| *** lyrrad has quit IRC | 00:03 | |
| *** kutija has quit IRC | 00:04 | |
| MooingLemur | or maybe it's the same fragment... but why doesn't it ever reconstruct? Nothing's logged about it | 00:06 |
|---|---|---|
| *** setmason has joined #openstack-swift | 00:16 | |
| MooingLemur | it's missing two fragments, 0 and 10. | 00:16 |
| *** gyee has quit IRC | 00:18 | |
| *** zhill_ has quit IRC | 00:19 | |
| *** remix_tj has quit IRC | 00:19 | |
| *** gyee has joined #openstack-swift | 00:25 | |
| *** zhill_ has joined #openstack-swift | 00:39 | |
| *** setmason_ has joined #openstack-swift | 00:39 | |
| *** setmason has quit IRC | 00:41 | |
| *** setmason_ is now known as setmason | 00:41 | |
| *** kota_ has joined #openstack-swift | 00:43 | |
| kota_ | morning | 00:43 |
| *** annegentle has joined #openstack-swift | 00:45 | |
| *** SkyRocknRoll has quit IRC | 00:53 | |
| *** zhill_ has quit IRC | 00:59 | |
| *** charlesw has joined #openstack-swift | 01:02 | |
| *** annegentle has quit IRC | 01:03 | |
| *** asettle has quit IRC | 01:11 | |
| *** SkyRocknRoll has joined #openstack-swift | 01:19 | |
| *** blmartin has quit IRC | 01:32 | |
| *** blmartin_ has joined #openstack-swift | 01:32 | |
| *** setmason_ has joined #openstack-swift | 01:36 | |
| *** asettle has joined #openstack-swift | 01:37 | |
| *** setmason has quit IRC | 01:39 | |
| *** setmason_ is now known as setmason | 01:39 | |
| *** zackmdavis has left #openstack-swift | 01:55 | |
| *** kota_ has quit IRC | 01:56 | |
| *** proteusguy has quit IRC | 02:12 | |
| *** proteusguy has joined #openstack-swift | 02:23 | |
| *** wbhuber has quit IRC | 02:26 | |
| *** setmason has quit IRC | 02:31 | |
| *** lyrrad has joined #openstack-swift | 02:34 | |
| *** blmartin_ has quit IRC | 02:53 | |
| *** mwheckmann has joined #openstack-swift | 03:01 | |
| *** gyee is now known as operator99 | 03:06 | |
| *** barra204 has quit IRC | 03:11 | |
| *** barra204 has joined #openstack-swift | 03:13 | |
| *** zhill_ has joined #openstack-swift | 03:17 | |
| *** zhill_ has quit IRC | 03:20 | |
| *** zhill has quit IRC | 03:45 | |
| *** zhill has joined #openstack-swift | 03:59 | |
| *** setmason has joined #openstack-swift | 04:00 | |
| *** mwheckmann has quit IRC | 04:08 | |
| *** proteusguy has quit IRC | 04:11 | |
| *** proteusguy has joined #openstack-swift | 04:14 | |
| *** janonymous has joined #openstack-swift | 04:15 | |
| *** kota_ has joined #openstack-swift | 04:19 | |
| *** asettle is now known as asettle-afk | 04:23 | |
| *** charlesw has quit IRC | 04:29 | |
| *** zhill has quit IRC | 04:34 | |
| *** SkyRocknRoll has quit IRC | 04:43 | |
| *** zhill has joined #openstack-swift | 04:46 | |
| *** zynisch_o7 has joined #openstack-swift | 04:48 | |
| *** setmason has quit IRC | 04:48 | |
| *** setmason has joined #openstack-swift | 04:50 | |
| *** Triveni has joined #openstack-swift | 04:55 | |
| MooingLemur | clayg: I wonder if it was really broken all along when the ring wasn't properly dispersed for EC and that's when I uploaded those objects. | 05:15 |
| MooingLemur | but I double-checked and made sure the current ring is good with validate_ec_ring_parts.py | 05:15 |
| MooingLemur | clayg: that's the only difference between the two containers I can think of (since one didn't lose anything) | 05:17 |
| *** SkyRocknRoll has joined #openstack-swift | 05:27 | |
| *** SkyRocknRoll has joined #openstack-swift | 05:27 | |
| *** zynisch_o7 has quit IRC | 05:37 | |
| *** nadeem has joined #openstack-swift | 05:40 | |
| *** Triveni_ has joined #openstack-swift | 05:40 | |
| *** setmason_ has joined #openstack-swift | 05:40 | |
| *** peluse has quit IRC | 05:41 | |
| *** peluse has joined #openstack-swift | 05:42 | |
| *** setmason has quit IRC | 05:43 | |
| *** Triveni has quit IRC | 05:43 | |
| *** setmason_ has quit IRC | 05:46 | |
| cschwede | Good Morning! | 06:12 |
| *** zackmdavis has joined #openstack-swift | 06:48 | |
| openstackgerrit | Merged openstack/swift: Remove simplejson from tests https://review.openstack.org/186590 | 06:57 |
| mattoliverau | cschwede: morning! | 07:02 |
| *** leopoldj has joined #openstack-swift | 07:04 | |
| *** mmcardle has joined #openstack-swift | 07:16 | |
| *** asettle-afk is now known as asettle | 07:23 | |
| *** geaaru has joined #openstack-swift | 07:27 | |
| *** jistr has joined #openstack-swift | 07:38 | |
| *** bkopilov has quit IRC | 07:39 | |
| *** hseipp has joined #openstack-swift | 07:42 | |
| *** hseipp has quit IRC | 07:42 | |
| *** hseipp has joined #openstack-swift | 07:42 | |
| openstackgerrit | Merged openstack/python-swiftclient: Stop decoding object content https://review.openstack.org/184659 | 08:13 |
| *** jordanP has joined #openstack-swift | 08:27 | |
| *** remix_tj has joined #openstack-swift | 08:29 | |
| *** nadeem has quit IRC | 08:45 | |
| *** silor has joined #openstack-swift | 09:24 | |
| *** km has quit IRC | 09:26 | |
| *** km has joined #openstack-swift | 09:31 | |
| *** aix has joined #openstack-swift | 09:41 | |
| *** acoles_away is now known as acoles | 09:44 | |
| *** km has quit IRC | 09:48 | |
| acoles | jrichli clayg: "object-post-sysmeta" was the one i couldn't remember from yvr. thx! | 09:49 |
| *** kei_yama has quit IRC | 09:50 | |
| *** bkopilov has joined #openstack-swift | 09:53 | |
| openstackgerrit | Alistair Coles proposed openstack/swift: Remove simplejson from staticweb https://review.openstack.org/186170 | 10:19 |
| *** ppai has joined #openstack-swift | 10:29 | |
| openstackgerrit | Christian Schwede proposed openstack/swift: Ignore invalid PIDs in swift-init https://review.openstack.org/186703 | 10:31 |
| *** asettle has quit IRC | 10:32 | |
| openstackgerrit | Christian Schwede proposed openstack/swift: Ignore invalid PIDs in swift-init https://review.openstack.org/186703 | 10:33 |
| *** chenhuayi has joined #openstack-swift | 10:44 | |
| *** asettle has joined #openstack-swift | 10:46 | |
| *** SkyRocknRoll has quit IRC | 10:53 | |
| *** SkyRocknRoll has joined #openstack-swift | 10:53 | |
| *** aix has quit IRC | 10:54 | |
| *** chenhuayi has quit IRC | 11:08 | |
| *** cdelatte has quit IRC | 11:38 | |
| *** cdelatte has joined #openstack-swift | 11:39 | |
| *** aix has joined #openstack-swift | 11:45 | |
| *** silor has quit IRC | 11:56 | |
| *** aix has quit IRC | 12:06 | |
| *** aix has joined #openstack-swift | 12:08 | |
| *** Shimoda has joined #openstack-swift | 12:08 | |
| *** SkyRocknRoll has quit IRC | 12:09 | |
| openstackgerrit | Christian Schwede proposed openstack/swift: Ignore invalid PIDs in swift-init https://review.openstack.org/186703 | 12:11 |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift: WIP: Fix missing container update https://review.openstack.org/186735 | 12:15 |
| *** SkyRocknRoll has joined #openstack-swift | 12:17 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift: WIP: Fix missing container update https://review.openstack.org/186735 | 12:20 |
| *** kota_ has quit IRC | 12:21 | |
| *** fthiagogv has joined #openstack-swift | 12:41 | |
| *** lpabon has joined #openstack-swift | 12:44 | |
| *** gsilvis has quit IRC | 12:50 | |
| *** SkyRocknRoll has quit IRC | 12:56 | |
| *** jkugel has joined #openstack-swift | 13:05 | |
| *** proteusguy has quit IRC | 13:10 | |
| *** lpabon has quit IRC | 13:12 | |
| tdasilva | good morning | 13:16 |
| acoles | tdasilva: good morning | 13:19 |
| *** zynisch_o7 has joined #openstack-swift | 13:22 | |
| *** proteusguy has joined #openstack-swift | 13:22 | |
| openstackgerrit | Merged openstack/swift: Remove simplejson from staticweb https://review.openstack.org/186170 | 13:23 |
| *** silor has joined #openstack-swift | 13:25 | |
| *** annegentle has joined #openstack-swift | 13:29 | |
| *** gsilvis has joined #openstack-swift | 13:29 | |
| *** acampbel11 has joined #openstack-swift | 13:44 | |
| *** acampbell has joined #openstack-swift | 13:44 | |
| *** annegent_ has joined #openstack-swift | 13:44 | |
| *** annegentle has quit IRC | 13:45 | |
| *** gsilvis has quit IRC | 13:47 | |
| *** aix has quit IRC | 13:49 | |
| *** aix has joined #openstack-swift | 13:49 | |
| *** CaioBren_ has joined #openstack-swift | 13:52 | |
| *** mwheckmann has joined #openstack-swift | 13:54 | |
| *** lpabon has joined #openstack-swift | 13:54 | |
| *** jrichli has joined #openstack-swift | 13:56 | |
| *** CaioBrentano has quit IRC | 13:56 | |
| *** annegent_ has quit IRC | 13:57 | |
| *** jkugel has left #openstack-swift | 13:58 | |
| *** CaioBren_ has quit IRC | 14:00 | |
| *** charlesw has joined #openstack-swift | 14:03 | |
| *** SkyRocknRoll has joined #openstack-swift | 14:05 | |
| *** CaioBrentano has joined #openstack-swift | 14:05 | |
| *** acampbel11 has quit IRC | 14:07 | |
| *** acampbell has quit IRC | 14:07 | |
| *** ppai has quit IRC | 14:07 | |
| *** acampbell has joined #openstack-swift | 14:08 | |
| *** acampbel11 has joined #openstack-swift | 14:08 | |
| *** zynisch_o7 has quit IRC | 14:09 | |
| *** annegentle has joined #openstack-swift | 14:10 | |
| *** CaioBrentano has quit IRC | 14:13 | |
| *** CaioBrentano has joined #openstack-swift | 14:13 | |
| *** gsilvis has joined #openstack-swift | 14:15 | |
| openstackgerrit | Christian Schwede proposed openstack/swift-specs: Add object undelete spec https://review.openstack.org/155421 | 14:16 |
| *** leopoldj has quit IRC | 14:17 | |
| *** breitz has quit IRC | 14:20 | |
| *** breitz has joined #openstack-swift | 14:21 | |
| *** gsilvis has quit IRC | 14:24 | |
| *** jistr has quit IRC | 14:26 | |
| *** acampbell has quit IRC | 14:31 | |
| *** acampbel11 has quit IRC | 14:31 | |
| zaitcev | Does anyone remember a time when someone submitted a humongous patch that swapped arguments in some kind of test assertion? The justification was that testtools mandated that order. We declined it because unittest was the opposite. | 14:32 |
| *** wbhuber has joined #openstack-swift | 14:32 | |
| cschwede | zaitcev: yes, i remember swapping arguments in tests - iirc unittest prefers the default as first argument, because if the test fails it will print more details then. do you have a specific patch in mind? | 14:35 |
| *** esker has joined #openstack-swift | 14:35 | |
| zaitcev | cschwede: someone picked the bug "move to testtools" today, so I expect we'll see a patch soon. | 14:36 |
| acoles | zaitcev: cschwede : was it 'expected value first, actual value second'? | 14:36 |
| cschwede | zaitcev: ah, ok. interesting | 14:36 |
| zaitcev | I hit it before with this, note that py3 is still the same as always was, it's just testtools being incompatible for no reason - http://bugs.python.org/issue20798 | 14:36 |
| cschwede | acoles: yes, at least for unittest | 14:37 |
| zaitcev | bug is https://bugs.launchpad.net/bugs/1179009 - jelly is coming for us, guys | 14:38 |
| openstack | Launchpad bug 1179009 in Trove "Use testtools for base test class" [Wishlist,In progress] - Assigned to Wu Wenxiang (wu-wenxiang) | 14:38 |
| zaitcev | it's not very important, I was just curious | 14:38 |
| *** acampbell has joined #openstack-swift | 14:40 | |
| *** acampbel11 has joined #openstack-swift | 14:40 | |
| *** acampbel11 has quit IRC | 14:40 | |
| *** acampbell has quit IRC | 14:40 | |
| *** acampbell has joined #openstack-swift | 14:40 | |
| *** acampbel11 has joined #openstack-swift | 14:41 | |
| acoles | zaitcev: you had a +2 on patch 177836 (thx!) but I needed to fix a trivial merge conflict, any chance you could vote again and get it landed? | 14:41 |
| patchbot | acoles: https://review.openstack.org/#/c/177836/ | 14:41 |
| zaitcev | acoles: sure | 14:42 |
| acoles | zaitcev: thank you | 14:42 |
| zaitcev | acoles: that patch caused me to spend countless minutes in the bathroom pondering if I ought to bikeshed that yield '' or not. Really, there's no way for this to leak across the TCP connection, so fixing it obviously fixes something unrelated, like a bug in test suite. | 14:43 |
| eikke | any cores willing to +2 trivial patches (cherry-picks) 181834 and 181836 ? | 14:45 |
| *** jistr has joined #openstack-swift | 14:47 | |
| acoles | zaitcev: hmmm, i was pretty sure that the yield '' was causing probe tests to fail, something about the implementation of readline in ssync_sender maybe | 14:50 |
| zaitcev | acoles: that's what I thought too | 14:50 |
| zaitcev | acoles: well, it's probably more tricky than just that... I suspect eventlet gets involved somehow by faking syscalls in an imperfect way | 14:51 |
| *** acampbell111 has joined #openstack-swift | 14:51 | |
| acoles | eikke: idk about other swift cores but I don't have +2 button on backport reviews | 14:52 |
| acoles | notmyname: ^^ how do we get those landed?? | 14:52 |
| *** acampbel11 has quit IRC | 14:52 | |
| *** acampbell has quit IRC | 14:52 | |
| *** acampbell111 has quit IRC | 14:56 | |
| *** CaioBrentano has quit IRC | 14:57 | |
| *** charlesw_ has joined #openstack-swift | 14:57 | |
| *** CaioBrentano has joined #openstack-swift | 14:57 | |
| openstackgerrit | Mark Seger proposed openstack/swift: make sure user is root https://review.openstack.org/186779 | 14:58 |
| eikke | acoles: interesting | 14:58 |
| *** CaioBren_ has joined #openstack-swift | 14:58 | |
| *** CaioBren_ has quit IRC | 14:59 | |
| *** charlesw has quit IRC | 14:59 | |
| acoles | eikke: hopefully notmyname can shed some light on that | 14:59 |
| *** charlesw_ is now known as charlesw | 14:59 | |
| *** esker has quit IRC | 14:59 | |
| *** annegentle has quit IRC | 14:59 | |
| *** acampbell has joined #openstack-swift | 15:00 | |
| *** acampbel11 has joined #openstack-swift | 15:00 | |
| *** esker has joined #openstack-swift | 15:00 | |
| *** CaioBre__ has joined #openstack-swift | 15:01 | |
| *** CaioBrentano has quit IRC | 15:02 | |
| *** hseipp has left #openstack-swift | 15:11 | |
| *** CaioBre__ has quit IRC | 15:17 | |
| *** jordanP has quit IRC | 15:23 | |
| *** CaioBrentano has joined #openstack-swift | 15:29 | |
| *** CaioBrentano has quit IRC | 15:29 | |
| *** annegentle has joined #openstack-swift | 15:30 | |
| *** blmartin has joined #openstack-swift | 15:31 | |
| *** CaioBrentano has joined #openstack-swift | 15:31 | |
| *** CaioBrentano has quit IRC | 15:34 | |
| *** lyrrad has quit IRC | 15:36 | |
| *** gyee has joined #openstack-swift | 15:41 | |
| *** annegentle has quit IRC | 15:42 | |
| *** annegentle has joined #openstack-swift | 15:43 | |
| *** kutija has joined #openstack-swift | 15:46 | |
| cschwede | acoles: eikke: only PTLs and the stable maintainers can +2/approve patches on stable branches | 15:46 |
| acoles | cschwede: ok. that explains. thx! | 15:46 |
| openstackgerrit | Merged openstack/swift: Make SSYNC receiver return a reponse when initial checks fail https://review.openstack.org/177836 | 15:47 |
| openstackgerrit | Merged openstack/swift: Remove _ensure_flush() from SSYNC receiver https://review.openstack.org/177837 | 15:50 |
| *** nadeem has joined #openstack-swift | 15:56 | |
| *** nadeem has quit IRC | 15:56 | |
| *** jistr has quit IRC | 15:58 | |
| *** theanalyst has quit IRC | 15:58 | |
| *** SkyRocknRoll has quit IRC | 15:59 | |
| *** annegentle has quit IRC | 16:01 | |
| *** lyrrad has joined #openstack-swift | 16:02 | |
| *** theanalyst has joined #openstack-swift | 16:02 | |
| *** jistr has joined #openstack-swift | 16:06 | |
| *** jistr has quit IRC | 16:09 | |
| *** acoles is now known as acoles_away | 16:30 | |
| *** hseipp has joined #openstack-swift | 16:34 | |
| *** hseipp has quit IRC | 16:35 | |
| *** hseipp has joined #openstack-swift | 16:35 | |
| *** SkyRocknRoll has joined #openstack-swift | 16:35 | |
| notmyname | good morning | 16:36 |
| *** SkyRocknRoll has quit IRC | 16:36 | |
| notmyname | eikke: acoles_away: patch 181834 | 16:37 |
| patchbot | notmyname: https://review.openstack.org/#/c/181834/ | 16:37 |
| notmyname | patch 181836 | 16:37 |
| patchbot | notmyname: https://review.openstack.org/#/c/181836/ | 16:37 |
| notmyname | I'll look at those and land as appropriate. seeing all the core +1s, it shouldn't be a problem | 16:38 |
| *** silor has quit IRC | 16:38 | |
| *** esker has quit IRC | 16:38 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/python-swiftclient: Updated from global requirements https://review.openstack.org/89250 | 16:42 |
| *** SkyRocknRoll has joined #openstack-swift | 16:42 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Updated from global requirements https://review.openstack.org/88736 | 16:43 |
| *** acoles_away is now known as acoles | 16:43 | |
| acoles | notmyname: thx | 16:43 |
| *** lpabon has quit IRC | 16:44 | |
| *** charlesw has quit IRC | 16:45 | |
| *** charlesw has joined #openstack-swift | 16:45 | |
| *** acoles is now known as acoles_away | 16:50 | |
| *** lpabon has joined #openstack-swift | 16:50 | |
| *** lpabon has quit IRC | 16:54 | |
| notmyname | acoles_away: eikke: done | 16:59 |
| notmyname | thanks for bringing it to my attention | 16:59 |
| *** lyrrad has quit IRC | 17:04 | |
| *** Fin1te has joined #openstack-swift | 17:07 | |
| *** Shimoda has quit IRC | 17:09 | |
| *** Fin1te has quit IRC | 17:12 | |
| *** barra204 has quit IRC | 17:26 | |
| openstackgerrit | Christian Schwede proposed openstack/swift-specs: Add object undelete spec https://review.openstack.org/155421 | 17:28 |
| *** aix has quit IRC | 17:31 | |
| *** CaioBrentano has joined #openstack-swift | 17:41 | |
| *** geaaru has quit IRC | 17:43 | |
| *** annegentle has joined #openstack-swift | 17:45 | |
| openstackgerrit | Merged openstack/swift: X-Auth-Token should be a bytestring. https://review.openstack.org/181834 | 17:49 |
| openstackgerrit | Merged openstack/swift: X-Auth-Token should be a bytestring. https://review.openstack.org/181836 | 17:49 |
| clayg | jrichli: sorry - got caught up in a issue or two yesterday - when you get online, try me again to look into those test failures for encryption | 18:07 |
| *** barra204 has joined #openstack-swift | 18:08 | |
| *** annegentle has quit IRC | 18:08 | |
| *** annegentle has joined #openstack-swift | 18:09 | |
| *** lyrrad has joined #openstack-swift | 18:09 | |
| *** Fin1te has joined #openstack-swift | 18:10 | |
| jrichli | clayg: thanks. i gotta give a phone interview now | 18:13 |
| clayg | jrichli: yeah i hear ya :( | 18:14 |
| clayg | jrichli: er... I mean - good luck! hope you catch a good one! | 18:14 |
| *** barra204 has quit IRC | 18:15 | |
| jrichli | lol :-) | 18:15 |
| *** lpabon has joined #openstack-swift | 18:16 | |
| *** barra204 has joined #openstack-swift | 18:19 | |
| *** lpabon has quit IRC | 18:21 | |
| *** barra204_ has joined #openstack-swift | 18:27 | |
| *** barra204 has quit IRC | 18:31 | |
| *** hseipp has left #openstack-swift | 18:34 | |
| *** SkyRocknRoll has quit IRC | 18:34 | |
| *** annegentle has quit IRC | 18:41 | |
| *** lpabon has joined #openstack-swift | 18:48 | |
| *** barra204_ is now known as shakamunyi | 18:52 | |
| *** fthiagogv has quit IRC | 18:55 | |
| *** aix has joined #openstack-swift | 19:00 | |
| *** CaioBrentano has quit IRC | 19:03 | |
| *** CaioBrentano has joined #openstack-swift | 19:03 | |
| *** mmcardle has quit IRC | 19:04 | |
| *** B4rker has joined #openstack-swift | 19:10 | |
| *** silor has joined #openstack-swift | 19:16 | |
| *** acampbel11 has quit IRC | 19:32 | |
| *** acampbell has quit IRC | 19:32 | |
| *** lpabon has quit IRC | 19:37 | |
| *** silor has quit IRC | 19:37 | |
| *** CaioBrentano has quit IRC | 19:43 | |
| *** acampbell has joined #openstack-swift | 19:48 | |
| *** acampbel11 has joined #openstack-swift | 19:48 | |
| openstackgerrit | Thiago da Silva proposed openstack/swift: versioned writes middleware https://review.openstack.org/134347 | 19:59 |
| tdasilva | zaitcev, clayg: ^^^ found a little bug concerning the 'allow_versioned_writes' flag, which is fixed now, overall I think it is good to go. | 20:01 |
| zaitcev | tdasilva: checking | 20:01 |
| tdasilva | oh, I also added some more documentation to explain the whole idea of backwards compatibility and how in a future release the container flag would be deprecated... | 20:02 |
| tdasilva | zaitcev: thanks! | 20:02 |
| *** bkopilov has quit IRC | 20:04 | |
| *** openstack has joined #openstack-swift | 20:06 | |
| *** tdasilva has quit IRC | 20:08 | |
| *** openstack has joined #openstack-swift | 20:31 | |
| *** Fin1te has quit IRC | 20:33 | |
| *** peluse has quit IRC | 20:40 | |
| *** acampbel11 has quit IRC | 20:41 | |
| *** acampbell has quit IRC | 20:41 | |
| *** peluse has joined #openstack-swift | 20:41 | |
| *** annegentle has joined #openstack-swift | 20:42 | |
| *** _bluev has joined #openstack-swift | 20:49 | |
| *** wbhuber has quit IRC | 20:55 | |
| *** Triveni_ has quit IRC | 20:56 | |
| *** charlesw has quit IRC | 20:56 | |
| *** openstackgerrit has quit IRC | 20:59 | |
| *** openstackgerrit has joined #openstack-swift | 20:59 | |
| *** annegentle has quit IRC | 21:17 | |
| *** jrichli has quit IRC | 21:24 | |
| *** shakamunyi has quit IRC | 21:26 | |
| *** lcurtis has joined #openstack-swift | 21:29 | |
| *** openstack has joined #openstack-swift | 21:32 | |
| lcurtis | hello all...question...if I have 3 replicas...two local and 1 at remote data center, and I lose both at local, will cluster pull from the remote when object is requsted? | 21:39 |
| lcurtis | and conversely, if i lose disk with 1 replica at the remote site, would it pull from my 'local' site? | 21:40 |
| *** B4rker has quit IRC | 21:42 | |
| notmyname | yes | 21:44 |
| *** annegentle has joined #openstack-swift | 21:46 | |
| lcurtis | notmyname: would it automatically replicate or simply pull when requested? | 21:48 |
| ctennis | both will happen, independent of each other | 21:50 |
| ctennis | GET requests will go to your 1 replica and continue going there until it is replicated elsewhere. That may happen sometime later depending on how fast replication occurs. | 21:51 |
| *** minwoob has quit IRC | 21:51 | |
| lcurtis | ctennis: thanks! | 21:57 |
| *** blmartin has quit IRC | 22:13 | |
| *** annegentle has quit IRC | 22:18 | |
| *** annegentle has joined #openstack-swift | 22:20 | |
| *** mwheckmann has quit IRC | 22:24 | |
| *** lpabon has joined #openstack-swift | 22:28 | |
| *** aix has quit IRC | 22:39 | |
| *** annegentle has quit IRC | 22:40 | |
| openstackgerrit | Darrell Bishop proposed openstack/swift: Allow one object-server per disk deployment https://review.openstack.org/184189 | 22:42 |
| *** lpabon has quit IRC | 22:42 | |
| *** B4rker has joined #openstack-swift | 22:43 | |
| swifterdarrell | for anyone following along at home, ^^^^^^^ consolidates my two patch sets and got marked WIP while I add the ability to do N workers per port | 22:45 |
| swifterdarrell | zaitcev: sure enough, benchmarking showed that more than one worker per port will be necessary to actually do better than "normal" when all disks are healthy | 22:46 |
| *** josed has joined #openstack-swift | 22:48 | |
| zaitcev | swifterdarrell: interesting | 22:56 |
| swifterdarrell | zaitcev: the approach will be to have the initial process inspect the config for the server_per_port option and additional server_per_port_worker_count (defaulting to something based on my benchmarking); I'm avoiding the existing worker count config var to avoid backwards compat biz and to ensure only adding server_per_port=True doesn't result in way too many processes | 22:58 |
| zaitcev | Hmm... | 22:58 |
| zaitcev | You know, I only noticed because we have too many confusing variables as it is. | 22:59 |
| openstackgerrit | Merged openstack/swift: go: replace ghetto getpwnam with os/user https://review.openstack.org/183635 | 22:59 |
| swifterdarrell | zaitcev: that first process will then bind server_per_port_worker_count listen sockets for each port and then fire off the N workers (where N = drive_count * spp_worker_count), telling each worker which socket to actually use for accepting. | 22:59 |
| swifterdarrell | zaitcev: that first proc will then handle HUP/TERM, and process restarting (kind of like how my patch does now, where it knows which PIDs map to which port, which allows it to know which socket for a newly-forked child to listen on) | 23:00 |
| zaitcev | swifterdarrell: okay | 23:01 |
| swifterdarrell | zaitcev: anyway, it'll be AWESOME! you'll see! | 23:06 |
| zaitcev | swifterdarrell: bring it on! | 23:06 |
| swifterdarrell | zaitcev: and whatever my next patch I stand by is, it'll come with some solid numbers from a 2 beefy proxies + 10G network + 2 29-disk storage nodes (the best test platform I could put together on short notice) | 23:07 |
| *** _bluev1 has joined #openstack-swift | 23:07 | |
| swifterdarrell | zaitcev: still decent non-trivial bare-metal | 23:07 |
| swifterdarrell | zaitcev: learned that our guy who saw good results was in AWS and also was using 4 workers unix procs per disk (port); so I'm not dead in the water yet just because my initial numbers weren't great | 23:08 |
| swifterdarrell | zaitcev: (btw, latencies went slightly up w/server-per-port, but req/s was about even--so it wasn't very much worse either, which can't be said for thread_per_disk whcih I'm benchmarking right now and its req/s is way lower) | 23:09 |
| swifterdarrell | zaitcev: hope to have new patchset + numbers early next week | 23:09 |
| *** _bluev has quit IRC | 23:09 | |
| zaitcev | wait, wait, what thread_per_disk | 23:12 |
| swifterdarrell | zaitcev: WAT? it's been in swift for a while | 23:16 |
| swifterdarrell | zaitcev: torgomatic added it | 23:16 |
| swifterdarrell | zaitcev: provides isolation for tarpit or otherwise hammered disks | 23:16 |
| swifterdarrell | zaitcev: but at the overhead of throwing all I/O calls through to a thread from a pool (with one threadpool per disk, in each object-server) | 23:17 |
| swifterdarrell | zaitcev: we'd planned for it to pretty much be able to be used all the time instead of "normal" style, but in practice overhead was too high (among other things the thread count is (drive_count * server_count * threads_per_disk) whcih can get quite high and thrash the kernel's scheduling | 23:18 |
| zaitcev | oh | 23:19 |
| *** B4rker has quit IRC | 23:21 | |
| *** charlesw has joined #openstack-swift | 23:29 | |
| *** lcurtis has quit IRC | 23:31 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!