| prometheanfire | offhand... is there a playbook or something to clear out old venvs (old install, updated 6 or so times, the lvm parts per container are getting a bit full) | 05:08 |
|---|---|---|
| *** akahat|rover is now known as akahat | 07:05 | |
| jrosser | prometheanfire: https://github.com/openstack/openstack-ansible-ops/blob/master/ansible_tools/playbooks/cleanup-venvs.yml | 08:15 |
| jrosser | prometheanfire: ymmv as it's the ops repo so best to test first | 08:15 |
| ThiagoCMC | Morning, folks! I'm seeing a problem with `dynamic_inventory.py` on CentOS 9 Stream, OSA `stable/zed`. It returns "hosts (0)". The very same `/etc/openstack_deploy` directory works on Ubuntu 20.04. Detailed error output: https://paste.openstack.org/show/bIGlB1G4R293ymbnhz40/ - Please, advise. | 10:47 |
| *** dviroel_ is now known as dviroel|rover | 11:24 | |
| jrosser | ThiagoCMC: you can run the dynamic inventory script standalong (without ansible) to see if there is any further useful error messages | 13:00 |
| jrosser | also it has a `--debug` which produces tons of output | 13:00 |
| ThiagoCMC | jrosser, hello! Yes, if I run `/opt/openstack-ansible/inventory/dynamic_inventory.py --list` it outputs the JSON. | 13:03 |
| ThiagoCMC | but it looks like that `ansible -i ...` is not running the Python script, instead, it looks like it's passing the Python script straight to `ansible`, hence the output "dynamic_inventory.py:18: Expected key=value host variable assignment, got: argparse". | 13:04 |
| jrosser | well i don't have a centos-9s system to try this on | 13:05 |
| jrosser | but it would be interesting to know if the same happens on a fresh AIO after doing bootstrap-aio.sh | 13:05 |
| ThiagoCMC | Sure, I'll try that. BTW, I also tried on Oracle Linux 9, same result. | 13:06 |
| jrosser | but otherwise the OSA playbooks are working, or is it all broken? | 13:07 |
| ThiagoCMC | It's broken (on RH9-based distros, AFAIK), it doesn't pass past the `scripts/bootstrap-ansible.sh`, because it can't find any hosts. | 13:07 |
| ThiagoCMC | Not even "ping all" works | 13:08 |
| ThiagoCMC | The `openstack-ansible setup-hosts.yml` after `scripts/bootstrap-aio.sh` seems to be proceeding (Oracle Linux 9), however, it printed: "[WARNING]: Unable to parse /etc/openstack_deploy/inventory.ini as an inventory source". - https://paste.openstack.org/show/bjefQxNkG3BOr2fkYmEl/ - I'll try again from scratch on CentOS 9 Stream. | 13:17 |
| jrosser | but thats a warning, not an error? | 13:29 |
| ThiagoCMC | Yes, it's only a warning. I'll reinstall the serves and try again, first AIO, then, multi-server. | 13:31 |
| jrosser | remember that you can look in any of the OSA CI jobs for centos-9 to see what happens | 13:33 |
| ThiagoCMC | Sure! Can you share the link? | 13:34 |
| jrosser | http://bit.ly/osa-review-board-v4_1 | 13:34 |
| ThiagoCMC | Thanks | 13:36 |
| ThiagoCMC | jrosser, I tried again from scratch, same problem (multi-node): https://paste.openstack.org/show/bJduUu4jPRSyQppV5mD3/ - I'll try again AIO. | 15:18 |
| jrosser | ThiagoCMC: now i look again `[WARNING]: Unable to parse /opt/openstack-ansible/inventory/dynamic_inventory.py as an inventory source` | 15:19 |
| jrosser | ^ not good | 15:19 |
| ThiagoCMC | Yeah, it's werid that exactly same procedure works on Ubuntu 20.04, but not on CentOS 9. | 15:19 |
| ThiagoCMC | Not sure how Ansible is even seeing that "got argparse"! | 15:20 |
| jrosser | well thats the first code in the script | 15:20 |
| jrosser | `import argparse` | 15:20 |
| ThiagoCMC | But it was supposed to be executed, so Ansible would only see its output, not its contents. | 15:21 |
| ThiagoCMC | Right? | 15:21 |
| jrosser | indeed | 15:21 |
| ThiagoCMC | Yeah, weird... lol | 15:21 |
| jrosser | not some selinux wierdness? | 15:22 |
| jrosser | where is your centos-9 from? the best i can do is try a centos-9 cloud image | 15:22 |
| ThiagoCMC | Hmm... Let me double-check. | 15:22 |
| ThiagoCMC | CentOS 9 Stream ISO full install | 15:23 |
| jrosser | hmm that won't really be the same | 15:23 |
| jrosser | well, or is not necessarily the same | 15:23 |
| ThiagoCMC | But I'm installing in bare metal machines with BMC, kind hard to use the cloud-based images. | 15:25 |
| ThiagoCMC | And it isn't the SELINUX, it's disabled. | 15:25 |
| jrosser | oh sure yes, i mean more that i have no centos at all so the best i could to to look at it is try a cloud image | 15:25 |
| ThiagoCMC | It's okay, don't worry about it... Thanks! ^_^ | 15:26 |
| ThiagoCMC | With Ubuntu, everything just works... lol | 15:27 |
| ThiagoCMC | But I need to learn this on CentOS too. :-/ | 15:27 |
| jrosser | ThiagoCMC: sorry :( https://paste.opendev.org/show/bHPrGL4839dl28XRWNep/ | 15:59 |
| ThiagoCMC | Yeah, the problem seems to be with multi-node, not AIO. I'll keep trying... | 16:03 |
| ThiagoCMC | Perhaps OSA AIO should actually be a multi-node but, all-in-one. If you know what I mean... lol | 16:04 |
| jrosser | it could be an error in your inventory that only appears with python3.9 | 16:04 |
| jrosser | well there is a multinode-aio in the openstack-ansible-ops repo | 16:04 |
| ThiagoCMC | Hmm... | 16:04 |
| jrosser | like data error | 16:05 |
| jrosser | and tbh you should be able to boostrap a vm as a deploy node and drop the /etc/openstack_deploy in and just try the ping playbook | 16:05 |
| jrosser | it should not matter if it's actually a multinode deploy or not just to try --list-hosts | 16:06 |
| ThiagoCMC | I'm kind of doing that! I have a `/etc/openstack_deploy` as a tarball, that I can easily use to "setup-everything" in Ubuntu 20.04, over and over. | 16:10 |
| ThiagoCMC | Howver, when I use the same "openstack_deploy tarball in CentOS 9, it fails. | 16:10 |
| ThiagoCMC | I appreciate your input! I'll keep digging :-) | 16:11 |
| *** dviroel|rover is now known as dviroel|rover|lunch | 16:15 | |
| *** dviroel|rover|lunch is now known as dviroel|rover | 17:20 | |
| opendevreview | Jonathan Rosser proposed openstack/ansible-role-uwsgi master: Allow extra arguments to be passed to the uwsgi application through pyargv https://review.opendev.org/c/openstack/ansible-role-uwsgi/+/872195 | 17:36 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-os_neutron master: Ensure that all neutron config files are used when deployed with uwsgi https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/872199 | 17:48 |
| opendevreview | Jonathan Rosser proposed openstack/ansible-role-uwsgi master: Allow extra arguments to be passed to the uwsgi application through pyargv https://review.opendev.org/c/openstack/ansible-role-uwsgi/+/872195 | 18:42 |
| jrosser | well why would the hash of https://github.com/jhunt/hatop/archive/v0.8.0.tar.gz have changed all of a sudden | 19:24 |
| mgariepy | it changed? | 19:32 |
| jrosser | my patches are all failing CI on a sha mismatch | 19:36 |
| jrosser | and I have a version from June on my laptop by chance which has the correct sha | 19:36 |
| jrosser | and download it today the sha is actually different | 19:37 |
| jrosser | mgariepy: it says `The checksum for /opt/cache/files/v0.8.0.tar.gz did not match bcdab1664358ec83027957df11bbeb322df1a96d414a3ccc4e211532b82c4ad2; it was 74c65b1e3e4587f7ab2b54100d4acedeb250e5a06744182b1ca4f90dbdfdc980` | 19:42 |
| mgariepy | interesting.. did they change the lib that generate the tar-gz ? | 20:45 |
| jrosser | could be yes | 21:29 |
| jrosser | i compared the contents and they look identical | 21:30 |
| *** dviroel|rover is now known as dviroel|out | 21:54 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!