| prometheanfire | is there a tag to do rolling restart of openstack services (kinda like galera has one)? | 00:16 |
|---|---|---|
| *** dviroel|afk is now known as dviroel | 00:41 | |
| *** dviroel is now known as dviroel|out | 00:50 | |
| NeilHanlon | i don't think so prometheanfire, but there's https://docs.openstack.org/openstack-ansible/latest/admin/troubleshooting.html#restarting-services if you've not seen it | 02:46 |
| opendevreview | OpenStack Proposal Bot proposed openstack/openstack-ansible master: Imported Translations from Zanata https://review.opendev.org/c/openstack/openstack-ansible/+/856910 | 04:12 |
| *** ysandeep|out is now known as ysandeep | 04:42 | |
| *** ysandeep is now known as ysandeep|brb | 05:02 | |
| *** ysandeep|brb is now known as ysandeep | 05:35 | |
| noonedeadpunk | prometheanfire: there's a playbook somewhere damiandabrowski contributed to for rolling restart of controller nodes | 07:22 |
| damiandabrowski | I'm afraid it was never merged :/ https://review.opendev.org/c/openstack/openstack-ansible-ops/+/806937 | 07:23 |
| noonedeadpunk | but rolling restart of a service is more a correct ad-hoc? | 07:23 |
| noonedeadpunk | ah, well. But smth you can contribute to if that's what you was looking for ;) | 07:24 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_ironic master: Bind http and tftp services to the bmaas network https://review.opendev.org/c/openstack/openstack-ansible-os_ironic/+/852122 | 07:33 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_ironic master: Bind http and tftp services to the bmaas network https://review.opendev.org/c/openstack/openstack-ansible-os_ironic/+/852122 | 07:35 |
| *** mgoddard- is now known as mgoddard | 08:11 | |
| *** ysandeep is now known as ysandeep|lunch | 09:06 | |
| ThiagoCMC | Morning, folks! In OSA, what's the status of Ironic, Masakari, and Trove? Ready for prime time? | 09:27 |
| noonedeadpunk | ThiagoCMC: I'd say they had some love quite recently | 10:00 |
| noonedeadpunk | With Masakari the only weak point can be pacemaker-remote inplementation | 10:00 |
| noonedeadpunk | while it should be supported by role I never had time to test it out for real | 10:01 |
| noonedeadpunk | But if your cluster is up to 24 computes - you should be good even without it | 10:02 |
| noonedeadpunk | (or will create multiple segments in masakari) | 10:02 |
| ThiagoCMC | noonedeadpunk, cool! I'm using `stable/yoga` now. Ironic is a priority for me now, Masaraki more of a nice to have, and Trove low. I'll give a try then, thanks! | 10:08 |
| opendevreview | Merged openstack/openstack-ansible-os_horizon master: Add the ability to define CSRF_TRUSTED_ORIGINS https://review.opendev.org/c/openstack/openstack-ansible-os_horizon/+/856363 | 10:19 |
| *** ysandeep|lunch is now known as ysandeep | 10:43 | |
| *** dviroel|out is now known as dviroel | 11:36 | |
| ThiagoCMC | In `stable/yoga`, when deploying OSA with Ceph, I noticed that it's also trying to install "gluster" for Glance, which isn't desirable now. Tasks related to it are failing but proceeding. Is there a way to disable/silent gluster tasks? | 11:46 |
| noonedeadpunk | ThiagoCMC: I believe gluster should be isntalled for repo containers only | 11:47 |
| noonedeadpunk | Yes, there's a variable to disable gluster deployment, but then you should provide some another shared filesystem (can be cephfs) | 11:48 |
| noonedeadpunk | But it should be available during repo containers deployment | 11:48 |
| noonedeadpunk | NeilHanlon: Regarding Rocky patch. Seems there's an issue with distro detection. | 11:59 |
| noonedeadpunk | based on /etc/os-release we detect it as "9.0", while here https://opendev.org/openstack/openstack-ansible/src/branch/master/scripts/bootstrap-ansible.sh#L76 case checks for 9. | 12:00 |
| noonedeadpunk | And CentOS has it as "9": https://zuul.opendev.org/t/openstack/build/f4389ac80ca04eeb9a949b80e01f4b68/log/job-output.txt#5681 | 12:01 |
| noonedeadpunk | so python3-devel is not installed and job get failed because of that | 12:01 |
| noonedeadpunk | I'm not sure tbh that it's supposed to be 9.0 As we check for DISTRO_VERSION_ID, which generally represents major version. And full version should be in $VERSION ? | 12:04 |
| noonedeadpunk | probably it can be that way according to https://www.freedesktop.org/software/systemd/man/os-release.html#VERSION_ID= | 12:05 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Add Rocky Linux 9 to zuul and docs https://review.opendev.org/c/openstack/openstack-ansible/+/857191 | 12:12 |
| *** frenzyfriday is now known as frenzyfriday|lunch | 12:32 | |
| jamesdenton | what is responsible for creating the /run/<service>/uwsgi directory if not the service itself? https://paste.opendev.org/show/bLfNvZ0jeCAtTBNEjTk0/ | 12:57 |
| jamesdenton | i have noticed that sometimes after a reboot those /run directories aren't there, and the respective service won';t start | 12:57 |
| mgariepy | https://www.freedesktop.org/software/systemd/man/systemd.exec.html#RuntimeDirectory= | 13:03 |
| mgariepy | jamesdenton, can you paste the systemd service files ? | 13:03 |
| jamesdenton | sure one sec | 13:03 |
| jamesdenton | https://paste.opendev.org/show/bveUUprS7rU6UtvfRp0S/ | 13:04 |
| jamesdenton | that's or glance, but it affects all of the uwsgi services, best i can tell | 13:04 |
| mgariepy | can you add the directive and test? | 13:06 |
| jamesdenton | yes, one sec | 13:07 |
| jamesdenton | no love, mgariepy | 13:15 |
| jamesdenton | mgariepy looks like we're using tmpfiles.d? | 13:18 |
| mgariepy | hmm yes indeed. | 13:19 |
| mgariepy | why doesn't it work then ? | 13:19 |
| jamesdenton | rerunning "systemd-tmpfiles --create --remove" did the needful | 13:19 |
| jamesdenton | not sure why that doesn't always happen at boot. | 13:19 |
| mgariepy | racing ? | 13:20 |
| mgariepy | .. :/ | 13:20 |
| jamesdenton | must be | 13:20 |
| mgariepy | is the service enabled and run at boot ? | 13:21 |
| mgariepy | maybe our service needs to depends on the tmpfile one. | 13:22 |
| mgariepy | It is possible to combine --create, --clean, and --remove in one invocation (in which case removal and cleanup are executed before creation of new files). For example, during boot the following command line is executed to ensure that all temporary and volatile directories are removed and created according to the configuration file: | 13:23 |
| mgariepy | systemd-tmpfiles --remove --create | 13:23 |
| jamesdenton | IIRC this is reproducible in an AIO, if you do a deploy and then reboot it, you might find the services won't start due to the /run dir missing. Easiest to replicate in an aio_metal | 13:25 |
| *** frenzyfriday|lunch is now known as frenzyfriday | 13:26 | |
| jamesdenton | not /run missing, but rather, the /run/<service> | 13:32 |
| cloudnull | dmsimard 👋 | 13:39 |
| opendevreview | Kevin Carter proposed openstack/openstack-ansible-os_horizon master: Add uwsgi option to horizon https://review.opendev.org/c/openstack/openstack-ansible-os_horizon/+/856393 | 13:46 |
| opendevreview | Kevin Carter proposed openstack/openstack-ansible-rabbitmq_server master: Update the heartbeat and handshake timeout https://review.opendev.org/c/openstack/openstack-ansible-rabbitmq_server/+/855996 | 13:51 |
| noonedeadpunk | jamesdenton: it's should be systemd tmpdirs or smth like that, sec | 14:14 |
| noonedeadpunk | ah, yes, thanks mgariepy :D | 14:15 |
| jamesdenton | yes, thanks. my systemd-tmpfiles-setup.service is dead on this node, but no indication why. | 14:15 |
| noonedeadpunk | I did saw that, actually | 14:15 |
| noonedeadpunk | But the reason when I digged was systemd-tmpfiles startup failure indeed due to race condition | 14:16 |
| noonedeadpunk | Like it was started before some bind mount was in place or smth like that | 14:16 |
| noonedeadpunk | maybe we should take care of starting it up after some service/target that we don't defined | 14:17 |
| ThiagoCMC | noonedeadpunk, thanks! I'm using Ceph as I always did (with Victoria, and older), but now with Yoga (same old /etc/openstack_deploy copied), it complains about Gluster... I'll check it out! | 14:17 |
| noonedeadpunk | ThiagoCMC: lsyncd on repo container has been replaced with gluster. Though despite some task fails during deployment, it's recovered with rescue block. So should not be an issue | 14:18 |
| jamesdenton | noonedeadpunk i will try to spend a few cycles on it. Any idea if this also impacts centos? | 14:18 |
| noonedeadpunk | nope, haven't tested centos for a while now | 14:23 |
| noonedeadpunk | well, on X I did rebooted bunch of production machines recently, and I haven't seen that issue there | 14:23 |
| jamesdenton | it's definitely intermittent | 14:27 |
| noonedeadpunk | and service restart just works iirc | 14:29 |
| jamesdenton | yes, it appears that way | 14:30 |
| noonedeadpunk | #startmeeting openstack_ansible_meeting | 15:00 |
| opendevmeet | Meeting started Tue Sep 13 15:00:17 2022 UTC and is due to finish in 60 minutes. The chair is noonedeadpunk. Information about MeetBot at http://wiki.debian.org/MeetBot. | 15:00 |
| opendevmeet | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 15:00 |
| opendevmeet | The meeting name has been set to 'openstack_ansible_meeting' | 15:00 |
| noonedeadpunk | #topic rollcall | 15:00 |
| noonedeadpunk | o/ | 15:00 |
| damiandabrowski | hi! | 15:00 |
| noonedeadpunk | I finally managed to set my alarm :D | 15:00 |
| * damiandabrowski waiting for winter time change :D | 15:02 | |
| noonedeadpunk | OH NO | 15:02 |
| rgunasekaran | Hi | 15:02 |
| * noonedeadpunk googles if Croatia has summer time | 15:03 | |
| rgunasekaran | Who is from Croatia? | 15:03 |
| noonedeadpunk | ¯\(◉◡◔)/¯ | 15:04 |
| noonedeadpunk | #topic office hours | 15:04 |
| noonedeadpunk | So, first thing is that rocky9 job has failed. And I broke it even more now :D | 15:05 |
| * noonedeadpunk somebody need to reacall bash | 15:05 | |
| noonedeadpunk | The issue is that Rocky 9 defines VERSION_ID as 9.0 while it used to have bare 8 before. And CentOS also does have bare 9 | 15:07 |
| noonedeadpunk | So we somehow need to have regular expression in case statement https://review.opendev.org/c/openstack/openstack-ansible/+/857191/3/scripts/bootstrap-ansible.sh | 15:08 |
| noonedeadpunk | and I obviosly did it wrong | 15:09 |
| noonedeadpunk | another thing is that I made release for yoga that is broken because of keepalived version | 15:10 |
| damiandabrowski | does case statement even support regex? | 15:10 |
| noonedeadpunk | it should support shel patterns... https://www.gnu.org/savannah-checkouts/gnu/bash/manual/bash.html#Pattern-Matching | 15:11 |
| NeilHanlon | o/ - double booked today but I saw your message earlier noonedeadpunk re: rocky, makes sense.. that's "expected" because Stream doesn't have minor releases, it's always just a whole number vs RHEL (and rocky/alma) which have the concept of minor releases | 15:12 |
| * jamesdenton reminder to discuss keepalived/vrrp 8 character auth truncation when the time is right | 15:13 | |
| noonedeadpunk | NeilHanlon: I'm pretty sure that full version used to be in CentOS elsewhere then in VERSION_ID? | 15:13 |
| mgariepy | 9*(.*[0-9])) | 15:14 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Add Rocky Linux 9 to zuul and docs https://review.opendev.org/c/openstack/openstack-ansible/+/857191 | 15:15 |
| noonedeadpunk | let's try it out :D | 15:15 |
| noonedeadpunk | ok, yes, 8 char auth | 15:15 |
| noonedeadpunk | do we want to leave auth at all :D? | 15:15 |
| mgariepy | or more: 9*(.+[0-9]) | 15:16 |
| noonedeadpunk | as 1 - it's not _really_ working. Given another keepalived will run in the same multicast domain, auth won't be really respected | 15:16 |
| noonedeadpunk | ah, yes, makes sense mgariepy | 15:16 |
| ThiagoCMC | Folks, my setup-infrastructure.yml (Ubuntu 20.04) is failing to mount /var/www/repo in the repo container. I can't see any gluster being configured by OSA, should I do this in advance before OSA? Or can I just ignore this? | 15:16 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Add Rocky Linux 9 to zuul and docs https://review.opendev.org/c/openstack/openstack-ansible/+/857191 | 15:17 |
| noonedeadpunk | and eventually auth has been dropped from VRRP 3 | 15:17 |
| noonedeadpunk | So maybe we should look into switching to VRRP 3 isntead? | 15:18 |
| jamesdenton | we could do that | 15:19 |
| noonedeadpunk | so what we saw one day - is keepalived was flapping because of another keepalived was running in same network with same router ids (but different auth password) | 15:21 |
| noonedeadpunk | so auth kind of useless anyway imo | 15:21 |
| noonedeadpunk | and switching to vrrp3 will be step forward into ipv6 only deployments | 15:21 |
| jamesdenton | that makes it easy, then | 15:21 |
| noonedeadpunk | Which I guess we should set as a goal one day | 15:21 |
| noonedeadpunk | ok, great then agreed? :) | 15:22 |
| jamesdenton | # Note: authentication was removed from the VRRPv2 specification by RFC3768 in 2004. | 15:22 |
| anskiy | noonedeadpunk: well, that's why I set unicast_peers :) | 15:22 |
| jamesdenton | lol, yes, agreed | 15:22 |
| noonedeadpunk | #agreed drop auth from keepalived and work on switching to VRRPv3 | 15:23 |
| noonedeadpunk | And I will release bugfix for yoga as it's now broken because of keepalived | 15:25 |
| jamesdenton | sounds good | 15:25 |
| jamesdenton | should we consider unicast or stick with multicast? | 15:26 |
| noonedeadpunk | Um, not really sure | 15:27 |
| noonedeadpunk | I wonder about possible overhead when having unicast | 15:27 |
| anskiy | https://review.opendev.org/c/openstack/openstack-ansible/+/854370 with this, there could be just a mention in docs | 15:28 |
| jamesdenton | good compromise | 15:28 |
| noonedeadpunk | well, feel free to add usecase to docs ;) | 15:28 |
| jamesdenton | is that how you configured yours? | 15:28 |
| anskiy | jamesdenton: I think, I'm still overriding the whole thing... But that was the purpose, yeah | 15:29 |
| jamesdenton | cool cool | 15:30 |
| noonedeadpunk | Eventually it's also time to fill in topics for upcoming PTG. | 15:33 |
| noonedeadpunk | I;ve jsut created etherpad page - will put some content soon | 15:34 |
| noonedeadpunk | #link https://etherpad.opendev.org/p/osa-antelope-ptg | 15:34 |
| noonedeadpunk | Regarding that - I think we might want to add operator hour as well | 15:37 |
| noonedeadpunk | This might be useful to hear back from operators. I'm not sure how that initiative will fly, but worth trying | 15:37 |
| noonedeadpunk | I'm thinking to sign-up osa on wednesday. wdyt? | 15:38 |
| damiandabrowski | fine for me | 15:38 |
| mgariepy | works for me also | 15:40 |
| mgariepy | wenesday is oct 19 tho ;) | 15:41 |
| noonedeadpunk | I'm talking about operators hours I was talking about. For the team PTG I've signed up on Tuesday as usual | 15:42 |
| noonedeadpunk | But it's not written in stone - I can change that if needed? | 15:42 |
| damiandabrowski | so PTG on Tuesday and operator hours on Wednesday? ok for me | 15:45 |
| mgariepy | ha ok . sorry not following 100%, i'm in a meeting | 15:45 |
| noonedeadpunk | yup, nice | 15:45 |
| *** dviroel is now known as dviroel|lunch | 15:51 | |
| *** ysandeep is now known as ysandeep|out | 15:57 | |
| noonedeadpunk | #endmeeting | 15:57 |
| opendevmeet | Meeting ended Tue Sep 13 15:57:58 2022 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 15:57 |
| opendevmeet | Minutes: https://meetings.opendev.org/meetings/openstack_ansible_meeting/2022/openstack_ansible_meeting.2022-09-13-15.00.html | 15:57 |
| opendevmeet | Minutes (text): https://meetings.opendev.org/meetings/openstack_ansible_meeting/2022/openstack_ansible_meeting.2022-09-13-15.00.txt | 15:57 |
| opendevmeet | Log: https://meetings.opendev.org/meetings/openstack_ansible_meeting/2022/openstack_ansible_meeting.2022-09-13-15.00.log.html | 15:57 |
| noonedeadpunk | ThiagoCMC: I have some doubts if it is _really_ failing to mount | 16:02 |
| noonedeadpunk | As the job that can fail got's rescued and should be done with follow-up task | 16:02 |
| noonedeadpunk | you can check if you have mount inside your repo container to verify it exists | 16:02 |
| noonedeadpunk | mgariepy: fwiw it seems not to work :( | 16:13 |
| mgariepy | hmm | 16:21 |
| ThiagoCMC | Somehow, the Gluster tasks worked! `/var/www/repo` is mounted with fuse.glusterfs, cool... ! I'm using MaaS, easy to "Release/Deploy", then try OSA again. Thanks! | 16:37 |
| *** dviroel|lunch is now known as dviroel | 16:57 | |
| mgariepy | noonedeadpunk, rockylinux9 bash seems weird. lol | 16:59 |
| BobZ_Annapolis | sorry if this is repeating a known issue but i had an up-running o-a yoga aio for a few weeks, wanted to stand up another one for a variety of reasons and i can't get past setup-hosts - consistently crashing at lxc_hosts : Build the base image using a CLI tool - wall of red, looks like it's having extracting images from official packages - sound familiar ? tia | 16:59 |
| mgariepy | case : 9|9.0) ? | 17:04 |
| noonedeadpunk | mgariepy: I bet they will change number wrt minor version. So doesn't make sense to hardcode `.0` | 17:34 |
| mgariepy | case : 9|9[.0-9]*) ? | 17:35 |
| mgariepy | https://paste.openstack.org/show/bfL8bEA0D8fwD9KZ8BeG/ | 17:36 |
| noonedeadpunk | nice! you want me to try it?:) | 17:43 |
| noonedeadpunk | BobZ_Annapolis: as of today we build lxc images using deb-bootstrap | 17:44 |
| mgariepy | sure | 17:44 |
| noonedeadpunk | BobZ_Annapolis: would be quite helpful if you could paste error through paste.openstack.org | 17:45 |
| noonedeadpunk | and no, it's not smth well-known | 17:45 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Add Rocky Linux 9 to zuul and docs https://review.opendev.org/c/openstack/openstack-ansible/+/857191 | 17:45 |
| BobZ_Annapolis | sorry, not familiar with paste.openstack.org - went there - pasted it - gave me this reply : #bq8js0q1GAHU9y9Sq7NS | 17:47 |
| BobZ_Annapolis | https://paste.openstack.org/raw/bq8js0q1GAHU9y9Sq7NS/ | 17:48 |
| BobZ_Annapolis | fyi, i stood up a successful o-a yoga aio on August 12 successfully following the online aio steps; just went back there today to run thru it again and am having this lxc_hosts build CLI issue, thx | 17:53 |
| jamesdenton | noonedeadpunk noticing this recently on multi-node deploys: https://paste.opendev.org/show/bj9UZ23B4Sb3rbOnCKBw/ --- is there a good way to avoid without setting venv_wheel_build_enable: false | 18:06 |
| noonedeadpunk | jamesdenton: re-run os-repo-install.yml | 18:08 |
| jamesdenton | repo-install.yml? | 18:08 |
| noonedeadpunk | ah, yes | 18:09 |
| noonedeadpunk | BobZ_Annapolis: anything interesting in /var/lib/machines/ubuntu-20-amd64/debootstrap/debootstrap.log ? | 18:09 |
| noonedeadpunk | BobZ_Annapolis: also maybe worth trying: a) running apt clean (as it's obviously going from cache) b) drop /var/lib/machines/ubuntu-20-amd64 | 18:11 |
| BobZ_Annapolis | debootstrap.log : https://paste.openstack.org/show/816713/ | 18:11 |
| noonedeadpunk | really, try to clean apt cache | 18:13 |
| BobZ_Annapolis | trying now, thx | 18:14 |
| noonedeadpunk | `/var/lib/dpkg/tmp.ci/preinst: 6: cannot create /etc/passwd: Permission denied` is quite weird... As it's in chroot.. | 18:14 |
| jamesdenton | same error. and to add to that, /var/www/repo/os-releases/25.1.0.dev68 is empty so there's seemingly nothing to sync | 18:14 |
| opendevreview | Merged openstack/openstack-ansible-os_ironic master: Bind http and tftp services to the bmaas network https://review.opendev.org/c/openstack/openstack-ansible-os_ironic/+/852122 | 18:15 |
| noonedeadpunk | jamesdenton: um, and gluster is mounted? | 18:15 |
| noonedeadpunk | or this is metal deploy? | 18:16 |
| noonedeadpunk | (but still gluster should be present for repo sync) | 18:16 |
| jamesdenton | it's a metal deploy, and test, gluster is mounted and working | 18:17 |
| BobZ_Annapolis | AWS vm on Ubuntu 20.04; after apt clean - same errors as pasted above :-( | 18:17 |
| jamesdenton | 3 infra nodes. | 18:17 |
| jamesdenton | BobZ_Annapolis you are running as root? | 18:19 |
| jamesdenton | maybe it won't even let you try as non-root, i dunno | 18:19 |
| BobZ_Annapolis | Yup, doing everything as root. Built w/o issue on Aug 12th - this ansible task completed with a nice simple "changed" | 18:19 |
| jamesdenton | built on an AWS VM the same way? | 18:19 |
| BobZ_Annapolis | Yup. Went to the yoga o-a site and followed the QuickStart. We have an AWS CLI utility that we just need to provide the name & number of machines we want to stand up and it does the rest. Used the same utility to stand up 1 vm, same resources, size, network, security groups, etc | 18:22 |
| noonedeadpunk | jamesdenton: hm... that is really weird.... reading paste again, seems that wheels built was skipped, but their build is enabled... | 18:25 |
| noonedeadpunk | are you running with some tags or some limit? | 18:25 |
| jamesdenton | no, but maybe there is an errant override in place | 18:25 |
| noonedeadpunk | So eventually https://opendev.org/openstack/ansible-role-python_venv_build/src/branch/master/tasks/main.yml#L62-L71 is skipped while it supposed to run | 18:26 |
| jamesdenton | ok, lemme grab a few vars in debug and see what we're not matching on | 18:28 |
| BobZ_Annapolis | Obviously somethingis amiss :-) . . .i'm just re-running thru the QuickStart AIO steps folks, word for word, not modifying any files yet - just want to get the basic functionality / system up-and-running again. | 18:29 |
| jamesdenton | not much to mess up with the AIO if you play by the book. It wouldn't be the first time the environment or image screwed with things, though. | 18:31 |
| noonedeadpunk | Well, we would see that in CI for sure though | 18:33 |
| jamesdenton | noonedeadpunk https://paste.opendev.org/show/bTs7A7RPieBBbmYDYGYp/ -- #140. Looks like it's being set to false for infra1, the deploy host. | 18:35 |
| jamesdenton | i'll have to dig and see why | 18:35 |
| jamesdenton | ./python_venv_build/defaults/main.yml:119:venv_wheel_build_enable: "{{ venv_build_host != inventory_hostname }}" | 18:36 |
| noonedeadpunk | um | 18:39 |
| jamesdenton | eh, that's not it. | 18:39 |
| noonedeadpunk | are you sure ?:) | 18:40 |
| noonedeadpunk | at it kind of looks like it | 18:40 |
| jamesdenton | well, we have this: "msg": "JD - venv_wheel_build_enable is False", but "msg": "JD - venv_build_host is infra1 and inventory_hostname is infra1" | 18:41 |
| jamesdenton | oh well | 18:41 |
| jamesdenton | seems reversed? | 18:41 |
| jamesdenton | lemme see... | 18:41 |
| jamesdenton | OK, with this --> venv_wheel_build_enable: "{{ venv_build_host == inventory_hostname }}", i get the wheels building on infra1 and gluster'ed over to infra2/3 | 18:45 |
| jamesdenton | utility playbook complete. | 18:47 |
| noonedeadpunk | I think, idea of "{{ venv_build_host != inventory_hostname }}" was not to build wheels either against localhost or when there're only 1 host available | 18:53 |
| noonedeadpunk | But I guess I've changed logic of venv_build_host generation | 18:53 |
| noonedeadpunk | so this condition broke for metal deploys | 18:53 |
| jamesdenton | aye | 18:53 |
| noonedeadpunk | as before we always choose last host for venv_build_host | 18:54 |
| noonedeadpunk | (like repo3 or infra3) | 18:54 |
| noonedeadpunk | and now it's first one that is chosen | 18:54 |
| jamesdenton | right, ok | 18:54 |
| noonedeadpunk | question what we should do with all that... | 18:57 |
| jamesdenton | i was hoping you knew :D | 18:59 |
| noonedeadpunk | hehe, I have another incident now :( so a bit distracted | 19:00 |
| jamesdenton | oh, no worries. | 19:00 |
| jamesdenton | i'll open a bug and we can deal later | 19:00 |
| BobZ_Annapolis | eeesh. . .would you believe my issue was an AWS availability-zone? Me neither but my working o-a yoga aio is in east-1b.All of today's failures were in east-1a.For the hell of it, changed a new vm to AZ east-1b and am running thru the steps now - got past setup-hosts and setu-infrastructure, now running thru setup-openstack ???? | 19:08 |
| jamesdenton | that's frustrating | 19:09 |
| BobZ_Annapolis | That's 1 word for it. Since this is just a quick prototype demo, not sure i want to go down the wormhole of opening up AWS Support tickets - does make me worry about future production possibilities tho. Least it's something we can document here to keep an eye on as we move forward - thanks for all your suggestions / help folks !! | 19:10 |
| jamesdenton | any time! good luck | 19:11 |
| mgariepy | An exception occurred during task execution. To see the full traceback, use -vvv. The error was: ModuleNotFoundError: No module named 'selinux' | 19:21 |
| mgariepy | at least the case seems to work lol | 19:22 |
| noonedeadpunk | BobZ_Annapolis: well, for me it felt like some de-synced mirrors, IMO | 19:26 |
| noonedeadpunk | for production you can always keep your if needed. | 19:26 |
| noonedeadpunk | mgariepy: yup, which is quite nice :D | 19:26 |
| noonedeadpunk | It's weird though as ansible can deal with selinux bindings now I guess.. | 19:27 |
| noonedeadpunk | Oh! | 19:27 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Bump ansible-core version to 2.13.4 https://review.opendev.org/c/openstack/openstack-ansible/+/857506 | 19:29 |
| noonedeadpunk | this should have fixed nasty apt issue ^ | 19:29 |
| BobZ_Annapolis | no1dp : thanks, it doesn't sit right w/me that it's an AWS AZ issue but . . . .after spending a few hours on this am anxious to move forward - will have some interns do some testing / try out some variations. thx again | 19:31 |
| opendevreview | James Denton proposed openstack/openstack-ansible master: Switch keepalived vrrp version from 2 to 3 https://review.opendev.org/c/openstack/openstack-ansible/+/857507 | 19:45 |
| jamesdenton | ^^ prob need a release note here for upgrades | 19:46 |
| jamesdenton | Also, https://bugs.launchpad.net/openstack-ansible/+bug/1989506. For later. | 19:56 |
| *** dviroel is now known as dviroel|brb | 20:10 | |
| opendevreview | Neil Hanlon proposed openstack/openstack-ansible master: Add Rocky Linux 9 to zuul and docs https://review.opendev.org/c/openstack/openstack-ansible/+/857191 | 21:56 |
| NeilHanlon | that's more of a shot in the dark tbh. i am not sure why it doesn't think python3-libselinux is available.. or rather, why it's not available in the virtualenv | 22:04 |
| opendevreview | Neil Hanlon proposed openstack/openstack-ansible master: Add Rocky Linux 9 to zuul and docs https://review.opendev.org/c/openstack/openstack-ansible/+/857191 | 22:39 |
| opendevreview | Ebbex proposed openstack/openstack-ansible-os_keystone master: Remove redundant vars line https://review.opendev.org/c/openstack/openstack-ansible-os_keystone/+/857521 | 22:54 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!