| *** nguyenhai has joined #openstack-swift | 01:59 | |
| openstackgerrit | Nguyen Hai proposed openstack/swift master: add lower-constraints job https://review.openstack.org/556255 | 02:04 |
|---|---|---|
| *** bkopilov has quit IRC | 02:14 | |
| *** psachin` has joined #openstack-swift | 02:29 | |
| kota_ | hello world | 03:12 |
| kota_ | yeah, it's quiet. | 03:12 |
| mattoliverau | kota_: morning | 03:35 |
| *** bkopilov has joined #openstack-swift | 03:48 | |
| openstackgerrit | Vu Cong Tuan proposed openstack/swift master: Replace unicode with six.text_type https://review.openstack.org/580277 | 04:09 |
| openstackgerrit | Pete Zaitcev proposed openstack/swift master: Make ShardRanges and its CLI py3-compliant https://review.openstack.org/570320 | 04:17 |
| *** cshastri has joined #openstack-swift | 04:59 | |
| *** cbartz has joined #openstack-swift | 05:11 | |
| *** ccamacho has quit IRC | 05:14 | |
| *** psachin` has quit IRC | 05:27 | |
| *** psachin` has joined #openstack-swift | 05:29 | |
| *** ianychoi has joined #openstack-swift | 05:34 | |
| openstackgerrit | Tim Burke proposed openstack/swift master: Make docs link consistent https://review.openstack.org/580291 | 06:05 |
| openstackgerrit | Tim Burke proposed openstack/swift master: Be consistent about capitalizing Swift in the README https://review.openstack.org/580292 | 06:06 |
| mattoliverau | timburke: is see your 4th of July festivities are over and your back at your laptop ;) | 06:07 |
| *** armaan has joined #openstack-swift | 06:13 | |
| *** armaan has quit IRC | 06:15 | |
| *** armaan has joined #openstack-swift | 06:16 | |
| *** ccamacho has joined #openstack-swift | 06:19 | |
| *** armaan has quit IRC | 06:21 | |
| *** hseipp has joined #openstack-swift | 06:38 | |
| *** silor has joined #openstack-swift | 07:05 | |
| *** gkadam has joined #openstack-swift | 07:11 | |
| *** rcernin has quit IRC | 07:54 | |
| *** mikecmpbll has joined #openstack-swift | 08:00 | |
| *** silor has quit IRC | 08:02 | |
| *** silor has joined #openstack-swift | 08:02 | |
| *** geaaru has joined #openstack-swift | 08:02 | |
| *** mikecmpbll has quit IRC | 08:48 | |
| *** mikecmpbll has joined #openstack-swift | 08:49 | |
| *** cbartz has quit IRC | 08:55 | |
| openstackgerrit | Merged openstack/swift master: Address some review comments https://review.openstack.org/558692 | 09:40 |
| *** armaan has joined #openstack-swift | 09:44 | |
| *** spsurya_ has quit IRC | 09:49 | |
| openstackgerrit | Merged openstack/swift master: Make docs link consistent https://review.openstack.org/580291 | 09:51 |
| openstackgerrit | Merged openstack/swift master: Be consistent about capitalizing Swift in the README https://review.openstack.org/580292 | 09:53 |
| *** silor1 has joined #openstack-swift | 10:02 | |
| *** silor has quit IRC | 10:04 | |
| *** silor1 is now known as silor | 10:04 | |
| openstackgerrit | Yuxin Wang proposed openstack/swift master: HEAD to check existence before container PUT https://review.openstack.org/580333 | 10:04 |
| *** armaan has quit IRC | 10:06 | |
| *** bkopilov has quit IRC | 10:28 | |
| *** cbartz has joined #openstack-swift | 10:38 | |
| *** silor has quit IRC | 10:56 | |
| *** hseipp has quit IRC | 11:20 | |
| *** armaan has joined #openstack-swift | 11:22 | |
| *** lifeless has quit IRC | 11:45 | |
| *** cbartz has quit IRC | 11:49 | |
| *** bkopilov has joined #openstack-swift | 11:57 | |
| *** kei_yama has quit IRC | 12:20 | |
| *** silor has joined #openstack-swift | 12:21 | |
| *** silor1 has joined #openstack-swift | 12:38 | |
| *** silor has quit IRC | 12:39 | |
| *** silor has joined #openstack-swift | 12:42 | |
| *** silor1 has quit IRC | 12:43 | |
| *** cshastri has quit IRC | 12:46 | |
| *** armaan has quit IRC | 12:46 | |
| *** armaan has joined #openstack-swift | 12:47 | |
| *** silor1 has joined #openstack-swift | 12:51 | |
| *** silor has quit IRC | 12:53 | |
| openstackgerrit | Alistair Coles proposed openstack/swift master: Make primary only and handoff only workers dump separate recon stats https://review.openstack.org/580215 | 12:55 |
| openstackgerrit | Alistair Coles proposed openstack/swift master: Make obj replicator config error messages more informative https://review.openstack.org/580216 | 12:55 |
| *** silor1 has quit IRC | 12:55 | |
| *** silor has joined #openstack-swift | 13:12 | |
| *** armaan has quit IRC | 13:16 | |
| *** armaan has joined #openstack-swift | 13:17 | |
| *** silor has quit IRC | 13:21 | |
| *** cbartz has joined #openstack-swift | 13:21 | |
| *** armaan has quit IRC | 13:21 | |
| *** armaan has joined #openstack-swift | 13:24 | |
| *** zaitcev has joined #openstack-swift | 13:40 | |
| *** ChanServ sets mode: +v zaitcev | 13:40 | |
| *** cbartz has quit IRC | 13:44 | |
| acoles | clayg: mattoliverau: I need some help knowing how to aggregate replicator timing stats when there are handoff only and primary only workers https://review.openstack.org/#/c/580215/ | 13:44 |
| patchbot | patch 580215 - swift - Make primary only and handoff only workers dump se... | 13:44 |
| acoles | i.e. what is a sensible value to report for the aggregate replication_time? | 13:45 |
| clayg | acoles: yes, this is a big problem | 13:45 |
| acoles | (see my comment on the patch) | 13:45 |
| clayg | the whole idea of cycle time needs to die | 13:45 |
| clayg | which isn't good - because i'm pretty sure a lot of ops have some ideas/tools built around it | 13:46 |
| acoles | yup | 13:46 |
| clayg | the time between first worker starting and last worker finishing is what we want | 13:46 |
| clayg | I honestly don't know if handoffs or primaries will finish first during a rebalance | 13:47 |
| acoles | ^^ this is what the reconstructor reports, I believe | 13:47 |
| clayg | also - do we still have a way to say "do not do any primaries" | 13:47 |
| acoles | p 578213 does not seem to allow replicator_primary_only_workers=0 if replicator_handoff_only_workers is set | 13:48 |
| patchbot | https://review.openstack.org/#/c/578213/ - swift - object-replicator: separate workers for primary/ha... | 13:48 |
| clayg | because we'll want that for saving iops, even if the handoff workers get better at looping on handoffs until they finish | 13:48 |
| clayg | acoles: but like - we could still say .. "handoffs_only = True" and that makes all workers handoff workers? maybe? | 13:49 |
| clayg | the interactoin there is probably akward :\ | 13:49 |
| clayg | acoles: sounds like there's still a fair amount of design to do on this patch? oops | 13:49 |
| clayg | thanks for looking at it. | 13:49 |
| acoles | clayg: for replicator - you mean handoffs_first? | 13:51 |
| clayg | acoles: LOL, yes I forgot replicator DOESN'T have a way to say "no primaries" :'( | 13:52 |
| acoles | I guess torgomatic had a reason to enforce minimum of one worker of each mode (handoffs/primaries) | 13:54 |
| clayg | keep in mind if it makes it simpler it would be great to make the replicator work like the reconstructor: https://github.com/openstack/swift/blob/master/swift/obj/reconstructor.py#L188 | 13:55 |
| clayg | acoles: I had asked about accidently configuring in "primaries only" mode - which afaik is not something anyone would ever want to do | 13:55 |
| clayg | we have on occasion wanted to do ONLY rebalance - but never ONLY primary sync | 13:56 |
| acoles | makes sense | 13:56 |
| acoles | clayg: I could imagine p 578213 landing and a follow up relaxing constraint to allow zero primary_only. But I think the recon stats need to be fixed somehow before it lands. | 13:58 |
| patchbot | https://review.openstack.org/#/c/578213/ - swift - object-replicator: separate workers for primary/ha... | 13:58 |
| clayg | i agree, good scoring - thanks for looking at it! | 13:59 |
| acoles | clayg: so question is whether p 578213 is sane enough for now, albeit not the end goal of (first start - last finish)? Then follow up with changing to (first start - last finish). | 14:01 |
| patchbot | https://review.openstack.org/#/c/578213/ - swift - object-replicator: separate workers for primary/ha... | 14:01 |
| acoles | argh, I mean https://review.openstack.org/#/c/580215/2 | 14:01 |
| patchbot | patch 580215 - swift - Make primary only and handoff only workers dump se... | 14:01 |
| acoles | ^^ if that is ok as a stop-gap then we could squash it and make progress. | 14:02 |
| *** cbartz has joined #openstack-swift | 14:13 | |
| *** zaitcev has quit IRC | 14:24 | |
| *** zaitcev has joined #openstack-swift | 14:26 | |
| *** ChanServ sets mode: +v zaitcev | 14:26 | |
| *** armaan has quit IRC | 14:29 | |
| *** itlinux has quit IRC | 14:40 | |
| *** ccamacho1 has joined #openstack-swift | 14:44 | |
| *** ccamacho1 has quit IRC | 14:44 | |
| *** ccamacho1 has joined #openstack-swift | 14:45 | |
| *** ccamacho has quit IRC | 14:45 | |
| openstackgerrit | Alistair Coles proposed openstack/swift master: Tighten up expirer tests https://review.openstack.org/352438 | 14:47 |
| openstackgerrit | Alistair Coles proposed openstack/swift master: Always fetch listings from shards when possible https://review.openstack.org/571531 | 14:55 |
| *** silor has joined #openstack-swift | 14:55 | |
| openstackgerrit | Alistair Coles proposed openstack/swift master: swift-ring-builder shows hint about composite builder file https://review.openstack.org/475727 | 14:57 |
| *** cbartz has quit IRC | 15:04 | |
| *** armaan has joined #openstack-swift | 15:08 | |
| *** itlinux has joined #openstack-swift | 15:25 | |
| *** gyee has joined #openstack-swift | 15:31 | |
| notmyname | good morning | 16:12 |
| timburke | good morning | 16:14 |
| *** silor has quit IRC | 16:25 | |
| *** mikecmpbll has quit IRC | 16:39 | |
| *** guimaluf has quit IRC | 16:45 | |
| *** armaan_ has joined #openstack-swift | 16:45 | |
| *** armaan has quit IRC | 16:45 | |
| *** armaan_ has quit IRC | 16:50 | |
| *** psachin` has quit IRC | 16:52 | |
| *** ccamacho1 has quit IRC | 17:03 | |
| openstackgerrit | Alistair Coles proposed openstack/swift master: Replace MIME with PUT+POST for EC and Encryption https://review.openstack.org/427911 | 17:10 |
| openstackgerrit | Alistair Coles proposed openstack/swift master: Enable FakeConn to handle multiple requests/responses https://review.openstack.org/579228 | 17:10 |
| openstackgerrit | Alistair Coles proposed openstack/swift master: WIP PUT+POST: test proxy object controller PUT path https://review.openstack.org/579229 | 17:10 |
| acoles | zaitcev: rebasing and added a few more tests in final patch ^^ | 17:10 |
| zaitcev | acoles: thanks a lot. Can't wait to see if you did anything about that extra raise ChunkReadError | 17:11 |
| acoles | zaitcev: which one is that? I haven't touched anything today, just added tests | 17:12 |
| acoles | oh, there were a few unnecessary self.app.use_put_v1 assignments in test_obj that I cleaned up | 17:12 |
| zaitcev | You wrote "However, I don't think raising ChunkWriteTimout is the correct strategy. Raising an exception here will cause the entire PUT to terminate early, whereas it is only this putter that has failed - a quorum may still succeed. I *think* the right thing may be to make the failed response be the putter's final response, and then return - no poont sending the POST requests." | 17:12 |
| acoles | oh yeah | 17:13 |
| zaitcev | I didn't realize what would happen when I put it there. Seemed like I needed something to signal to the main loop that a putter failed. | 17:13 |
| acoles | zaitcev: this was in the last patchset actually https://review.openstack.org/#/c/579229/4/swift/proxy/controllers/obj.py | 17:13 |
| patchbot | patch 579229 - swift - WIP PUT+POST: test proxy object controller PUT path | 17:13 |
| acoles | and the tests I've been adding assert the behaviour | 17:14 |
| zaitcev | yay let me see | 17:14 |
| zaitcev | go go gadget git review -d 579229 | 17:15 |
| acoles | TBH the putter final_resp dance can get confusing me but the tests seem to confirm things | 17:15 |
| *** geaaru has quit IRC | 17:41 | |
| *** bkopilov has quit IRC | 17:43 | |
| *** bkopilov has joined #openstack-swift | 17:52 | |
| *** SkyRocknRoll has joined #openstack-swift | 17:55 | |
| *** SkyRocknRoll has quit IRC | 17:57 | |
| notmyname | clayg: timburke: based on comments in gerrit, I'm tempted to go ahead and land both https://review.openstack.org/#/c/337960/ and https://review.openstack.org/#/c/575860/ | 18:18 |
| patchbot | patch 337960 - swift - Send correct SLO ETag for container updates | 18:18 |
| patchbot | patch 575860 - swift - Include '-' in multipart ETags | 18:18 |
| timburke | notmyname: i could also try another approach today -- basically, take the s3api approach and apply it to slo, too (while being careful to preserve any other override params) | 18:19 |
| timburke | i think it'd make clay happier, and still provide value for timur and others who might want to know about SLO-ness in listings | 18:20 |
| *** ccamacho has joined #openstack-swift | 18:23 | |
| *** armaan has joined #openstack-swift | 18:32 | |
| *** armaan has quit IRC | 18:44 | |
| *** armaan has joined #openstack-swift | 18:45 | |
| *** armaan has quit IRC | 18:49 | |
| *** ccamacho has quit IRC | 18:54 | |
| *** UnfairFunction has joined #openstack-swift | 19:20 | |
| UnfairFunction | /msg NickServ identify maosSo8e | 19:21 |
| *** Chealion has quit IRC | 19:21 | |
| DHE | protip: set your nickserv password as your IRC connection password to auto-identify on connect without risk of this happening. | 19:21 |
| UnfairFunction | How do I open a bug for pyeclib? | 19:22 |
| UnfairFunction | thank you, DHE :) | 19:22 |
| tdasilva | UnfairFunction: https://bugs.launchpad.net/pyeclib | 19:23 |
| UnfairFunction | the pyeclib repo does not have a CONTRIBUTING page.. | 19:23 |
| *** Chealion has joined #openstack-swift | 19:24 | |
| tdasilva | UnfairFunction: good point, little busy atm, but will take a look at that | 19:27 |
| tdasilva | even if we just need to add the info about bugs here: https://github.com/openstack/pyeclib#code-maintenance | 19:27 |
| *** UnfairFunction has quit IRC | 19:31 | |
| *** gkadam has quit IRC | 19:32 | |
| *** lifeless has joined #openstack-swift | 19:39 | |
| *** UnfairFunction has joined #openstack-swift | 20:12 | |
| *** UnfairFunction has quit IRC | 20:15 | |
| *** UnfairFunction has joined #openstack-swift | 20:15 | |
| *** ccamacho has joined #openstack-swift | 20:19 | |
| *** mikecmpbll has joined #openstack-swift | 21:21 | |
| *** armaan has joined #openstack-swift | 21:23 | |
| *** itlinux has quit IRC | 21:24 | |
| openstackgerrit | Ehud Kaldor proposed openstack/swift master: added Alpine Linux (apk) entries to bindep.txt. https://review.openstack.org/580511 | 21:25 |
| openstackgerrit | Ehud Kaldor proposed openstack/swift master: added Alpine Linux (apk) entries to bindep.txt. closes-bug: #1780163 Change-Id: Iff24a2b02b8927113d616a5af06331f1a1e31170 https://review.openstack.org/580511 | 21:29 |
| openstack | bug 1780163 in OpenStack Object Storage (swift) "no Alpine Linux (apk) entries in Swift bindep.txt" [Undecided,New] https://launchpad.net/bugs/1780163 | 21:29 |
| *** rcernin has joined #openstack-swift | 22:00 | |
| *** armaan has quit IRC | 22:36 | |
| *** armaan has joined #openstack-swift | 22:37 | |
| *** armaan has quit IRC | 22:41 | |
| mattoliverau | morning | 22:49 |
| notmyname | mattoliverau: hello | 22:52 |
| *** rcernin has quit IRC | 22:58 | |
| *** rcernin has joined #openstack-swift | 23:01 | |
| *** ianychoi_ has joined #openstack-swift | 23:01 | |
| *** ianychoi has quit IRC | 23:04 | |
| mattoliverau | notmyname: o/ enjoy the holiday? | 23:04 |
| notmyname | yeah. went on a hike with the kids. it was great | 23:06 |
| *** kei_yama has joined #openstack-swift | 23:22 | |
| mattoliverau | nice | 23:27 |
| mattoliverau | UnfairFunction: do mind if I quickly update your commit message for p 580511? Just some OCD. I can do it via gerrit, you'd still remain the patch author. It would mean I can then go land it :) | 23:58 |
| patchbot | https://review.openstack.org/#/c/580511/ - swift - added Alpine Linux (apk) entries to bindep.txt. cl... | 23:58 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!