*** altlogbot_2 has joined #openstack-swift | 00:29 | |
*** altlogbot_2 has quit IRC | 00:37 | |
*** altlogbot_1 has joined #openstack-swift | 01:07 | |
openstackgerrit | Pete Zaitcev proposed openstack/swift master: py3: port dlo func tests https://review.opendev.org/642920 | 01:12 |
---|---|---|
*** altlogbot_1 has quit IRC | 01:15 | |
*** gyee has quit IRC | 01:22 | |
*** altlogbot_1 has joined #openstack-swift | 02:19 | |
*** irclogbot_1 has joined #openstack-swift | 02:19 | |
*** altlogbot_1 has quit IRC | 02:25 | |
*** irclogbot_1 has quit IRC | 02:26 | |
*** tkajinam has quit IRC | 02:29 | |
*** tkajinam has joined #openstack-swift | 02:29 | |
*** psachin has joined #openstack-swift | 03:03 | |
openstackgerrit | Merged openstack/swift master: py3: Specify an encoding when loading db.pending pickles https://review.opendev.org/665486 | 03:37 |
*** mahatic has joined #openstack-swift | 03:49 | |
*** ChanServ sets mode: +v mahatic | 03:49 | |
kota_ | good morning mattoliverau | 03:51 |
mattoliverau | kota_: o/ | 03:51 |
*** irclogbot_0 has joined #openstack-swift | 04:07 | |
*** irclogbot_0 has quit IRC | 04:12 | |
*** altlogbot_3 has joined #openstack-swift | 04:13 | |
*** altlogbot_3 has quit IRC | 04:17 | |
openstackgerrit | Merged openstack/swift master: Increase some middleware test coverage https://review.opendev.org/652185 | 04:39 |
*** altlogbot_3 has joined #openstack-swift | 05:03 | |
*** irclogbot_2 has joined #openstack-swift | 05:07 | |
*** altlogbot_3 has quit IRC | 05:21 | |
*** irclogbot_2 has quit IRC | 05:22 | |
*** pcaruana has joined #openstack-swift | 05:37 | |
*** rcernin has quit IRC | 06:10 | |
*** rcernin has joined #openstack-swift | 06:11 | |
*** pcaruana has quit IRC | 06:20 | |
*** new_student1411 has joined #openstack-swift | 06:35 | |
new_student1411 | Hi, can anyone help me with how to set `X-Account-Sysmeta-Global-Write-Ratelimit` for account ratelimit. | 06:36 |
openstackgerrit | HCLTech-SSW proposed openstack/swift master: Add ability to undelete an account. https://review.opendev.org/507808 | 06:56 |
*** altlogbot_1 has joined #openstack-swift | 07:03 | |
*** altlogbot_1 has quit IRC | 07:08 | |
*** new_student1411 has quit IRC | 07:25 | |
*** irclogbot_1 has joined #openstack-swift | 07:40 | |
*** irclogbot_1 has quit IRC | 07:44 | |
openstackgerrit | zengjia proposed openstack/swift master: Update auth_url in install docs https://review.opendev.org/670005 | 07:53 |
*** pcaruana has joined #openstack-swift | 07:54 | |
*** gkadam has joined #openstack-swift | 08:13 | |
*** tkajinam has quit IRC | 08:14 | |
*** gkadam has quit IRC | 08:22 | |
*** rdejoux has joined #openstack-swift | 08:23 | |
*** Fidde has joined #openstack-swift | 08:31 | |
*** zaitcev__ has joined #openstack-swift | 08:48 | |
*** ChanServ sets mode: +v zaitcev__ | 08:48 | |
*** zaitcev_ has quit IRC | 08:52 | |
*** irclogbot_3 has joined #openstack-swift | 08:58 | |
*** irclogbot_3 has quit IRC | 09:02 | |
*** ccamacho has joined #openstack-swift | 09:03 | |
*** baojg has joined #openstack-swift | 09:09 | |
*** rcernin has quit IRC | 09:17 | |
*** altlogbot_2 has joined #openstack-swift | 09:26 | |
*** altlogbot_2 has quit IRC | 09:28 | |
*** altlogbot_2 has joined #openstack-swift | 09:30 | |
*** new_student1411 has joined #openstack-swift | 09:33 | |
*** altlogbot_2 has quit IRC | 09:34 | |
openstackgerrit | Merged openstack/swift master: py3: Be able to read and write non-ASCII headers https://review.opendev.org/662546 | 09:35 |
*** altlogbot_3 has joined #openstack-swift | 09:46 | |
*** altlogbot_3 has quit IRC | 09:50 | |
*** zaitcev_ has joined #openstack-swift | 09:56 | |
*** ChanServ sets mode: +v zaitcev_ | 09:56 | |
*** new_student1411 has quit IRC | 09:59 | |
*** zaitcev__ has quit IRC | 09:59 | |
*** irclogbot_2 has joined #openstack-swift | 10:24 | |
*** irclogbot_2 has quit IRC | 10:26 | |
*** new_student1411 has joined #openstack-swift | 10:56 | |
*** tesseract has joined #openstack-swift | 11:08 | |
*** zaitcev__ has joined #openstack-swift | 11:36 | |
*** ChanServ sets mode: +v zaitcev__ | 11:36 | |
*** zaitcev_ has quit IRC | 11:39 | |
*** tomha has joined #openstack-swift | 11:51 | |
*** tomha has quit IRC | 11:57 | |
openstackgerrit | Alexandre Lécuyer proposed openstack/swift feature/losf: Add tests for vfile.py https://review.opendev.org/666378 | 12:02 |
*** henriqueof has joined #openstack-swift | 12:51 | |
*** zaitcev__ has quit IRC | 12:52 | |
*** zaitcev__ has joined #openstack-swift | 13:04 | |
*** ChanServ sets mode: +v zaitcev__ | 13:04 | |
*** altlogbot_3 has joined #openstack-swift | 13:12 | |
*** zaitcev_ has joined #openstack-swift | 13:12 | |
*** ChanServ sets mode: +v zaitcev_ | 13:12 | |
*** zaitcev__ has quit IRC | 13:15 | |
*** altlogbot_3 has quit IRC | 13:16 | |
*** irclogbot_2 has joined #openstack-swift | 13:18 | |
*** irclogbot_2 has quit IRC | 13:22 | |
*** BjoernT has joined #openstack-swift | 13:56 | |
*** Fidde has quit IRC | 14:52 | |
timburke | \o/ one step closer to the release! | 14:55 |
timburke | on an unrelated note.... now that we can UPDATE container DBs... i wonder if we should revisit how the replicator enqueues reconciler work... | 14:56 |
*** tdasilva has joined #openstack-swift | 15:05 | |
*** ChanServ sets mode: +v tdasilva | 15:05 | |
*** altlogbot_0 has joined #openstack-swift | 15:09 | |
*** zaitcev__ has joined #openstack-swift | 15:11 | |
*** ChanServ sets mode: +v zaitcev__ | 15:11 | |
*** altlogbot_0 has quit IRC | 15:12 | |
*** zaitcev_ has quit IRC | 15:15 | |
*** henriqueof has quit IRC | 15:29 | |
*** BjoernT_ has joined #openstack-swift | 15:33 | |
*** BjoernT has quit IRC | 15:36 | |
*** BjoernT has joined #openstack-swift | 15:36 | |
*** BjoernT_ has quit IRC | 15:38 | |
*** henriqueof has joined #openstack-swift | 15:43 | |
*** henriqueof has quit IRC | 15:47 | |
*** viks___ has quit IRC | 15:55 | |
*** altlogbot_3 has joined #openstack-swift | 15:57 | |
*** altlogbot_3 has quit IRC | 15:58 | |
openstackgerrit | Tim Burke proposed openstack/swift master: functests: make container creation less flakey in test_object https://review.opendev.org/670111 | 16:13 |
*** gyee has joined #openstack-swift | 16:16 | |
DHE | so, I'm considering making a custom diskfile scheme, mainly because I will have a lot of sorta small (100k to 4M) files and the scheme of "1 object" per directory sounds like it's gonna hurt. but at the same time swift-on-file is described as being unsuitable for swift general use since it sounds like replication doesn't work. is that right? | 16:18 |
timburke | new_student1411, you won't be able to do it through the proxy-server, but you can issue a curl request straight to the account server (assuming you're inside the cluster's network) | 16:20 |
tdasilva | DHE: if you are concerned about lots of small files, you should checkout the losf branch, that's a project that rledisez and alecuyer are leading | 16:21 |
*** henriqueof has joined #openstack-swift | 16:21 | |
timburke | to figure out *which* account server to POST to, run something like `swift-get-nodes <path to account ring> <account name>` | 16:21 |
tdasilva | DHE: I'm not sure how that concern has anything to do with the swift-on-file project. The project had another goal and it's actually pretty dead | 16:21 |
*** psachin has quit IRC | 16:26 | |
*** henriqueof has quit IRC | 16:29 | |
DHE | tdasilva: it seems to be the go-to alternative diskfile implementation | 16:32 |
DHE | I was under the impression losf was intended for really small files, like 10k or smaller | 16:33 |
*** henriqueof has joined #openstack-swift | 16:37 | |
clayg | I went ahead and marked lp bug #1836082 as "high" because I don't understand exactly how bad the failure mode is. Obviously if you have a split brain and a sharded container you could have object data that's inaccesible in the wrong policy until you manually re-enqueue the rows to move. | 16:48 |
openstack | Launchpad bug 1836082 in OpenStack Object Storage (swift) "Reconciler-enqueuing needs to be shard-aware" [High,Confirmed] https://launchpad.net/bugs/1836082 | 16:48 |
clayg | But in the meantime the reconciler could be off it's rocker trying to do GETs for objects that don't exist up to a reclaim age only to eventually drop the records on the floor - at which point ... would the sharded container re-enqueue them? | 16:49 |
clayg | Maybe after an upgrade that fixes the problem you could just reset the row_id pointer and all things would sort themselves out 🤷♂️ | 16:49 |
openstackgerrit | Tim Burke proposed openstack/swift master: py3: port SLO func tests https://review.opendev.org/659673 | 16:50 |
*** altlogbot_2 has joined #openstack-swift | 16:50 | |
openstackgerrit | Tim Burke proposed openstack/swift master: py3: port staticweb and domain_remap func tests https://review.opendev.org/647923 | 16:51 |
openstackgerrit | Tim Burke proposed openstack/swift master: py3: Cover account/container func tests https://review.opendev.org/645388 | 16:53 |
*** altlogbot_2 has quit IRC | 16:54 | |
openstackgerrit | Tim Burke proposed openstack/swift master: py3: Bring functional/test_object.py under test; add func-ec-py37 job https://review.opendev.org/645895 | 16:55 |
timburke | aha! that's what i get for just looking at master's requirements list: https://github.com/openstack/swift/commit/9d980e2#diff-b4ef698db8ca845e5845c4618278f29a | 16:59 |
*** henriqueof has quit IRC | 17:03 | |
*** henriqueof has joined #openstack-swift | 17:03 | |
openstackgerrit | Tim Burke proposed openstack/swift feature/losf: Back out urllib3 requirement https://review.opendev.org/670139 | 17:13 |
*** henriqueof has quit IRC | 17:34 | |
*** henriqueof has joined #openstack-swift | 17:35 | |
*** BjoernT has quit IRC | 17:41 | |
*** Fidde has joined #openstack-swift | 17:41 | |
*** new_student1411 has quit IRC | 17:47 | |
*** rdejoux has quit IRC | 18:04 | |
*** irclogbot_3 has joined #openstack-swift | 18:06 | |
*** irclogbot_3 has quit IRC | 18:08 | |
*** zaitcev__ has quit IRC | 18:18 | |
*** zaitcev__ has joined #openstack-swift | 18:30 | |
*** ChanServ sets mode: +v zaitcev__ | 18:30 | |
timburke | clayg, yeah, resetting the high-water mark should certainly help... but i'm a little nervous about deleted rows -- they'll only hang around so long, and after that we're out of luck :-/ | 18:54 |
openstackgerrit | Merged openstack/swift master: container: Unify storage-policy error paths https://review.opendev.org/660525 | 18:54 |
openstackgerrit | Merged openstack/python-swiftclient master: Optionally display listings in raw json https://review.opendev.org/668705 | 18:54 |
*** altlogbot_2 has joined #openstack-swift | 19:03 | |
*** altlogbot_2 has quit IRC | 19:04 | |
*** Fidde has quit IRC | 19:09 | |
*** BjoernT has joined #openstack-swift | 19:11 | |
*** tesseract has quit IRC | 19:16 | |
*** irclogbot_3 has joined #openstack-swift | 19:20 | |
*** irclogbot_3 has quit IRC | 19:22 | |
*** zaitcev__ has quit IRC | 19:29 | |
clayg | i feel like there's some ambiguity with a PUT hardlink 404 - it could mean the container doesn't exist - or it could mean the target doesn't exist... we have the HEAD resp, instead of returning it directly we could spin up a 409 like we do for etag doesn't match? 🤷♂️ | 19:35 |
clayg | I'm only worried about it because of versioned writes wanting to do a stack mode delete when the top of the queue has expired... you can't make the hardlink to an expired object | 19:36 |
*** zaitcev__ has joined #openstack-swift | 19:41 | |
*** ChanServ sets mode: +v zaitcev__ | 19:41 | |
timburke | there's another thing we oughta think hard about... how can/should s3-style versioning / history mode versioning interact with object expiration? | 19:43 |
clayg | you mean if someone is using both api's on the same objects? or you're talking about bucket policies? | 19:45 |
clayg | or do you just mean swift versioning + swift expiring features? cause yeah that's an existing mess of itself 😁 | 19:47 |
*** openstackgerrit has quit IRC | 19:49 | |
timburke | that last one, basically | 20:26 |
timburke | s3 is impacted insofar as s3 has its own notion of how expiration works that we ought to think about as we decide how *swift* should work | 20:27 |
*** altlogbot_2 has joined #openstack-swift | 20:28 | |
*** altlogbot_2 has quit IRC | 20:32 | |
*** zaitcev_ has joined #openstack-swift | 20:36 | |
*** ChanServ sets mode: +v zaitcev_ | 20:36 | |
*** zaitcev__ has quit IRC | 20:40 | |
*** pcaruana has quit IRC | 20:48 | |
kota_ | morning | 20:58 |
mattoliverau | Morning | 20:59 |
kota_ | timburke: meeting | 21:01 |
kota_ | ? | 21:01 |
timburke | yep! sorry | 21:01 |
timburke | zaitcev_, clayg ^^^ | 21:02 |
clayg | already! where does the day go :'( | 21:02 |
*** zaitcev_ is now known as zaitcev | 21:02 | |
*** altlogbot_3 has joined #openstack-swift | 21:04 | |
*** altlogbot_3 has quit IRC | 21:08 | |
*** altlogbot_2 has joined #openstack-swift | 21:18 | |
*** altlogbot_2 has quit IRC | 21:22 | |
*** openstackgerrit has joined #openstack-swift | 21:26 | |
openstackgerrit | Merged openstack/swift master: functests: make container creation less flakey in test_object https://review.opendev.org/670111 | 21:26 |
*** irclogbot_1 has joined #openstack-swift | 21:49 | |
*** irclogbot_1 has quit IRC | 21:52 | |
clayg | queue all the tasks | 21:59 |
clayg | rledisez: we're gunna queue reconstructor rebuilds for sure 👍 | 21:59 |
rledisez | clayg: i'm so glad you say that! i dreamt of it for few month now :D | 21:59 |
rledisez | i gotta go, but i'll follow up on that topic | 22:00 |
clayg | well apparently we have to fix the reconciler first | 22:00 |
clayg | kthxbai! | 22:00 |
clayg | i don't think we need to write down the epoch - as long as the task has the obj name in it we could always rehash and be like "this part isn't right anymore" | 22:01 |
clayg | then we can either re-enqueue it where it goes or just deal with it | 22:02 |
*** henriqueof has quit IRC | 22:06 | |
DHE | so the default diskfile implementation makes a directory per object, and the reason seems to be so that rsync et al can handle multiple versions of an object nicely. right? I'm contemplating making a variant that avoids it and instead has 1 file per object in a small number of hash directories (squid (?) style) | 22:12 |
openstackgerrit | Tim Burke proposed openstack/swift master: functests: Make test_PUT_metadata less flakey https://review.opendev.org/670209 | 22:37 |
timburke | ok, so now that we've got https://review.opendev.org/#/c/668705/ ... should the next swiftclient still be 3.7.1, or 3.8.0? | 22:44 |
patchbot | patch 668705 - python-swiftclient - Optionally display listings in raw json (MERGED) - 9 patch sets | 22:44 |
timburke | i kinda really want a release with https://review.opendev.org/#/c/651666/ in it | 22:45 |
patchbot | patch 651666 - python-swiftclient - Fix SLO re-upload (MERGED) - 3 patch sets | 22:45 |
timburke | plus it's been like 4 months | 22:45 |
*** rcernin has joined #openstack-swift | 22:48 | |
*** altlogbot_3 has joined #openstack-swift | 23:04 | |
*** tkajinam has joined #openstack-swift | 23:06 | |
*** altlogbot_3 has quit IRC | 23:08 | |
*** altlogbot_2 has joined #openstack-swift | 23:14 | |
*** altlogbot_2 has quit IRC | 23:18 | |
*** altlogbot_2 has joined #openstack-swift | 23:20 | |
*** altlogbot_2 has quit IRC | 23:24 | |
*** altlogbot_3 has joined #openstack-swift | 23:26 | |
openstackgerrit | Merged openstack/python-swiftclient master: Support pdb in tests better https://review.opendev.org/473925 | 23:27 |
openstackgerrit | Merged openstack/swift feature/losf: Back out urllib3 requirement https://review.opendev.org/670139 | 23:27 |
lifeless | /win goto ##flyingcircus | 23:28 |
*** altlogbot_3 has quit IRC | 23:30 | |
*** zaitcev has quit IRC | 23:33 | |
*** hoonetorg has quit IRC | 23:38 | |
*** hoonetorg has joined #openstack-swift | 23:40 | |
*** zaitcev has joined #openstack-swift | 23:45 | |
*** ChanServ sets mode: +v zaitcev | 23:45 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!