| hberaud[m] | #startmeeting openstack-eventlet-removal | 15:00 |
|---|---|---|
| opendevmeet | Meeting started Tue Jan 21 15:00:03 2025 UTC and is due to finish in 60 minutes. The chair is hberaud[m]. 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_eventlet_removal' | 15:00 |
| hberaud[m] | Ping amorin, JayF, gouthamr, ralonsoh | 15:00 |
| amorin | o/ | 15:00 |
| hberaud[m] | #link https://etherpad.opendev.org/p/epoxy-eventlet-tracking | 15:00 |
| hberaud[m] | We are at line 140 | 15:00 |
| hberaud[m] | I'm waiting 1 or 2 minutes to allow folks to join | 15:01 |
| hberaud[m] | ok lets go | 15:03 |
| ralonsoh | hello (I was in the Neutron meeting) | 15:03 |
| hberaud[m] | #topic review task completion | 15:03 |
| hberaud[m] | no tasks from previous weeks | 15:03 |
| hberaud[m] | #topic updates | 15:04 |
| hberaud[m] | There is one great news, Neutron finished to implement their new socket server to replace the eventlet wsgi one https://review.opendev.org/c/openstack/neutron/+/937545 Kudos ralonsoh | 15:06 |
| ralonsoh | yes but needs to be approved | 15:06 |
| ralonsoh | at least we are replacing the metdata server | 15:06 |
| ralonsoh | now I'm trying to deal with the Neutron API issues since we migrate to wsgi | 15:06 |
| hberaud[m] | yes the base implementation seems already merged if I'm right | 15:06 |
| hberaud[m] | ack | 15:06 |
| ralonsoh | just a heads-up | 15:07 |
| ralonsoh | https://review.opendev.org/c/openstack/neutron/+/938659 | 15:07 |
| hberaud[m] | Great job Rodolfo | 15:07 |
| ralonsoh | we are running Neutron API without eventlet | 15:07 |
| ralonsoh | (but we have many problems with OVN) | 15:07 |
| hberaud[m] | oh nice | 15:07 |
| hberaud[m] | can we prefix all these series of patches with "eventlet-removal"? | 15:07 |
| hberaud[m] | as topic | 15:08 |
| ralonsoh | sure, I'll do it | 15:08 |
| hberaud[m] | it will allow them to appear here https://review.opendev.org/q/prefixtopic:%22eventlet-removal%22 | 15:08 |
| hberaud[m] | thanks | 15:08 |
| hberaud[m] | in parallel of that I started to deprecate more eventlet things in the oslo side by example futurist's green executor, oslo.utils's eventletutils etc... | 15:09 |
| hberaud[m] | And glance started to prepare some refactor of their functional tests without eventlet | 15:10 |
| hberaud[m] | We could also note that the requirements bump is now merged and openstack use the latest version of eventlet | 15:11 |
| hberaud[m] | this version land the support of python 3.13, so in theory OSP now support Python 3.13 | 15:11 |
| hberaud[m] | takashi proposed to drop the support of python 3.8, is this version is now EOL | 15:12 |
| hberaud[m] | but for now the patch is stuck by our gates | 15:13 |
| hberaud[m] | Do you have other updates that you want to share with us? | 15:13 |
| ralonsoh | no, thanks | 15:13 |
| amorin | nop | 15:14 |
| hberaud[m] | ack,thanks | 15:14 |
| hberaud[m] | #topic oslo.service's migration | 15:15 |
| hberaud[m] | concerning the migration of oslo.service, Daniel, Takashi, and Mike recently had an interesting discussion concerning the fact to use or not a config option to implement the backend selector | 15:16 |
| hberaud[m] | and finally they agreed to move from a config option to an env var to setup the backend in use in oslo.service | 15:17 |
| amorin | an env var? This may be difficult to implement on operators side | 15:18 |
| amorin | am I wrong? | 15:18 |
| hberaud[m] | good question | 15:19 |
| hberaud[m] | apparently a couple of other deliverables already function this way | 15:19 |
| hberaud[m] | https://review.opendev.org/c/openstack/oslo.service/+/935783 | 15:19 |
| hberaud[m] | the discussion started around patch set 26-27 | 15:20 |
| hberaud[m] | https://review.opendev.org/c/openstack/oslo.service/+/935783/comments/0f71b887_594e76a8 | 15:20 |
| hberaud[m] | concerning operators, I first glance I do not see the difficulty to define an env var, but I'm far to be an expert deployer... | 15:21 |
| amorin | I'll take a look at the patchset | 15:22 |
| amorin | thanks | 15:22 |
| hberaud[m] | thanks | 15:22 |
| hberaud[m] | I think the development of this part of the migration of oslo.service is almost done, only remains cosmetic changes, I'll put a comment to see if we can address the remaining tiny changes into a follow-up patch | 15:23 |
| hberaud[m] | #topic open discussion | 15:24 |
| hberaud[m] | Anything else that you want to discuss today? | 15:24 |
| amorin | just for the record, I have not been able to work further with mistral eventlet removal, topic is paused on my side | 15:25 |
| hberaud[m] | ack, thanks | 15:25 |
| hberaud[m] | Then I think we are done | 15:27 |
| hberaud[m] | Thanks everyone | 15:27 |
| hberaud[m] | #endmeeting | 15:27 |
| opendevmeet | Meeting ended Tue Jan 21 15:27:12 2025 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 15:27 |
| opendevmeet | Minutes: https://meetings.opendev.org/meetings/openstack_eventlet_removal/2025/openstack_eventlet_removal.2025-01-21-15.00.html | 15:27 |
| opendevmeet | Minutes (text): https://meetings.opendev.org/meetings/openstack_eventlet_removal/2025/openstack_eventlet_removal.2025-01-21-15.00.txt | 15:27 |
| opendevmeet | Log: https://meetings.opendev.org/meetings/openstack_eventlet_removal/2025/openstack_eventlet_removal.2025-01-21-15.00.log.html | 15:27 |
| ralonsoh | bye | 15:27 |
| amorin | thansk, bye | 15:28 |
| JayF | hberaud[m]: amorin: I -1'd that patch, env vars for that use are just pain waiting to happen | 15:54 |
| JayF | I'm not a core so who knows if it'll be honored, but that seems like an operational headache I'd want no part of with either my Ironic dev or downstream operator hats on | 15:55 |
| amorin | I added a comment regarding this with exactly the same opinion | 15:55 |
| amorin | before -1 I'd like to know if some exemple exists somewhere | 15:55 |
| JayF | it shouldn't be configurable AT ALL | 15:55 |
| JayF | in CONF. or in env vars | 15:56 |
| amorin | that was said in the comemnts | 15:56 |
| amorin | why not in conf? | 15:56 |
| JayF | it's not an operator decision; it's a project decision | 15:56 |
| JayF | if I ship an ironic that depends on eventlet, why should I give a CONF.break_everything? | 15:56 |
| JayF | same from the post-migration perspective | 15:56 |
| amorin | so there is not point to have two backends then? Or maybe to let project switch on their own timeline | 15:56 |
| JayF | the project has to be the one to trigger a switch | 15:57 |
| JayF | not the operator | 15:57 |
| amorin | ok, make sense | 15:57 |
| JayF | CONF / env vars are operator facing, not upstraem-of-library facing | 15:57 |
| hberaud[m] | can you put your arguments in a comment of the review? | 15:57 |
| JayF | I already did | 15:58 |
| JayF | if it's not clear let me know and I can elaborate, I try not to writte books on PRs :) | 15:59 |
| hberaud[m] | I was just thinking that your "project decision" arguments were well expressed and useful and your actual comment do not reflect that | 16:01 |
| hberaud[m] | actual comment = the one in the PR | 16:01 |
| hberaud[m] | but as you want | 16:01 |
| JayF | I'll go put that wording on the comment then :) | 16:02 |
| hberaud[m] | thanks | 16:02 |
| JayF | I tend to ramble and am not always good at figuring out the useful bits from the stream of conciousness :D | 16:04 |
| JayF | so ty for the comment :) | 16:04 |
| itamarst | today in why eventlet needs to die: https://github.com/eventlet/eventlet/issues/1020#issuecomment-2605181748 | 16:21 |
| JayF | we are trying | 16:22 |
| itamarst | that's how this issue came about, yes | 16:24 |
| JayF | does someone share a channel with Mike to share that issue link with him? | 16:25 |
| JayF | I can't remember his IRC name off the top of my head | 16:25 |
| hberaud[m] | Mike? Bayer? | 16:26 |
| hberaud[m] | yes Mike Bayer | 16:26 |
| JayF | yeah I was like 75% sure that was his last name so omitted it to avoid the 1/4 failure :P | 16:26 |
| hberaud[m] | his irc name is zzzeek | 16:27 |
| hberaud[m] | and you can reach him on #openstack-oslo | 16:27 |
| itamarst | did someone share it, or should I? | 16:44 |
| hberaud[m] | itamarst: go ahead | 16:46 |
| itamarst | gonna do it privately, then | 16:58 |
| JayF | I linked it to him in the channel | 16:59 |
| JayF | more chat doesn't hurt :) | 16:59 |
| itamarst | ah ok | 16:59 |
| itamarst | I'll wait a bit | 16:59 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!