| zaitcev | timburke: I really need that sharder thing. https://review.openstack.org/#/c/570320/ | 01:17 |
|---|---|---|
| patchbot | patch 570320 - swift - py3: port the container sharder - 15 patch sets | 01:17 |
| openstackgerrit | Pete Zaitcev proposed openstack/swift master: py3: port proxy account controller https://review.openstack.org/637653 | 01:19 |
| timburke | zaitcev, i'm actively looking at it. still not sure that i like having names/bounds as bytes across the board... i feel like we should be able to use native strings since everything has a well-defined encoding at that point... i'll see what kinda of a patch i can get together before heading out tonight | 01:19 |
| zaitcev | timburke: I thought the separation of .lower and .lower_str asked for that representation. | 01:20 |
| openstackgerrit | Clay Gerrard proposed openstack/swift master: Simplify empty suffix handling https://review.openstack.org/637662 | 02:06 |
| clayg | lp bug #1816501 is pretty annoying, any EC cluster that delete's data and adds capacity should be seeing these empty handoff dirs getting stuck to drives | 02:19 |
| openstack | Launchpad bug 1816501 in OpenStack Object Storage (swift) "reconstructor doesn't remove empty handoff dirs with reclaimed tombstones" [Undecided,New] https://launchpad.net/bugs/1816501 | 02:19 |
| openstackgerrit | Tim Burke proposed openstack/swift master: Follow-up to sharder; port container backend and reconciler https://review.openstack.org/637663 | 02:21 |
| timburke | all right, i gotta go... i think it's a start, though? | 02:21 |
| *** psachin has joined #openstack-swift | 03:00 | |
| openstackgerrit | Pete Zaitcev proposed openstack/swift master: py3: port copy middleware https://review.openstack.org/635998 | 03:32 |
| *** psachin has quit IRC | 03:50 | |
| *** psachin has joined #openstack-swift | 04:00 | |
| kota_ | clayg: yup, I like p 637572. Perhaps, we should propose the change to the master tho. | 04:14 |
| patchbot | https://review.openstack.org/#/c/637572/ - swift (feature/losf) - Use more clear module imports - 1 patch set | 04:14 |
| kota_ | timburke: thanks for pointing out the fix for the probetests failure. I'll take a time to look at. | 04:15 |
| kota_ | OIC, that patch has been merged 9 hours ago, so I should make a merge commit from the master asap. | 04:15 |
| *** irclogbot_2 has quit IRC | 04:55 | |
| *** kei-ichi has joined #openstack-swift | 05:24 | |
| *** kei-ichi has quit IRC | 05:28 | |
| *** itlinux has quit IRC | 05:45 | |
| *** itlinux has joined #openstack-swift | 06:11 | |
| openstackgerrit | Merged openstack/swift feature/losf: Use more clear module imports https://review.openstack.org/637572 | 06:13 |
| *** itlinux has quit IRC | 06:16 | |
| *** e0ne has joined #openstack-swift | 06:25 | |
| *** e0ne has quit IRC | 06:26 | |
| *** ccamacho has quit IRC | 06:35 | |
| *** itlinux has joined #openstack-swift | 07:13 | |
| *** itlinux has quit IRC | 07:17 | |
| *** e0ne has joined #openstack-swift | 07:23 | |
| *** rcernin has quit IRC | 07:25 | |
| *** takamatsu has joined #openstack-swift | 07:40 | |
| *** ccamacho has joined #openstack-swift | 07:47 | |
| *** hseipp has joined #openstack-swift | 07:49 | |
| *** pcaruana has joined #openstack-swift | 08:11 | |
| *** tkajinam has quit IRC | 08:12 | |
| *** e0ne has quit IRC | 08:13 | |
| *** itlinux has joined #openstack-swift | 08:22 | |
| *** itlinux has quit IRC | 08:22 | |
| *** itlinux has joined #openstack-swift | 08:41 | |
| *** itlinux has quit IRC | 08:42 | |
| *** itlinux has joined #openstack-swift | 09:03 | |
| *** itlinux has quit IRC | 09:03 | |
| *** mikecmpbll has joined #openstack-swift | 09:04 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift feature/losf: Merge remote-tracking branch 'remotes/origin/master' into losf-feature https://review.openstack.org/637795 | 09:22 |
| *** ccamacho has quit IRC | 09:31 | |
| *** [diablo]5 has quit IRC | 09:42 | |
| *** e0ne has joined #openstack-swift | 09:53 | |
| *** mvkr has quit IRC | 12:09 | |
| *** itlinux has joined #openstack-swift | 13:12 | |
| *** mvkr has joined #openstack-swift | 13:13 | |
| *** itlinux has quit IRC | 13:21 | |
| *** itlinux has joined #openstack-swift | 13:22 | |
| *** itlinux has quit IRC | 13:28 | |
| *** itlinux has joined #openstack-swift | 13:32 | |
| *** itlinux has quit IRC | 13:37 | |
| *** ccamacho has joined #openstack-swift | 14:05 | |
| *** ccamacho has quit IRC | 14:18 | |
| *** ccamacho has joined #openstack-swift | 14:22 | |
| *** itlinux has joined #openstack-swift | 14:22 | |
| *** itlinux has quit IRC | 14:30 | |
| *** itlinux has joined #openstack-swift | 14:53 | |
| *** itlinux has quit IRC | 14:56 | |
| *** itlinux has joined #openstack-swift | 14:56 | |
| *** itlinux has quit IRC | 14:57 | |
| *** e0ne has quit IRC | 16:00 | |
| *** ianychoi has quit IRC | 16:00 | |
| *** e0ne has joined #openstack-swift | 16:03 | |
| *** ianychoi has joined #openstack-swift | 16:12 | |
| *** pcaruana has quit IRC | 16:57 | |
| openstackgerrit | Alexandre Lécuyer proposed openstack/swift feature/losf: Lots Of Small Files - alternate diskfile https://review.openstack.org/636578 | 17:10 |
| timburke | good morning | 17:16 |
| timburke | kota_, my patch didn't actually fix the particular failure -- i'll see about getting a patch up that should | 17:16 |
| timburke | in general, there are a whole bunch of tests that get flakey in a resource-constrained environment like the gate. we should certainly fix them as we figure them out, but idk that you necessarily need to merge right after -- whatever regular merge cadence you settle on for the branch is probably fine | 17:18 |
| notmyname | good morning | 17:24 |
| *** hseipp has quit IRC | 17:27 | |
| *** e0ne has quit IRC | 17:35 | |
| *** mikecmpbll has quit IRC | 17:36 | |
| *** psachin has quit IRC | 17:39 | |
| *** ccamacho has quit IRC | 17:46 | |
| zaitcev | timburke: Do you want me to pull parts of 637663 into 570320? You can rebase easily, right? | 17:55 |
| timburke | yeah, don't worry about my patch | 17:57 |
| timburke | sorry i forgot about the `raise` i put in there to help me sniff out where else we needed to patch out dump_recon_cache... | 17:58 |
| zaitcev | "don't worry" in what way? That sharder is a major hurdle and I take it you need the experiments in 637663 to resolve either way before you can approve 570320 (e.g. can you find a way to live without bounds2str, make .lower&.upper strings rather than bytes). | 18:00 |
| *** mvkr has quit IRC | 18:36 | |
| timburke | "don't worry" == take what you need and push up what makes sense. i can rebase. i *do* think we can get away without bounds2str, by making lower and upper native strings (or MIN/MAX) | 18:38 |
| *** e0ne has joined #openstack-swift | 18:41 | |
| *** e0ne has quit IRC | 18:42 | |
| *** ccamacho has joined #openstack-swift | 18:42 | |
| zaitcev | That would be nice. I'm still worried about str(). It's safer to get rid of them completely. | 18:43 |
| timburke | omg -- wtf are we doing?? https://github.com/openstack/swift/blob/2.20.0/test/unit/container/test_sync.py#L790-L792 | 18:52 |
| timburke | produces "TypeError: function() argument 1 must be code, not function" on py37 | 18:53 |
| *** e0ne has joined #openstack-swift | 18:54 | |
| timburke | because mock.patch is just *too hard* | 18:59 |
| *** mvkr has joined #openstack-swift | 19:07 | |
| openstackgerrit | Pete Zaitcev proposed openstack/swift master: py3: port the container sharder https://review.openstack.org/570320 | 19:15 |
| *** e0ne has quit IRC | 19:39 | |
| *** e0ne has joined #openstack-swift | 19:39 | |
| *** [diablo] has joined #openstack-swift | 19:41 | |
| *** e0ne has quit IRC | 19:45 | |
| openstackgerrit | Tim Burke proposed openstack/swift master: Follow-up to sharder; port rest of container https://review.openstack.org/637663 | 19:51 |
| *** openstackgerrit has quit IRC | 20:09 | |
| *** openstackgerrit has joined #openstack-swift | 20:12 | |
| openstackgerrit | Tim Burke proposed openstack/swift master: probe tests: wait to start replicators until after verifying initial state https://review.openstack.org/637986 | 20:12 |
| openstackgerrit | Tim Burke proposed openstack/swift master: py3: Follow-up to sharder; port rest of container https://review.openstack.org/637663 | 20:23 |
| *** e0ne has joined #openstack-swift | 20:37 | |
| openstackgerrit | Merged openstack/pyeclib master: Move experimental jobs to be gating https://review.openstack.org/636771 | 20:44 |
| openstackgerrit | Merged openstack/pyeclib master: Make TestBackendsEnabled work on py3 https://review.openstack.org/636787 | 20:46 |
| openstackgerrit | Tim Burke proposed openstack/swift master: py3: port versioned_writes https://review.openstack.org/638010 | 21:07 |
| *** e0ne has quit IRC | 21:28 | |
| openstackgerrit | Tim Burke proposed openstack/swift master: versioned_writes: stop trying to catch HTTPPreconditionFailed https://review.openstack.org/638015 | 21:29 |
| openstackgerrit | Tim Burke proposed openstack/swift master: Clean up how we walk through ranges in ECAppIter https://review.openstack.org/638019 | 21:54 |
| openstackgerrit | Tim Burke proposed openstack/swift master: py3: Follow-up to sharder; port rest of container https://review.openstack.org/637663 | 22:10 |
| *** rcernin has joined #openstack-swift | 22:26 | |
| mattoliverau | morning | 22:52 |
| *** tkajinam has joined #openstack-swift | 22:55 | |
| notmyname | good morning mattoliverau | 23:04 |
| mattoliverau | notmyname: o/ | 23:04 |
| *** ccamacho has quit IRC | 23:13 | |
| *** ccamacho has joined #openstack-swift | 23:14 | |
| zaitcev | timburke: are you done with patch 637663? should I squash it into 570320? | 23:22 |
| patchbot | https://review.openstack.org/#/c/637663/ - swift - py3: Follow-up to sharder; port rest of container - 4 patch sets | 23:22 |
| timburke | zaitcev, yeah, i think so? double check that it all actually makes sense... | 23:27 |
| zaitcev | timburke: seems okay. For the record, I strongly disagree with your course on embracing the WSGI cancer instead of containing it inside the environ[], but I am not in a position to bargain about it. I just want this shit to be over. | 23:28 |
| timburke | i'm definitely on board with limiting it to the WSGI parts of our stack, though -- i definitely *don't* want backend.py or diskfile.py to need to know about it | 23:32 |
| timburke | but as long as we're ostensibly a WSGI application... i feel like we ought to adhere to WSGI, in all its madness | 23:33 |
| timburke | it sucks that internal_client users have to call a wsgi app directly... | 23:34 |
| zaitcev | Okay. I'm going to squash it and the get mattoliverau go over it one last time. | 23:35 |
| mattoliverau | yeah do that. I'm just wanna test the sharder in my SAIO to test functionally. But otherwise it's looking good. so if you squish I'll be ready for a final parse and vote :) | 23:38 |
| zaitcev | Well, more like squash... But I'm known to squish when talking to my spouse. | 23:39 |
| mattoliverau | sorry it's taken so long. work has been trying to distract me with other stuff :( | 23:39 |
| mattoliverau | zaitcev: lol | 23:39 |
| zaitcev | + print(broker.get_shard_ranges()) | 23:46 |
| zaitcev | I don't thinks so | 23:46 |
| zaitcev | I always mark my debug inserts with "# P3", then grep one last time. | 23:46 |
| timburke | heh. sorry | 23:47 |
| timburke | i think there may have been another bit where i added a traceback.print_exc(), but it may have been more tolerable since it was in a test | 23:47 |
| zaitcev | So... should I re-add it back or remove in all other tests? | 23:51 |
| zaitcev | Also I found a place where you added it (in test_server.py) | 23:52 |
| timburke | ah! i think i got confused about where i'd added it, so i removed it from the wrong place. adding to both makes sense to me | 23:53 |
| zaitcev | thanks | 23:53 |
| zaitcev | Not sure why we're importing in-place, but it's replicated everywhere | 23:54 |
| timburke | probably for reasons like why i'd done it that way -- i wasn't convinced it should be permanent and i couldn't be bothered losing my place in the file to add the import to the top | 23:55 |
| timburke | so, laziness | 23:55 |
| mattoliverau | notmyname: any chance you can take a look at https://review.openstack.org/#/c/636595 timburke has already made the changes to master. And this is causing troubles for asettle when she builds docs. | 23:57 |
| patchbot | patch 636595 - swift (stable/rocky) - Fixing target for cross-ref 'get_data_dir' - 2 patch sets | 23:57 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!