*** tosky has quit IRC | 00:44 | |
*** piotrowskim has quit IRC | 01:28 | |
*** zenkuro has quit IRC | 01:44 | |
*** jamesmcarthur has joined #zuul | 03:02 | |
*** jamesmcarthur has quit IRC | 04:19 | |
*** jamesmcarthur has joined #zuul | 04:27 | |
*** jamesmcarthur has quit IRC | 04:32 | |
*** jamesmcarthur has joined #zuul | 04:43 | |
openstackgerrit | James E. Blair proposed zuul/nodepool master: Add a state machine driver framework https://review.opendev.org/c/zuul/nodepool/+/779071 | 04:44 |
---|---|---|
openstackgerrit | James E. Blair proposed zuul/nodepool master: Add an example state machine driver https://review.opendev.org/c/zuul/nodepool/+/779072 | 04:45 |
corvus | mordred, tobiash, avass: ^ that's a WIP idea that i have for a driver framework that tries to incorporate everything we've learned over the past few years; it's inspired by tobiash's de-threading work and the simple driver interface i did for gce that isolates cloud implementation from nodepool logic. i think we can accomodate all of the clouds with that framework, which means more shared code that's easier | 04:48 |
corvus | to understand and maintain. | 04:48 |
corvus | mordred, tobiash, avass: the easiest way to check it out is probably to look at adapter.py in 779072. that's a complete (no-op) driver implementation using it. then take a look at the bottom part of statemachine.py in 779071 for the interface documentation (the top half is not user-servicable). | 04:52 |
*** jamesmcarthur has quit IRC | 05:00 | |
*** evrardjp has quit IRC | 05:33 | |
*** evrardjp has joined #zuul | 05:33 | |
*** bhavikdbavishi has joined #zuul | 05:56 | |
openstackgerrit | Merged zuul/nodepool master: Offload waiting for server creation/deletion https://review.opendev.org/c/zuul/nodepool/+/775438 | 06:09 |
openstackgerrit | Merged zuul/nodepool master: Optimize list_servers call in cleanupLeakedInstances https://review.opendev.org/c/zuul/nodepool/+/775796 | 06:14 |
*** jamesmcarthur has joined #zuul | 07:01 | |
*** jamesmcarthur has quit IRC | 07:05 | |
*** bhavikdbavishi1 has joined #zuul | 07:16 | |
*** bhavikdbavishi has quit IRC | 07:18 | |
*** bhavikdbavishi1 is now known as bhavikdbavishi | 07:18 | |
*** harrymichal has joined #zuul | 07:20 | |
*** harrymichal has quit IRC | 07:25 | |
*** bhavikdbavishi has quit IRC | 08:06 | |
openstackgerrit | Albin Vass proposed zuul/zuul master: Use shell-type config from nodepool https://review.opendev.org/c/zuul/zuul/+/775382 | 10:17 |
*** jangutter has quit IRC | 10:41 | |
*** jangutter has joined #zuul | 10:41 | |
*** jangutter has quit IRC | 10:43 | |
*** jangutter has joined #zuul | 10:44 | |
*** jangutter has quit IRC | 11:06 | |
*** jangutter has joined #zuul | 11:07 | |
*** jangutter has quit IRC | 11:29 | |
*** jangutter has joined #zuul | 11:30 | |
avass | zuul-client doesn't seem to work very well with single tenant zuul installations :/ | 11:31 |
openstackgerrit | Sorin Sbârnea proposed zuul/zuul-jobs master: Upgrade ansible-lint to 5.0 https://review.opendev.org/c/zuul/zuul-jobs/+/773245 | 12:04 |
*** tosky has joined #zuul | 13:06 | |
*** bhavikdbavishi has joined #zuul | 13:33 | |
*** jangutter has quit IRC | 13:33 | |
*** jangutter has joined #zuul | 13:34 | |
*** jangutter has quit IRC | 13:43 | |
*** jangutter has joined #zuul | 13:44 | |
openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: Zuul Cache role with s3 implementation. https://review.opendev.org/c/zuul/zuul-jobs/+/764808 | 13:51 |
openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: Zuul Cache role with s3 implementation. https://review.opendev.org/c/zuul/zuul-jobs/+/764808 | 13:53 |
*** bhavikdbavishi has quit IRC | 14:02 | |
*** jangutter has quit IRC | 14:16 | |
*** jangutter has joined #zuul | 14:17 | |
openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: Zuul Cache role with s3 implementation. https://review.opendev.org/c/zuul/zuul-jobs/+/764808 | 14:20 |
corvus | avass: how so? | 14:21 |
avass | corvus: I think half of it was sort of my fault since I was routing /api/tenant/<tenant> wrong (it went to /api/tenant/<tenant>/tenant/<tenant) but I was still required to supply a --tenant flag | 14:38 |
avass | so imo if --tenant isn't supplied it should try /api/<project>.pub when trying to encrypt for example | 14:39 |
avass | so I ended up fetching the public key with curl and encrypting, but then I was still required to supply --zuul-url even though it was a local operation. | 14:40 |
corvus | avass: i agree. regarding tenant auto detection, we provide the /info endpoint so programs can do that, so should be possible | 14:46 |
corvus | mhu: ^ fyi | 14:46 |
corvus | avass: we are talking about a whitelabel setup, right? | 14:47 |
avass | yeah | 14:47 |
avass | https://zuul.vassast.org | 14:47 |
corvus | thought so, just checking :) | 14:47 |
openstackgerrit | James E. Blair proposed zuul/zuul master: Use ZooKeeper TLS in tests https://review.opendev.org/c/zuul/zuul/+/777489 | 15:00 |
openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: Zuul Cache role with s3 implementation. https://review.opendev.org/c/zuul/zuul-jobs/+/764808 | 15:06 |
avass | corvus: found a nice way to batch uploads without resorting to python scripts ^ :) | 15:07 |
corvus | avass: ah yep! is there no upper limit on the number in parallel? | 15:16 |
corvus | oh there i see it batch(10) | 15:16 |
avass | yeah :) | 15:16 |
openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: Zuul Cache role with s3 implementation. https://review.opendev.org/c/zuul/zuul-jobs/+/764808 | 15:17 |
avass | tests were successful, just needed to remove some debug tasks. so that should now be ready to review | 15:18 |
openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: Zuul Cache role with s3 implementation. https://review.opendev.org/c/zuul/zuul-jobs/+/764808 | 15:21 |
mordred | corvus, tobiash: I +2'd but did not +A the global repo state patch - wasn't sure if you wanted anyone else to look at it too | 15:49 |
*** jangutter has quit IRC | 15:50 | |
*** jangutter has joined #zuul | 15:51 | |
mordred | corvus: on the statemachine change - looking nice! should the machines have an unwind() method (or similar?) or is the idea that a CreateStateMachine that fails halfway through should have any intermediate resources cleaned up by the cleanup thread? | 15:59 |
mordred | corvus: or maybe the higher level is going to call cleanupNode for a node launch failure - so the delete state machine should be written such that it could also delete partially created nodes? | 16:02 |
mordred | corvus: also - I left a couple of comments on the example change | 16:10 |
mordred | I really like that whole approach | 16:10 |
*** tosky has quit IRC | 16:12 | |
*** vishalmanchanda has quit IRC | 16:22 | |
corvus | mordred: the latter -- my thought is the state machine should set external_id as early as possible, and if it exists, then the higher level will hand it to a DeleteStateMachine. if it doesn't, then anything created will have to be cleaned up by the leak cleanup. i was also thinking we should handle external_id changing. so, for example, if a really bad cloud didn't have any metadata support, maybe | 16:40 |
corvus | external_id could be a json serialization of uuids of things as they are created (eg: "{fip: uuid1}" ... "{fui: uuid1, port:uuid2}" ... etc, until it finally gets to a vm uuid that ties it all together) | 16:40 |
corvus | mordred: (i had an earlier version where i wrote an unwind method but it didn't feel right sitting right next to a DeleteStateMachine) | 16:42 |
corvus | also, as currently written, the state machine will have to internally handle retries, but i think we shoud move that into the higher level (it will involve retrying a node launch as a new node) | 16:44 |
corvus | (that's not too hard, it's almost what we do right now in case we hit a quota limit) | 16:44 |
mordred | ++ | 16:44 |
openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: Zuul Cache role with s3 implementation. https://review.opendev.org/c/zuul/zuul-jobs/+/764808 | 16:52 |
openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: Zuul Cache role with s3 implementation. https://review.opendev.org/c/zuul/zuul-jobs/+/764808 | 16:55 |
*** zenkuro has joined #zuul | 17:24 | |
openstackgerrit | James E. Blair proposed zuul/zuul master: Use ZooKeeper TLS in tests https://review.opendev.org/c/zuul/zuul/+/777489 | 17:57 |
avass | err, I got some weird zuul-web console bug, try opening the last of the lists in the archived files and then try to close it: https://zuul.vassast.org/build/fe46800c203f4439859463cbe6e6fe08/console | 18:35 |
avass | when I try to close the list it instead opens the items from 2300 to 2400 | 18:36 |
avass | closing that one opens the items 2400 to 2500 | 18:36 |
corvus | avass: which task? | 18:38 |
corvus | avass: this one? https://zuul.vassast.org/build/fe46800c203f4439859463cbe6e6fe08/console#3/0/3/fedora | 18:39 |
avass | oh yep, I got the wrong link | 18:40 |
corvus | avass: i get weird similar (but not exactly the same) behavior. i can't close it, but attempting to opens the earliest unopened list | 18:41 |
corvus | avass: we just upgraded that component recently to deal with the string-sorted array index issue, wonder if a new bug crept into that release | 18:42 |
avass | yeah i'm either getting a random list opened to the one i'm trying to close is closed | 18:42 |
avass | corvus: yeah I was thinking the same | 18:42 |
corvus | avass: heh, it seems like they didn't test with numbers >10 before, what are the chances they test with numbers > 3500? :) | 18:42 |
avass | :D | 18:43 |
avass | who needs more than 10 numbers anyway | 18:43 |
corvus | yep: 0, 1. done. | 18:57 |
openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: zuul-cache-s3: Allow setting compression level https://review.opendev.org/c/zuul/zuul-jobs/+/779081 | 18:58 |
*** tosky has joined #zuul | 19:02 | |
*** jamesmcarthur has joined #zuul | 19:27 | |
*** jamesmcarthur has quit IRC | 21:02 | |
*** jamesmcarthur has joined #zuul | 21:03 | |
*** jamesmcarthur has quit IRC | 21:08 | |
openstackgerrit | James E. Blair proposed zuul/zuul master: Use ZooKeeper TLS in tests https://review.opendev.org/c/zuul/zuul/+/777489 | 21:20 |
openstackgerrit | James E. Blair proposed zuul/zuul master: Make ConnectionRegistry mandatory for Scheduler https://review.opendev.org/c/zuul/zuul/+/757095 | 21:20 |
openstackgerrit | James E. Blair proposed zuul/zuul master: Instantiate executor client, merger, nodepool and app within Scheduler https://review.opendev.org/c/zuul/zuul/+/757149 | 21:21 |
openstackgerrit | James E. Blair proposed zuul/zuul master: Move setupZK() helper function to BaseTestCase class https://review.opendev.org/c/zuul/zuul/+/779082 | 21:21 |
openstackgerrit | James E. Blair proposed zuul/zuul master: Initialize ZooKeeper connection in server rather than in cmd classes https://review.opendev.org/c/zuul/zuul/+/779083 | 21:21 |
*** jamesmcarthur has joined #zuul | 21:21 | |
*** jamesmcarthur has quit IRC | 21:33 | |
*** jamesmcarthur has joined #zuul | 21:34 | |
*** tosky has quit IRC | 21:42 | |
openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: Prepend ansible_user_dir when relative path is supplied https://review.opendev.org/c/zuul/zuul-jobs/+/779084 | 22:19 |
*** jamesmcarthur has quit IRC | 22:21 | |
*** jamesmcarthur has joined #zuul | 22:35 | |
*** jamesmcarthur has quit IRC | 23:12 | |
avass | nice. with the zuul-cache our rust project build went from 25 min to 2 min 50 sec | 23:18 |
*** jamesmcarthur has joined #zuul | 23:43 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!