opendevreview | Radosław Piliszek proposed openstack/kolla-ansible master: [DNM] Test MariaDB 10.6 https://review.opendev.org/c/openstack/kolla-ansible/+/813202 | 15:44 |
urimeba | Hello everyone! I have just a quick question for you guys. If you could help I'll appreciate it | 15:46 |
urimeba | In Kolla-Ansible, the namespaces in Neutron are shared between neutron containers? | 15:47 |
*** amoralej is now known as amoralej|off | 16:47 | |
frickler | yoctozepto: interesting. we've always deployed the nodes serially with chef, so we wouldn't have seen such an issue there | 16:58 |
opendevreview | Michal Arbet proposed openstack/kolla-ansible master: Add weight of haproxy backend also for custom members https://review.opendev.org/c/openstack/kolla-ansible/+/814027 | 17:18 |
yoctozepto | frickler: hehe, we used to not serialise it past the first node at all; then I added it for the expansion/upgrade case https://review.opendev.org/c/openstack/kolla-ansible/+/701010 | 17:55 |
yoctozepto | now dealing the final blow :D | 17:56 |
-opendevstatus- NOTICE: The Gerrit service on review.opendev.org will be offline starting in 5 minutes, at 18:00 UTC, for scheduled project rename maintenance, which should last no more than an hour (but will likely be much shorter): http://lists.opendev.org/pipermail/service-announce/2021-October/000024.html | 17:59 | |
yoctozepto | frickler: looks really fixed | 19:21 |
yoctozepto | I think you can release the held nodes | 19:21 |
yoctozepto | have a nice weekend | 19:22 |
* yoctozepto off | 19:22 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!