| opendevreview | Takashi Kajinami proposed openstack/puppet-openstack-integration master: Disable mariadb:10.11 module https://review.opendev.org/c/openstack/puppet-openstack-integration/+/909064 | 03:39 |
|---|---|---|
| opendevreview | Takashi Kajinami proposed openstack/puppet-openstack-integration master: Disable mariadb:10.11 module https://review.opendev.org/c/openstack/puppet-openstack-integration/+/909064 | 04:23 |
| zigo | tkajinam: Hi man! How are you? | 07:49 |
| zigo | Ahead of the swift patch here: https://review.opendev.org/c/openstack/swift/+/907523 | 07:49 |
| zigo | I started writing that one: https://review.opendev.org/c/openstack/puppet-swift/+/909004 Any idea why this last one fails its puppet-openstack-litmus-7.27-centos-9-stream job? I don't get it. | 07:49 |
| zigo | It feels unrelated to me. | 07:49 |
| tkajinam | that is another fun centos gives us :-) | 08:05 |
| tkajinam | they fixed one bug in libvirt but released a new bug in mariadb :-) | 08:05 |
| zigo | Oh ok. :) | 08:05 |
| zigo | So I can consider my patch to be working then? :) | 08:06 |
| zigo | (and I can write further ...) | 08:06 |
| zigo | tkajinam: Another point: currently, there's no mountpoints fact, because a Ruby package seems missing. | 08:07 |
| tkajinam | I guess so based on the unit tests added but I may add acceptance test to run these resources actually. | 08:07 |
| zigo | Should I write a custom fact to get the free space available for mount points? | 08:07 |
| zigo | Yeah, that's not completely finished, will do. | 08:07 |
| tkajinam | do we need any logic in the module side dependent on mount fact ? | 08:07 |
| tkajinam | I thought all checks should be done by that tooling. | 08:07 |
| tkajinam | I mean, we do not re-implement the same check but use the tool shipped by swift | 08:08 |
| zigo | Well, you don't want puppet to revert what the disk-full-checker cron will do. | 08:08 |
| zigo | If you believe we shouldn't re-do the work, ok, but I have no idea how we can avoid it. | 08:08 |
| tkajinam | you can generate a file in a temp path and run the tool over that file and install it to the final destination | 08:09 |
| tkajinam | this is how I'd implement that logic | 08:09 |
| tkajinam | having the same logic in two places are just adding another maintenance hell | 08:09 |
| zigo | Agreed. | 08:09 |
| zigo | Though the swift tool currently doesn't parse the CLI, only the config file. So I'll have to add such feature. | 08:10 |
| zigo | Will do that, that's IMO a very good idea. | 08:10 |
| zigo | Thanks for sharing. | 08:10 |
| tkajinam | it'd be probably easier if the tool can read template and generate the final file | 08:11 |
| tkajinam | I mean, use different path for source and destination | 08:11 |
| tkajinam | then we can use that in deployment tools more easily, I guess | 08:11 |
| tkajinam | we may still have to check if the puppet resource we currently use allows us to put the file out of the rsync config directory | 08:12 |
| tkajinam | ok we may override rsync::server::conf_file | 08:16 |
| zigo | Yeah, that's what I thought too. It has multiple advantages to do that. | 08:21 |
| zigo | One of them being that drive-full-checker is modifying the number of spaces in each directive, which would fool puppet to believe the file has changed. | 08:22 |
| zigo | That way we avoid this problem. | 08:22 |
| zigo | The only thing I need, as you wrote, is adding a --source-file option in drive-full-checker ! :) | 08:23 |
| zigo | tkajinam: We added the --source-file option to the /usr/bin/drive-full-checker https://review.opendev.org/c/openstack/swift/+/907523 | 09:23 |
| zigo | tkajinam: In puppet-module-puppetlabs-rsync we don't have conf_file unless we use latest commit, the latest release 1.2.0 doesn't have conf_file as parameter of rsync::server | 12:51 |
| zigo | tkajinam: Can we base our thing on a commit, rather than a release of the module? | 12:58 |
| zigo | If so, how? | 13:08 |
| opendevreview | Thomas Goirand proposed openstack/puppet-swift master: Per daemon log name https://review.opendev.org/c/openstack/puppet-swift/+/909114 | 13:38 |
| *** finnekit8 is now known as finnekit | 15:49 | |
| opendevreview | Merged openstack/puppet-openstack-guide master: Remove puppet-qdr https://review.opendev.org/c/openstack/puppet-openstack-guide/+/907974 | 20:24 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!