| kota_ | good morning | 00:29 |
|---|---|---|
| *** KANSAS has joined #openstack-swift | 00:29 | |
| *** KANSAS has quit IRC | 00:29 | |
| *** zhurong has joined #openstack-swift | 00:45 | |
| *** tovin07_ has joined #openstack-swift | 00:52 | |
| mattoliverau | kota_: morning | 01:08 |
| *** infurno has joined #openstack-swift | 01:15 | |
| *** cshastri has joined #openstack-swift | 01:20 | |
| *** zhurong has quit IRC | 02:01 | |
| openstackgerrit | Matthew Oliver proposed openstack/swift master: Bind SAIO services on different loopback addresses https://review.openstack.org/475202 | 02:48 |
| *** esnyder has quit IRC | 02:49 | |
| *** gkadam has joined #openstack-swift | 03:15 | |
| *** kei_yama has quit IRC | 04:35 | |
| *** links has joined #openstack-swift | 04:40 | |
| *** kei_yama has joined #openstack-swift | 05:14 | |
| *** silor has joined #openstack-swift | 05:29 | |
| *** silor1 has joined #openstack-swift | 05:36 | |
| *** silor has quit IRC | 05:37 | |
| *** silor1 is now known as silor | 05:37 | |
| *** psachin has joined #openstack-swift | 05:44 | |
| *** ChubYann has quit IRC | 05:47 | |
| *** rcernin has joined #openstack-swift | 05:59 | |
| *** cschwede has joined #openstack-swift | 06:07 | |
| *** ChanServ sets mode: +v cschwede | 06:07 | |
| *** skudlik has joined #openstack-swift | 06:11 | |
| *** tesseract has joined #openstack-swift | 06:24 | |
| *** hseipp has joined #openstack-swift | 06:45 | |
| *** pcaruana has joined #openstack-swift | 06:45 | |
| *** klrmn has quit IRC | 06:53 | |
| *** oshritf has joined #openstack-swift | 07:33 | |
| *** pcaruana has quit IRC | 07:58 | |
| *** rcernin has quit IRC | 07:59 | |
| *** geaaru has joined #openstack-swift | 07:59 | |
| *** rcernin has joined #openstack-swift | 08:12 | |
| *** pcaruana has joined #openstack-swift | 08:12 | |
| *** cbartz has joined #openstack-swift | 08:21 | |
| *** oshritf has quit IRC | 08:34 | |
| *** psachin has quit IRC | 08:43 | |
| *** oshritf has joined #openstack-swift | 08:56 | |
| *** rcernin has quit IRC | 09:46 | |
| *** mvk has joined #openstack-swift | 09:47 | |
| *** tovin07_ has quit IRC | 09:59 | |
| *** rcernin has joined #openstack-swift | 10:02 | |
| *** cshastri has quit IRC | 10:19 | |
| *** jaosorior has joined #openstack-swift | 10:30 | |
| *** pcaruana has quit IRC | 10:53 | |
| *** pcaruana has joined #openstack-swift | 11:05 | |
| *** logan- has joined #openstack-swift | 11:13 | |
| *** kei_yama has quit IRC | 11:32 | |
| *** rcernin has quit IRC | 12:02 | |
| *** rcernin has joined #openstack-swift | 12:04 | |
| *** hoonetorg has quit IRC | 12:16 | |
| *** rcernin has quit IRC | 12:21 | |
| *** rcernin has joined #openstack-swift | 12:21 | |
| *** hoonetorg has joined #openstack-swift | 12:28 | |
| *** mvk has quit IRC | 12:50 | |
| *** gkadam has quit IRC | 13:06 | |
| *** lucasxu has joined #openstack-swift | 13:13 | |
| *** mvk has joined #openstack-swift | 13:20 | |
| *** MVenesio has joined #openstack-swift | 13:25 | |
| *** klamath has joined #openstack-swift | 13:34 | |
| *** klamath has quit IRC | 13:34 | |
| *** klamath has joined #openstack-swift | 13:35 | |
| *** chlong has joined #openstack-swift | 13:51 | |
| *** NM has joined #openstack-swift | 13:54 | |
| *** links has quit IRC | 13:58 | |
| *** hseipp has quit IRC | 13:59 | |
| *** chlong has quit IRC | 14:01 | |
| *** NM has quit IRC | 14:01 | |
| *** NM has joined #openstack-swift | 14:06 | |
| *** chlong has joined #openstack-swift | 14:14 | |
| *** stewie925 has quit IRC | 14:28 | |
| *** NM has quit IRC | 14:32 | |
| *** NM has joined #openstack-swift | 14:36 | |
| *** klrmn has joined #openstack-swift | 14:46 | |
| *** skudlik has quit IRC | 14:48 | |
| *** aselius has joined #openstack-swift | 14:48 | |
| *** vint_bra has joined #openstack-swift | 14:56 | |
| *** rcernin has quit IRC | 15:04 | |
| *** links has joined #openstack-swift | 15:17 | |
| *** links has quit IRC | 15:25 | |
| *** chlong has quit IRC | 15:27 | |
| *** NM has quit IRC | 15:37 | |
| *** chlong has joined #openstack-swift | 15:40 | |
| *** NM has joined #openstack-swift | 15:44 | |
| *** mvk has quit IRC | 16:03 | |
| *** pcaruana has quit IRC | 16:03 | |
| *** lucasxu has quit IRC | 16:08 | |
| *** klamath has quit IRC | 16:09 | |
| *** klrmn has quit IRC | 16:10 | |
| *** gyee has joined #openstack-swift | 16:11 | |
| *** cbartz has quit IRC | 16:12 | |
| *** pcaruana has joined #openstack-swift | 16:14 | |
| *** mvk has joined #openstack-swift | 16:16 | |
| *** zaitcev has joined #openstack-swift | 16:20 | |
| *** ChanServ sets mode: +v zaitcev | 16:20 | |
| *** mvk has quit IRC | 16:21 | |
| notmyname | good morning | 16:30 |
| *** klamath has joined #openstack-swift | 16:38 | |
| notmyname | FYI (and I'll say this a few times before then) let's skip the 2100 meeting this wednesday. the 0700 meeting is still on (chaired by cschwede). I'll be updating the meeting agenda today | 16:50 |
| *** NM has quit IRC | 16:53 | |
| *** NM has joined #openstack-swift | 16:56 | |
| *** NM has quit IRC | 17:01 | |
| *** chlong has quit IRC | 17:02 | |
| *** pcaruana has quit IRC | 17:03 | |
| *** NM has joined #openstack-swift | 17:08 | |
| *** vint_bra has quit IRC | 17:10 | |
| *** klrmn has joined #openstack-swift | 17:11 | |
| *** chlong has joined #openstack-swift | 17:13 | |
| *** oshritf has quit IRC | 17:16 | |
| *** vint_bra has joined #openstack-swift | 17:16 | |
| *** lucasxu has joined #openstack-swift | 17:24 | |
| *** MVenesio has quit IRC | 17:27 | |
| *** MVenesio has joined #openstack-swift | 17:27 | |
| *** tonanhngo has joined #openstack-swift | 17:48 | |
| *** mvk has joined #openstack-swift | 17:58 | |
| *** cschwede has quit IRC | 18:00 | |
| *** NM has quit IRC | 18:07 | |
| *** geaaru has quit IRC | 18:14 | |
| *** vint_bra has quit IRC | 18:50 | |
| *** vint_bra has joined #openstack-swift | 19:01 | |
| *** chlong has quit IRC | 19:02 | |
| *** vint_bra has quit IRC | 19:10 | |
| *** ChubYann has joined #openstack-swift | 19:15 | |
| *** chlong has joined #openstack-swift | 19:16 | |
| clayg | did everyone have a nice weekend!? | 19:20 |
| *** chsc has joined #openstack-swift | 19:21 | |
| *** ChanServ has quit IRC | 19:30 | |
| *** ChanServ has joined #openstack-swift | 19:35 | |
| *** card.freenode.net sets mode: +o ChanServ | 19:35 | |
| *** NM has joined #openstack-swift | 19:36 | |
| *** vint_bra has joined #openstack-swift | 19:37 | |
| *** ChanServ sets mode: +v tdasilva | 19:39 | |
| *** silor has quit IRC | 19:40 | |
| *** klamath has quit IRC | 19:53 | |
| *** klamath has joined #openstack-swift | 19:59 | |
| *** tonanhngo has quit IRC | 20:19 | |
| *** tesseract has quit IRC | 20:21 | |
| notmyname | meeting agenda has been updated. cschwede is chairing the 0700 meeting this week. no 2100 meeting this week. | 20:22 |
| notmyname | https://wiki.openstack.org/wiki/Meetings/Swift | 20:22 |
| *** tonanhngo has joined #openstack-swift | 20:23 | |
| *** tonanhngo has quit IRC | 20:28 | |
| *** tonanhngo has joined #openstack-swift | 20:31 | |
| *** tonanhngo has quit IRC | 20:35 | |
| *** tonanhngo has joined #openstack-swift | 20:35 | |
| *** tonanhngo has quit IRC | 20:55 | |
| *** chsc has quit IRC | 21:01 | |
| *** tonanhngo has joined #openstack-swift | 21:02 | |
| *** chsc has joined #openstack-swift | 21:02 | |
| *** chsc has joined #openstack-swift | 21:02 | |
| *** chlong has quit IRC | 21:04 | |
| *** tonanhngo has quit IRC | 21:07 | |
| mattoliverau | Cools, I'm unsure of I'll make the 007 meeting, I hope to, but I'll be doing the long drive to Melbourne so may be stuck in a car.. but I can always pull over I guess :) | 21:10 |
| *** tonanhngo has joined #openstack-swift | 21:26 | |
| *** tonanhngo has quit IRC | 21:27 | |
| *** tonanhngo has joined #openstack-swift | 21:29 | |
| *** vint_bra has quit IRC | 21:38 | |
| clayg | mattoliverau: are you around right now - I can't tell if we're 100% in agreement on the write-affinity delete patch 470158 or only 90%? | 21:41 |
| patchbot | https://review.openstack.org/#/c/470158/ - swift - Write-affinity aware object deletion | 21:41 |
| *** lucasxu has quit IRC | 21:41 | |
| clayg | I feel like maybe I missed a comment where you said "the default should be BLAH, duh" and I'm missing something obvious? | 21:42 |
| mattoliverau | clayg: lol, I think the default should be what ever write affinity is, that is to say, they should use the same setting, and if we need a delete write affinity setting then fine, you can use it to override.. however if we want to just default to off (like we do to not distrupt upgrades), then I'm down as theres a precident and I don't really want to break existing clusters :) | 21:51 |
| clayg | so like the default write_affinity_node_count is 2 * replicas - which is the number of handoffs you will iterate to "fill" the needed replica count of "primary" connections with local devices - but in a three replica cluster that's 6 - which is way too many handoffs to blast tombstones onto I think - is it not? | 21:54 |
| mattoliverau | Yeah, true, I'm just trying to make it simple bad less knobby :) but I hope you into blast TS on the first replica count you find noting primaries will always be firdt | 21:56 |
| *** NM has quit IRC | 21:56 | |
| mattoliverau | a/bad/and/ | 21:56 |
| mattoliverau | Also, sorry on phone and auto complete is a $&#$& | 21:57 |
| clayg | so... the default value could be "replica"? Is *that* correct? [204, 204, 204, 404, 404, 404] would translate to 204 because of the override 404 translation... maybe that works? | 21:58 |
| clayg | i could definitely get tombstones onto the local handoffs... that's good... I think in the author's case it's one useless request - more useless requests on a 2 region cluster... | 22:00 |
| clayg | well... two region is special.... it's at least one useless request maybe more | 22:00 |
| *** tonanhngo has quit IRC | 22:07 | |
| *** tonanhngo has joined #openstack-swift | 22:11 | |
| *** stewie925 has joined #openstack-swift | 22:13 | |
| *** tonanhngo has quit IRC | 22:13 | |
| *** tonanhngo has joined #openstack-swift | 22:14 | |
| *** mrhillsman is now known as codebauss | 22:22 | |
| *** codebauss is now known as mrhillsman | 22:22 | |
| mattoliverau | clayg: ok, moved to my laptop and playing with it some more.. yeah there are alot of ts dropped. question: why do we drop ts on 404s on delete? | 22:46 |
| mattoliverau | I was assuming if we get a 404 a ts wouldn't be dropped. but I can see in the code it is. | 22:48 |
| mattoliverau | Im asusming there is a reason, probably some eventual consistant reason that I'm missing this early in morning (without any coffee) :) | 22:49 |
| clayg | yeah something like that - write down the intent - if three primaries all said 404 and did nothing, then later some old rebalanced off primaries shipped over .data's we could loose a delete | 23:00 |
| *** chsc has quit IRC | 23:03 | |
| mattoliverau | Hmm, right, well doing TS everywhere is annoying. So maybe just replica does make more sense. I mean still an extra request but it will grow with replicas. | 23:08 |
| *** chsc has joined #openstack-swift | 23:11 | |
| *** chsc has joined #openstack-swift | 23:11 | |
| clayg | well, no but I mean 2 * replicas is defnately wrong - [204, 204, 204, 404, 404, 404, 404, 404, 404] would *definitely* 404 - so that's exactly what the author is trying to avoid? it just so happens that the value of 2 or 3 have very similar behaviors... | 23:22 |
| *** chsc has quit IRC | 23:23 | |
| clayg | the only downside of using a default of 'replica' vs. a configured value of 2 is the extra request [204, 204, 204, 404, 404] is necessary and sufficient - [204, 204, 204, 404, 404, 404] is probably also sufficient but not obviously necessary... it seems like most of the time that third local handoff would wiff | 23:24 |
| *** kei_yama has joined #openstack-swift | 23:30 | |
| *** chsc has joined #openstack-swift | 23:30 | |
| *** chsc has joined #openstack-swift | 23:30 | |
| *** tonanhngo has quit IRC | 23:30 | |
| *** klamath has quit IRC | 23:30 | |
| *** tonanhngo has joined #openstack-swift | 23:32 | |
| *** chsc has quit IRC | 23:35 | |
| mattoliverau | Yeah, but hard coded 2 only works in the author's case, what about other configurations and replica numbers. I'd rather have something that works with a caveat for all rather then for just 1. | 23:37 |
| *** tonanhngo has quit IRC | 23:42 | |
| clayg | ah, ok, I can see that - so your suggestion is "make default replica # instead of 0" - I think that's worth considering | 23:44 |
| clayg | and it's still configurable - so you can dump the extra request if needed... | 23:44 |
| clayg | and it only kicks in when write_affinity is turned on (normally you'd not be putting any tombstones on handoffs) ... and in that case write "replica count tombstones to local handoffs" ... i think that makes some sense... | 23:46 |
| *** tonanhngo has joined #openstack-swift | 23:46 | |
| *** tonanhngo has quit IRC | 23:51 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!