| *** mtreinish_ is now known as mtreinish | 00:22 | |
| *** mhen_ is now known as mhen | 01:09 | |
| tkajinam | stephenfin, yes and that's why I added -c pyproject.toml to bandit args defined in .pre-commit-config.yaml | 01:36 |
|---|---|---|
| amorin | hello, quick question regarding config options. Mistral rely on oslo.messaging, oslo.messaging rely on oslo_concurrency (lock_path). Should I have oslo_concurrency options in the config generator namespace list? | 09:19 |
| amorin | related to https://review.opendev.org/c/openstack/mistral/+/962890 | 09:19 |
| amorin | cc zigo | 09:19 |
| zigo | amorin: IMO, you should *if* Mistral is loading it, which wasn't the case previously. | 09:27 |
| amorin | the thing is, I may not be aware of all things that could be loaded by libraries, so for sure, we will always be missing some params | 09:28 |
| tkajinam | amorin, the load part may need to be fixed in a more core place (within oslo.concurrency or oslo.messaging) | 13:14 |
| tkajinam | amorin, regarding the options exposed in list_opts we can technically include oslo.concurrency options in oslo.messaging but that cause conflict with existing projects using both libraries. so I'm hesitant to do and we may need to update all generator config instead. | 13:15 |
| tkajinam | these generator config files have been maintained best-effort basis and may lack some options from indirect deps or even direct deps | 13:16 |
| tkajinam | what is strange is, although these options are registered during import these are read only when the relevant codes are actually run. I don't know what is causing that "read before load" behavior zigo mentioned | 13:18 |
| tkajinam | we really need clear bug report with details for further work. | 13:19 |
| tkajinam | only having that commit message does not give us full view about the problem and we can't judge if that fix is made in an appropriate place | 13:19 |
| tkajinam | we should not fix "a mystery" without clear understanding about its cause IMHO | 13:25 |
| * tkajinam haven't seen so far any hard code which may cause that nova path in oslo libs | 13:25 | |
| amorin | ok, so maybe for now, I should on register the oslo.concurrency namespace in mistral config, but thats all | 13:39 |
| opendevreview | Arnaud Morin proposed openstack/oslo.messaging master: Switch to another rabbitmq host on connection error https://review.opendev.org/c/openstack/oslo.messaging/+/962932 | 14:16 |
| tkajinam | amorin, you mean add oslo.concurrency namespace to config file for oslo-config-generator, right ? | 14:18 |
| amorin | yes, in mistral | 14:19 |
| tkajinam | I just wanted to make sure that you don't attempt some register_opts stuffs in the code | 14:19 |
| tkajinam | +1 :-) | 14:19 |
| amorin | zigo would you mind splitting your change in two? One to add the namespace in config, and second for your patch? At least the first one we will approve | 14:20 |
| dansmith | has anyone seen rootwrap failing to start in a devstack due to debtcollector being missing? | 16:42 |
| dansmith | https://paste.opendev.org/show/bJiyOLKYUrzvztCcqCS8/ | 16:43 |
| dansmith | its oslo privsep failing to spawn the helper with rootwrap, because the latter fails to import debtcollector, | 16:43 |
| dansmith | but it's clearly in the venv and the venv is being used for rootwrap | 16:44 |
| dansmith | I'm very confused | 16:44 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!