| opendevreview | Stanislav Dmitriev proposed openstack/kolla master: Add Swift lock path in Swift containers https://review.opendev.org/c/openstack/kolla/+/818497 | 01:15 |
|---|---|---|
| opendevreview | Stanislav Dmitriev proposed openstack/kolla master: Add Swift lock path in Swift containers https://review.opendev.org/c/openstack/kolla/+/788073 | 01:21 |
| opendevreview | Merged openstack/kolla-ansible master: haproxy: remove unused tls check condition in config https://review.opendev.org/c/openstack/kolla-ansible/+/818179 | 02:29 |
| mnasiadka | yoctozepto: do you mean we would go to untagged release model? but then people doing pip install kolla-ansible would be in some weird place :) | 06:29 |
| jingvar | someone used masakari? | 06:46 |
| *** EugenMayer4 is now known as EugenMayer | 07:34 | |
| *** brinzhang_ is now known as brinzhang | 07:44 | |
| *** amoralej|off is now known as amoralej | 08:10 | |
| opendevreview | Merged openstack/kolla stable/xena: reno: polish release notes for Xena https://review.opendev.org/c/openstack/kolla/+/818470 | 08:34 |
| opendevreview | Pierre Riteau proposed openstack/kayobe master: Document that extra kernel parameters are important for inspection https://review.opendev.org/c/openstack/kayobe/+/818491 | 08:38 |
| DK4 | what would be the way to enter the mariadb database as a root? is the password saved in the passwords file? | 09:05 |
| opendevreview | Pierre Riteau proposed openstack/kayobe master: Clean up release notes for Xena release https://review.opendev.org/c/openstack/kayobe/+/818521 | 09:11 |
| jingvar | DK4: yes | 09:14 |
| DK4 | jingvar: but which one is it? i cannot seem to find it | 09:30 |
| jingvar | :) I've got the same at first time , password.yml database_password: | 09:34 |
| jingvar | I mean grep ^database_password: | 09:35 |
| DK4 | yes its *suddenly* there.. thanks :p | 09:41 |
| jingvar | yaw | 09:43 |
| DK4 | a complete cloud reboot (all controllers) will always result in the db cluster broken - atleast on my end - a db_recovery will fix that problem and gets the db operational - but is there anyway to make that reboot safe in case of disaster? | 09:44 |
| JohnB | No. A full shutdown (or more that quorum) of a database cluster results in a broken cluster. You can't do that safe. | 09:48 |
| opendevreview | Dr. Jens Harbott proposed openstack/kolla-ansible master: Stop creating non-keystone admin endpoints https://review.opendev.org/c/openstack/kolla-ansible/+/816164 | 09:51 |
| frickler | shutting down the controllers one by one and then starting them again in reverse order might work | 09:55 |
| frickler | mnasiadka: should I just add kolla to the existing bug or create a new one for the OVN/dns issue? (818362) I guess a new one would be better | 09:59 |
| mnasiadka | I think a new one would be cleaner | 10:01 |
| hrw | https://review.opendev.org/c/openstack/kolla/+/818240 - anyone to review 'binary is deprecated' doc/reno? | 10:04 |
| frickler | hrw: commented, I can also propose an update myself if you prefer | 10:09 |
| hrw | frickler: once patch is in gerrit it can be fixed/updated by others | 10:09 |
| hrw | frickler: dropping non-debian source builds is next cycle | 10:10 |
| hrw | this step is for 'binary is deprecated' | 10:10 |
| frickler | I was just thinking that it would help to increase the urgency of the warning. "consider using source" is much softer than "what you are doing will soon stop working" | 10:12 |
| jingvar | someone used masakari? | 10:15 |
| hrw | frickler: replied to comments | 10:15 |
| hrw | frickler: reread '5s delay' patch please | 10:17 |
| frickler | hrw: I'm still reading that as trying to build debian/aarch/binary will result in: warning msg,sleep(5),failure msg, exit() and I'm arguing just doing failure msg,exit() would be better in that case | 10:21 |
| hrw | frickler: debian/aarch64/binary was not supported combo due to lack of aarch64 packages. but as binary is deprecated now it will not be reenabled | 10:22 |
| hrw | frickler: 5sdelay patch adds block above that one | 10:23 |
| hrw | ah. you mean "please move that check after aarch64 one" | 10:23 |
| hrw | frickler: right? | 10:23 |
| * hrw out | 10:25 | |
| bkra | is there a way to skip net allocation validation in Kayobe? "Found invalid existing allocations in network X"? our seed IP and management IPs are in a different subnet.. | 10:33 |
| mgoddard | bkra: you could make the relevant *_net_name variables into group_vars, with a different network for the seed | 10:43 |
| mgoddard | bkra: similar to https://docs.openstack.org/kayobe/latest/configuration/reference/routed-control-plane-networks.html | 10:44 |
| bkra | ok will try, thx | 10:55 |
| DK4 | s there a master db in kolla HA setup? i think all are master? | 10:56 |
| bkra | mgoddard: tried to replace oc_managment in seed_network_interfaces with a custom interface but then "Host seed has no address configured on the admin network." keeps popping up at ssh-known-host | 11:04 |
| jingvar | DK4: what do you mean, mysql galera cluster itself HA | 11:23 |
| opendevreview | Seena Fallah proposed openstack/kolla-ansible stable/xena: haproxy: remove unused tls check condition in config https://review.opendev.org/c/openstack/kolla-ansible/+/818483 | 11:25 |
| DK4 | jingvar: i have a 3 controller node setup. im testing to backup and restore a full backup. i've choosen one node to perform a database restore. after the restored node comes backup online it receices a sst tranfer from a donor of the remaining nodes | 11:25 |
| opendevreview | Seena Fallah proposed openstack/kolla-ansible stable/wallaby: haproxy: remove unused tls check condition in config https://review.opendev.org/c/openstack/kolla-ansible/+/818484 | 11:26 |
| DK4 | thus im back at the point before the restore | 11:26 |
| DK4 | which is i guess expected behaviour. are there any docs that help restoring db in multinode setups? | 11:28 |
| opendevreview | Seena Fallah proposed openstack/kolla-ansible stable/wallaby: haproxy: remove unused tls check condition in config https://review.opendev.org/c/openstack/kolla-ansible/+/818484 | 11:30 |
| jingvar | the doc has info about db save, backup and restore | 11:32 |
| opendevreview | Merged openstack/kayobe master: CI: always return host configure test results https://review.opendev.org/c/openstack/kayobe/+/812941 | 11:33 |
| jingvar | let me rephrase my question, masakari project is dead? | 11:33 |
| mgoddard | jingvar: yoctozepto (PTL) would probably dispute that | 11:41 |
| jingvar | you are first who heard me | 11:43 |
| jingvar | maybe Is a different instance HA product preferred? | 11:55 |
| mnasiadka | Yes, VMWare | 11:57 |
| mnasiadka | :) | 11:57 |
| mnasiadka | (I'm arguing a cloud doesn't need that) | 11:58 |
| holtgrewe | I'm deploying with current in-dev stable/xena branches. Has anyone seen "kayobe overcloud service deploy" fail like this before? https://paste.openstack.org/show/811189/ 'dict object' has no attribute 'hostname' | 11:59 |
| jingvar | mnasiadka: it is philosophical question | 12:00 |
| mgoddard | jingvar: something wrong with your facts? using a cache? | 12:04 |
| jingvar | mgoddard: sorry I don't got it | 12:05 |
| jingvar | I just need an HA for instances. I'v found Masakari project. but I have issues with it. Maybe there is better choise | 12:07 |
| mgoddard | jingvar: sorry, that should have gone to holtgrewe | 12:07 |
| holtgrewe | mgoddard, not that I'm aware of... let me check a few things | 12:27 |
| holtgrewe | mgoddard, the problem is above this, apparently some hosts are not reachable any more... - what do you know, multi-bit memory error... | 12:36 |
| DK4 | jingvar: the doc has some inmfo about restoring, but i guess it doesnt apply to galera cluster | 12:44 |
| yoctozepto | mgoddard, jingvar: yeah, but it seems there is only PTL left there to support and he's (read: I am) busy with other duties that are sadly not relevant to Masakari :-( that said, Masakari works, as of wallaby, just fine and it keeps working in xena too | 12:57 |
| yoctozepto | mnasiadka: well, always make them follow the docs :-) | 12:58 |
| frigo | Hello Kolla! | 13:06 |
| frigo | I was wondering if anyone was running the kolla built containers on top of Fedora CoreOS (meaning, having Fedora CoreOS on the hypervisors). With regards to the discussion on openstack discuss, it would make sense to have libvirt and ovs packaged as containers for CoreOS... | 13:06 |
| mgoddard | Hi frigo. We don't actively support coreos as a host OS | 13:20 |
| frigo | yes I read that. But I think it would be nice :) and fitting the purpose since OpenStack is mostly deployed as containers. Just wondering if maybe someone somewhere tried it and had feedback | 13:22 |
| *** amoralej is now known as amoralej|lunch | 13:24 | |
| opendevreview | Mark Goddard proposed openstack/kayobe stable/xena: CI: always return host configure test results https://review.opendev.org/c/openstack/kayobe/+/818489 | 13:25 |
| opendevreview | Mark Goddard proposed openstack/kayobe stable/wallaby: CI: always return host configure test results https://review.opendev.org/c/openstack/kayobe/+/818570 | 13:25 |
| opendevreview | Mark Goddard proposed openstack/kayobe stable/victoria: CI: always return host configure test results https://review.opendev.org/c/openstack/kayobe/+/818571 | 13:25 |
| opendevreview | Mark Goddard proposed openstack/kayobe master: Import kolla-ansible baremetal role https://review.opendev.org/c/openstack/kayobe/+/818290 | 13:27 |
| yoctozepto | frigo: help welcome :-) | 13:48 |
| frigo | yoctozepto haha yes I might give it a try :) | 13:50 |
| holtgrewe | Is it possible to configure keystone to use an upstream LDAP for authentication with kolla/kayobe? | 14:00 |
| mgoddard | holtgrewe: yes | 14:07 |
| holtgrewe | mgoddard, I could find no documentation and no obvious ansible variable for spiking in configuration into the keystone configuration file. | 14:07 |
| mgoddard | yeah, I don't think it's documented current;y | 14:07 |
| holtgrewe | ;-) | 14:08 |
| *** amoralej|lunch is now known as amoralej | 14:15 | |
| jingvar | victoria/stable kayobe type install source , I have in masakari container masakari-base-source/masakari-10.0.3.dev1 - is it OK? dev version in stable? | 14:17 |
| mgoddard | jingvar: just means you're using the stable branch, 1 patch after 10.0.3 | 14:28 |
| jingvar | I try to understand why it not works at all | 14:29 |
| jingvar | #openstack-masakari is dead? I don't see any activities | 14:32 |
| jingvar | https://docs.openstack.org/kolla/latest/support_matrix x86_64 images Image Centos / masakari N T - is it realy Tested | 14:35 |
| bkra | Created a custom network for seed node; getting "Required network 'oc_management' (Bifrost provisioning network) is not mapped to this host;" Is there a way to remap this network to my custom network? can't find it anywhere tbf. cheers! | 15:07 |
| mgoddard | bkra: did you see my suggestion earlier? make a second network, and use group_vars to choose which one each host uses for provision_oc_net_name | 15:09 |
| priteau | Has anyone noticed that any change to prometheus, even a small config change to a single exporter, triggers a restart of *all* the prometheus containers? | 15:17 |
| priteau | I've looked at the handler code and I don't understand why | 15:17 |
| bkra | mgoddard: yes i did, but i now see that provision_oc_net_name was still set in globals.yml hehe. thank you for the swift replies :) | 15:21 |
| mgoddard | priteau: ansible 'bug/feature' | 15:22 |
| priteau | Sounds like a bug to me! | 15:22 |
| mgoddard | https://bugs.launchpad.net/kolla-ansible/+bug/1863510 | 15:22 |
| mgoddard | feel free to explain that to the ansible devs | 15:22 |
| mgoddard | we could not | 15:22 |
| hrw | have a nice weekend | 15:25 |
| *** amoralej is now known as amoralej|off | 17:29 | |
| opendevreview | Marcin Juszkiewicz proposed openstack/kolla master: build: 5s delay + note when binary images are used https://review.opendev.org/c/openstack/kolla/+/818378 | 17:50 |
| opendevreview | Merged openstack/kolla-ansible stable/xena: haproxy: remove unused tls check condition in config https://review.opendev.org/c/openstack/kolla-ansible/+/818483 | 19:39 |
| opendevreview | Merged openstack/kolla-ansible stable/wallaby: haproxy: remove unused tls check condition in config https://review.opendev.org/c/openstack/kolla-ansible/+/818484 | 19:39 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!