| @mhuin:matrix.org | Hey there, I've got a quick question about nodepool: it looks like the drivers engine offers the possibility to look for drivers in extra paths besides /path/to/nodepool/driver/: https://opendev.org/zuul/nodepool/src/branch/master/nodepool/driver/__init__.py#L60 - this is even tested in https://opendev.org/zuul/nodepool/src/branch/master/nodepool/tests/unit/test_drivers.py#L26 - but this feature isn't exposed? There's no way to tell nodepool via its config to look for extra drivers elsewhere? | 10:27 |
|---|---|---|
| @jim:acmegating.com | mhu: correct, nodepool does not support external drivers | 13:38 |
| @mhuin:matrix.org | I guess it was the plan to support them at some point and the possibility just stayed there? | 14:07 |
| @jim:acmegating.com | no it was never the plan | 14:07 |
| @tristanc_:matrix.org | Well I added this drivers_paths, hoping that it could be an option, but I remember we said we should not make any sort of guarantee with regard to a plugin API and it was in the user interest to disable out of tree drivers. | 14:18 |
| @mhuin:matrix.org | thanks both of you, that makes things clearer | 14:19 |
| @jangutter:matrix.org | I wonder if one can make a generalisation here: third party plugins of any form tend to cause the majority of maintenance cost, failures and hair loss. And in a dynamic language it's a nightmare. | 15:07 |
| @jangutter:matrix.org | I say that having contributed to some and still feeling a like now I owe modern civilization some recompence. | 15:09 |
| @mhuin:matrix.org | your pain was from the standpoint of a third party plugin developer right? | 15:10 |
| @jangutter:matrix.org | no, from the standpoint of inflicting a maintenance burden on our users.... | 15:10 |
| @mhuin:matrix.org | or from the standpoint of the maintainer of a project allowing third party plugins? | 15:11 |
| @jangutter:matrix.org | our hardware and software drivers had some sharp edges. | 15:11 |
| @jangutter:matrix.org | If any vendor asks you to a) install this unsigned package, b) install this kernel module or c) recompile your kernel with this patch, that's usually a signal that you're going to help them debug stuff. | 15:13 |
| @mhuin:matrix.org | totally unrelated question, is there a branch/review already up somewhere with zuul-launcher's future documentation, or is the plan to write it later when the component is deemed ready for showtime? | 16:02 |
| -@gerrit:opendev.org- Clark Boylan proposed: [zuul/zuul] 961817: Add debug logging for openstack network lookups https://review.opendev.org/c/zuul/zuul/+/961817 | 16:12 | |
| @clarkb:matrix.org | I don't think it has been written yet as we're still working through the rough edges (see the change I just pushed) | 16:12 |
| @jim:acmegating.com | yeah, the docs will come near the end | 16:13 |
| @mhuin:matrix.org | good to know, as always the code is the most up to date doc of course | 16:15 |
| @jim:acmegating.com | well, we're also saying no one should run it yet :) | 16:15 |
| @mhuin:matrix.org | even good-er to know | 16:16 |
| @jim:acmegating.com | the internals are very much in flux, and i'm still making changes that require manual zk surgery to keep opendev running | 16:17 |
| @mhuin:matrix.org | so despite running it on opendev you consider it too early for others to make the switch? | 16:17 |
| @jim:acmegating.com | yep due to previous comment :) | 16:17 |
| @mhuin:matrix.org | ok ... we're willing to offer some of our deployments as guinea pigs but I think we put our foot down at zk surgery | 16:18 |
| @mhuin:matrix.org | maybe later then | 16:18 |
| @jim:acmegating.com | yeah, there will be a period where we think it's stable enough for that, while nodepool is still supported, and feedback then would be useful. once we don't actually *expect* the internals and config syntax to be completely changed. :) | 16:19 |
| @mhuin:matrix.org | just say the word and we will break our production envs, for the sake of SCIENCE! | 16:23 |
| @mhuin:matrix.org | I mean we'll test and provide feedback in controlled environments | 16:24 |
| @jim:acmegating.com | will do! | 16:30 |
| -@gerrit:opendev.org- Clark Boylan proposed: [zuul/zuul] 961817: Add debug logging for openstack network lookups https://review.opendev.org/c/zuul/zuul/+/961817 | 18:42 | |
| @clarkb:matrix.org | corvus: ^ the existing Exception in that method tripped me up as that isn't a valid attribute I guess. I've gone ahead and fixed my debug line entry and then also updated the Exception to avoid confusion | 18:43 |
| @jim:acmegating.com | Clark: ack, but if you want to fix that, i suggested an alternate | 19:11 |
| @clarkb:matrix.org | corvus: updating now | 19:17 |
| -@gerrit:opendev.org- Clark Boylan proposed: [zuul/zuul] 961817: Add debug logging for openstack network lookups https://review.opendev.org/c/zuul/zuul/+/961817 | 19:18 | |
| @clarkb:matrix.org | corvus: ^ updated | 19:21 |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed on behalf of Clark Boylan: [zuul/zuul] 961817: Add debug logging for openstack network lookups https://review.opendev.org/c/zuul/zuul/+/961817 | 19:29 | |
| @clarkb:matrix.org | corvus: that latest patchset is failing on an image upload wait for artifacts to settle timeout | 20:38 |
| @clarkb:matrix.org | I have to pop out soon to do a school run. Do you want me to dequeue then reenqueue or should we just wait and reapprove? | 20:39 |
| @jim:acmegating.com | dequeue is fine, not sure if you left, i'll do it | 21:00 |
| @clarkb:matrix.org | Ya sorry on my way out now | 21:01 |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul] 961760: Launcher: move state machine info to separate znodes https://review.opendev.org/c/zuul/zuul/+/961760 | 21:56 | |
| -@gerrit:opendev.org- Zuul merged on behalf of Clark Boylan: [zuul/zuul] 961817: Add debug logging for openstack network lookups https://review.opendev.org/c/zuul/zuul/+/961817 | 22:35 | |
| @clarkb:matrix.org | corvus: I guess at this point we can let the weekly updates pick that up for us? | 22:51 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!