| *** gyee has quit IRC | 00:07 | |
| *** renich has joined #openstack-swift | 01:14 | |
| *** nanzha has joined #openstack-swift | 01:14 | |
| *** diablo_rojo has quit IRC | 02:18 | |
| *** diablo_rojo has joined #openstack-swift | 02:33 | |
| *** tkajinam has quit IRC | 03:04 | |
| *** diablo_rojo has quit IRC | 03:05 | |
| *** psachin has joined #openstack-swift | 03:33 | |
| *** tkajinam has joined #openstack-swift | 04:14 | |
| *** nanzha has quit IRC | 05:30 | |
| *** nanzha has joined #openstack-swift | 05:31 | |
| *** zaitcev_ has joined #openstack-swift | 05:51 | |
| *** ChanServ sets mode: +v zaitcev_ | 05:51 | |
| *** zaitcev has quit IRC | 05:55 | |
| *** ccamacho has quit IRC | 06:39 | |
| *** tesseract has joined #openstack-swift | 06:45 | |
| *** renich has quit IRC | 06:56 | |
| *** nanzha has quit IRC | 06:57 | |
| *** nanzha has joined #openstack-swift | 07:09 | |
| openstackgerrit | Thiago da Silva proposed openstack/swift master: Set new lower-bound for several packages https://review.opendev.org/692124 | 07:15 |
|---|---|---|
| *** pcaruana has joined #openstack-swift | 07:50 | |
| *** nanzha has quit IRC | 08:02 | |
| *** nanzha has joined #openstack-swift | 08:14 | |
| *** rdejoux has joined #openstack-swift | 08:31 | |
| *** tkajinam has quit IRC | 08:39 | |
| *** ccamacho has joined #openstack-swift | 08:51 | |
| *** rpittau|afk is now known as rpittau | 08:52 | |
| *** mikecmpbll has joined #openstack-swift | 09:00 | |
| *** ccamacho has quit IRC | 09:09 | |
| *** nanzha has quit IRC | 09:57 | |
| *** nanzha has joined #openstack-swift | 09:57 | |
| *** rpittau is now known as rpittau|bbl | 11:00 | |
| *** tkajinam has joined #openstack-swift | 11:17 | |
| *** ccamacho has joined #openstack-swift | 11:30 | |
| *** FlorianFa has joined #openstack-swift | 12:13 | |
| openstackgerrit | Thiago da Silva proposed openstack/swift master: Set new lower-bound for several packages https://review.opendev.org/692124 | 12:49 |
| *** tkajinam has quit IRC | 13:06 | |
| *** rpittau|bbl is now known as rpittau | 13:16 | |
| *** psachin has quit IRC | 13:35 | |
| *** nanzha has quit IRC | 15:05 | |
| *** nanzha has joined #openstack-swift | 15:06 | |
| *** renich has joined #openstack-swift | 15:12 | |
| *** FlorianFa has quit IRC | 15:54 | |
| *** mikecmpbll has quit IRC | 16:01 | |
| *** mikecmpbll has joined #openstack-swift | 16:02 | |
| openstackgerrit | Thiago da Silva proposed openstack/swift master: WIP: New Object Versioning mode https://review.opendev.org/682382 | 16:14 |
| *** nanzha has quit IRC | 16:20 | |
| *** rdejoux has quit IRC | 16:21 | |
| openstackgerrit | Gilles Biannic proposed openstack/swift master: Stop retrying every deletes in container-sync https://review.opendev.org/691729 | 16:22 |
| *** tesseract has quit IRC | 16:33 | |
| *** renich has quit IRC | 16:35 | |
| *** renich has joined #openstack-swift | 16:36 | |
| *** gyee has joined #openstack-swift | 16:50 | |
| *** mikecmpbll has quit IRC | 16:55 | |
| *** mikecmpbll has joined #openstack-swift | 16:58 | |
| *** rpittau is now known as rpittau|afk | 17:14 | |
| *** mikecmpbll has quit IRC | 17:33 | |
| *** renich has quit IRC | 18:07 | |
| clayg | timburke: I disagree with your suggestion to return a list to indicate an "acceptable" http error was raised - we now have a function that will return None, a tuple, or raise at least 4 different kinds of exceptions that we know about | 18:09 |
| clayg | timburke: Also "acceptable_statuses" is a terrible name for this new behavior - since any is_success status is allowed in *this* context by default SimpleClient now has an interface that *smells* like InternalClient's acceptable_statuses but is wildly different | 18:11 |
| timburke | eh? we either return a list (because that's the only thing base_request returns) or we raise -- where are you seeing None? | 18:12 |
| clayg | check the default values for acceptable_statuses on e.g. InternalClient.get_account_info - just sayin' cc gmann | 18:12 |
| timburke | clayg, i don't think that's Gilles ;-) | 18:12 |
| clayg | oops! | 18:12 |
| clayg | i see, we never fall out of the loop - yeah that helps | 18:13 |
| timburke | i *do* think that we should have a way to say "these statuses are not exceptional" though, very much like internal_client. i originally thought that it *did* work like internal_client (in part because we're *in the internal_client module*!) and was disappointed that it didn't | 18:14 |
| clayg | that loop should probably read `while True:` - i don't know what the type of err.info() vs. conn.info() | 18:14 |
| timburke | both HTTPMessages -- basically, all the headers from the response (but not the status code :-( ) | 18:15 |
| clayg | I think simple client was just thin enough that we didn't want to create a seperate module to hang it off at the time | 18:15 |
| clayg | lol @ not the status code | 18:15 |
| timburke | i think we could set it easily enough, and in both cases | 18:16 |
| timburke | *separately* (and maybe this would be enough for Gilles's problem) we should *almost certainly* prevent SimpleClient from retrying on *client* errors | 18:17 |
| clayg | I don't think we're going to make SimpleClient.retry_request work anything like InternalClient.make_request - not sure why we latched onto acceptable_statuses as a concept except that we're in the same module and didn't realize we have different classes/interfaces | 18:17 |
| clayg | oh, sure I like that pleanty - just have it raise 4XX immediately and only do backoff on 5XX and other kinds of http/socket errors 👍 | 18:18 |
| clayg | so with splitting the the HTTPError - then all we have to do is `if attempts > retries or is_client_error(e): raise` - no new kwarg | 18:19 |
| timburke | and the 409 behavior could be a separate patch | 18:19 |
| *** mvkr has quit IRC | 18:20 | |
| clayg | I see no reason to add the string to pull - i may not understand the situation with 409 - should it not get re-raised immediately like 404? | 18:20 |
| clayg | in sync we just say `if err.http_status in (...)`? | 18:21 |
| timburke | 👍 | 18:21 |
| timburke | i was mainly just thinking that they seem like orthogonal fixes -- i probably shouldn't have brought it up on the bug, was just kinda thinking out loud | 18:22 |
| clayg | I'm having a bit of trouble with `test/probe/test_object_expirer.py:TestObjectExpirer.test_expirer_doesnt_make_async_pendings` - anyone in a environment to spot check master, or have seen that recently? I haven't looked into it. Might be evironmental. | 18:26 |
| clayg | I don't really see the 404 vs 409 as orthogonal - I think you hit the nail on the head - retry_request shouldn't be using the same backoff/retry logic for socket errors that it's using for client errors | 18:28 |
| clayg | how interesting that it doesn't have any authentication handling - is that because of something in realms!? | 18:29 |
| clayg | WHO USES CONTAINER SYNC 🤣 | 18:29 |
| *** renich has joined #openstack-swift | 18:32 | |
| timburke | i was thinking more about the fact that 409 is treated as a failure and prevents sync_point2 from advancing -- that's orthogonal from the retry behavior issue | 18:37 |
| *** paladox has quit IRC | 18:52 | |
| *** paladox has joined #openstack-swift | 18:55 | |
| *** diablo_rojo has joined #openstack-swift | 19:51 | |
| rledisez | clayg: we do use container-sync, and we are going to work to improve it cause customers are asking for it (note: Gilles is from my team) | 19:55 |
| timburke | quick question -- did we *always* have py27 support? i know we supported py26 through 2.3.0, but the earliest mention i see in the changelog for py27 was way back in 1.4.0 and it just says "Fixed a Python 2.7 compatibility problem." | 19:55 |
| timburke | i guess my question is, was the scope of that problem "absolutely nothing works"? 🤣 | 19:56 |
| timburke | rledisez, you should get him to come on irc :-) | 19:59 |
| rledisez | timburke: i'll tell him :) | 20:01 |
| clayg | rledisez: 🤗 | 20:07 |
| *** diablo_rojo has quit IRC | 20:31 | |
| mattoliverau | Did we ever land those container sync improvements from Eran years ago? | 20:40 |
| *** pcaruana has quit IRC | 21:31 | |
| mattoliverau | morning | 21:49 |
| *** renich has quit IRC | 21:55 | |
| *** renich has joined #openstack-swift | 22:32 | |
| *** threestrands has joined #openstack-swift | 22:42 | |
| *** tkajinam has joined #openstack-swift | 23:00 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!