| *** mmcardle has joined #openstack-swift | 00:19 | |
| *** bach has joined #openstack-swift | 00:19 | |
| *** shakamunyi has joined #openstack-swift | 00:23 | |
| *** zhiyan_ is now known as zhiyan | 00:26 | |
| *** yuan has quit IRC | 00:27 | |
| *** mmcardle has quit IRC | 00:30 | |
| *** peluse has joined #openstack-swift | 00:33 | |
| *** ams0 has joined #openstack-swift | 00:38 | |
| *** ams0 has quit IRC | 00:43 | |
| *** shakamunyi has quit IRC | 00:48 | |
| *** bach has quit IRC | 00:51 | |
| *** shakamunyi has joined #openstack-swift | 01:13 | |
| *** saschpe has quit IRC | 01:22 | |
| *** saschpe has joined #openstack-swift | 01:24 | |
| *** mmcardle has joined #openstack-swift | 01:27 | |
| *** ams0 has joined #openstack-swift | 01:38 | |
| *** mmcardle has quit IRC | 01:39 | |
| *** ams0 has quit IRC | 01:43 | |
| *** bach has joined #openstack-swift | 01:44 | |
| *** mmcardle has joined #openstack-swift | 02:36 | |
| *** mmcardle has quit IRC | 02:49 | |
| *** praveenkumar has joined #openstack-swift | 02:55 | |
| *** shakamunyi has quit IRC | 02:57 | |
| *** piousbox_ has quit IRC | 03:34 | |
| *** tzumainn has quit IRC | 03:36 | |
| *** ams0 has joined #openstack-swift | 03:39 | |
| *** ams0 has quit IRC | 03:43 | |
| *** mmcardle has joined #openstack-swift | 03:46 | |
| *** mmcardle has quit IRC | 04:00 | |
| *** ppai has joined #openstack-swift | 04:26 | |
| *** haomaiwang has joined #openstack-swift | 04:31 | |
| *** haomaiwang has quit IRC | 04:32 | |
| *** haomaiwang has joined #openstack-swift | 04:33 | |
| *** mmcardle has joined #openstack-swift | 04:56 | |
| *** ams0 has joined #openstack-swift | 05:08 | |
| *** mmcardle has quit IRC | 05:10 | |
| *** mlipchuk has joined #openstack-swift | 05:11 | |
| *** nshaikh has joined #openstack-swift | 05:15 | |
| *** haomai___ has joined #openstack-swift | 05:22 | |
| *** haomaiwang has quit IRC | 05:22 | |
| *** ams0 has quit IRC | 05:26 | |
| *** haomaiwang has joined #openstack-swift | 05:26 | |
| *** chandan_kumar has joined #openstack-swift | 05:26 | |
| *** ams0 has joined #openstack-swift | 05:26 | |
| *** haomai___ has quit IRC | 05:29 | |
| *** ams0 has quit IRC | 05:30 | |
| *** ams0 has joined #openstack-swift | 05:56 | |
| *** mmcardle has joined #openstack-swift | 06:06 | |
| *** mmcardle has quit IRC | 06:19 | |
| *** ams0 has quit IRC | 06:32 | |
| *** cheri has joined #openstack-swift | 06:39 | |
| *** bach has quit IRC | 07:11 | |
| *** mmcardle has joined #openstack-swift | 07:15 | |
| *** bach has joined #openstack-swift | 07:17 | |
| *** psharma has joined #openstack-swift | 07:23 | |
| *** jamie_h has joined #openstack-swift | 07:24 | |
| *** jamie_h has quit IRC | 07:25 | |
| *** RockKuo_Office has joined #openstack-swift | 07:26 | |
| *** mmcardle has quit IRC | 07:27 | |
| *** a1|away is now known as JelleB | 07:37 | |
| *** ams0 has joined #openstack-swift | 07:38 | |
| *** bach has quit IRC | 07:43 | |
| *** openstackgerrit has quit IRC | 07:53 | |
| *** nacim has joined #openstack-swift | 08:00 | |
| *** Diddi has quit IRC | 08:14 | |
| *** Diddi has joined #openstack-swift | 08:14 | |
| *** mmcardle has joined #openstack-swift | 08:16 | |
| *** foexle has joined #openstack-swift | 08:19 | |
| *** ams0 has quit IRC | 08:27 | |
| *** tzumainn has joined #openstack-swift | 08:34 | |
| *** ams0 has joined #openstack-swift | 08:34 | |
| *** ams0 has quit IRC | 08:39 | |
| *** ams0 has joined #openstack-swift | 08:40 | |
| *** ams0 has quit IRC | 08:44 | |
| *** ams0_ has joined #openstack-swift | 08:47 | |
| *** mmcardle has quit IRC | 08:48 | |
| *** ams0_ has quit IRC | 08:59 | |
| *** ams0 has joined #openstack-swift | 09:00 | |
| *** nacim has quit IRC | 09:03 | |
| *** ams0 has quit IRC | 09:05 | |
| *** ams0 has joined #openstack-swift | 09:05 | |
| *** ams0_ has joined #openstack-swift | 09:07 | |
| *** ams0 has quit IRC | 09:10 | |
| *** mmcardle has joined #openstack-swift | 09:16 | |
| *** nacim has joined #openstack-swift | 09:17 | |
| *** haomaiwang has quit IRC | 09:25 | |
| *** mmcardle has quit IRC | 09:27 | |
| *** mkollaro has joined #openstack-swift | 09:33 | |
| *** mmcardle has joined #openstack-swift | 10:25 | |
| *** mmcardle has quit IRC | 10:37 | |
| *** cheri has quit IRC | 10:52 | |
| *** ppai has quit IRC | 10:57 | |
| *** ams0_ has quit IRC | 11:02 | |
| *** tdasilva has joined #openstack-swift | 11:02 | |
| *** ams0 has joined #openstack-swift | 11:02 | |
| *** RockKuo_Office has quit IRC | 11:04 | |
| *** ams0 has quit IRC | 11:06 | |
| *** bvandenh has joined #openstack-swift | 11:09 | |
| *** ppai has joined #openstack-swift | 11:15 | |
| *** haomaiwa_ has joined #openstack-swift | 11:21 | |
| *** bvandenh has quit IRC | 11:26 | |
| *** mmcardle has joined #openstack-swift | 11:33 | |
| *** ppai has quit IRC | 11:42 | |
| *** mmcardle has quit IRC | 11:44 | |
| *** ppai has joined #openstack-swift | 11:55 | |
| *** ppai has quit IRC | 12:00 | |
| *** ams0 has joined #openstack-swift | 12:00 | |
| *** ams0_ has joined #openstack-swift | 12:02 | |
| *** ams0 has quit IRC | 12:06 | |
| *** peluse has quit IRC | 12:15 | |
| *** ams0_ has quit IRC | 12:15 | |
| *** ams0 has joined #openstack-swift | 12:16 | |
| *** ams0_ has joined #openstack-swift | 12:30 | |
| *** ams0 has quit IRC | 12:33 | |
| *** bvandenh has joined #openstack-swift | 12:35 | |
| *** mmcardle has joined #openstack-swift | 12:42 | |
| *** mmcardle has quit IRC | 12:52 | |
| *** mlipchuk has quit IRC | 13:13 | |
| *** foexle has quit IRC | 13:13 | |
| *** foexle has joined #openstack-swift | 13:19 | |
| *** zul has joined #openstack-swift | 13:31 | |
| *** openstackgerrit has joined #openstack-swift | 13:35 | |
| *** mmcardle has joined #openstack-swift | 13:49 | |
| *** openstackgerrit has quit IRC | 13:51 | |
| *** ams0_ has quit IRC | 13:54 | |
| *** ams0 has joined #openstack-swift | 13:54 | |
| *** ams0 has quit IRC | 13:57 | |
| *** ams0 has joined #openstack-swift | 13:57 | |
| *** ams0 has quit IRC | 13:58 | |
| *** mmcardle has quit IRC | 13:58 | |
| *** ams0 has joined #openstack-swift | 13:58 | |
| *** nacim has quit IRC | 14:01 | |
| *** ChanServ sets mode: +v dfg_ | 14:10 | |
| *** dfg_ is now known as dfg | 14:11 | |
| *** nacim has joined #openstack-swift | 14:19 | |
| *** mmcardle has joined #openstack-swift | 14:24 | |
| *** wasmum has joined #openstack-swift | 14:27 | |
| *** bvandenh has quit IRC | 14:27 | |
| wasmum | Frontending Keystone with SSL using Apache mod_wsgi vs Nginx. Thought, performance issues, concerns? | 14:28 |
|---|---|---|
| *** wasmum has quit IRC | 14:31 | |
| *** psharma has quit IRC | 14:33 | |
| *** mmcardle has quit IRC | 14:39 | |
| *** foexle has quit IRC | 14:45 | |
| *** foexle has joined #openstack-swift | 14:46 | |
| *** mwstorer has joined #openstack-swift | 14:48 | |
| *** praveenkumar has quit IRC | 14:49 | |
| *** nacim has quit IRC | 14:50 | |
| *** byeager has joined #openstack-swift | 14:57 | |
| *** nacim has joined #openstack-swift | 14:58 | |
| *** mmcardle has joined #openstack-swift | 15:07 | |
| *** zul has quit IRC | 15:17 | |
| *** mmcardle has quit IRC | 15:18 | |
| *** wklely is now known as wkelly | 15:18 | |
| *** mlipchuk has joined #openstack-swift | 15:19 | |
| *** zul has joined #openstack-swift | 15:19 | |
| *** praveenkumar has joined #openstack-swift | 15:23 | |
| *** zhiyan is now known as zhiyan_ | 15:25 | |
| *** haomaiwa_ has quit IRC | 15:29 | |
| *** haomaiwang has joined #openstack-swift | 15:30 | |
| *** foexle has quit IRC | 15:35 | |
| *** haomaiwang has quit IRC | 15:37 | |
| *** haomaiwa_ has joined #openstack-swift | 15:38 | |
| *** openstackgerrit has joined #openstack-swift | 15:42 | |
| *** haomaiwa_ has quit IRC | 15:43 | |
| *** gyee has joined #openstack-swift | 15:43 | |
| *** byeager has quit IRC | 15:57 | |
| *** byeager has joined #openstack-swift | 15:58 | |
| *** byeager has quit IRC | 15:58 | |
| *** byeager has joined #openstack-swift | 15:59 | |
| *** cheri has joined #openstack-swift | 16:07 | |
| *** ams0__ has joined #openstack-swift | 16:09 | |
| *** ams0 has quit IRC | 16:09 | |
| *** mmcardle has joined #openstack-swift | 16:17 | |
| *** chandan_kumar has quit IRC | 16:19 | |
| notmyname | good morning, world | 16:24 |
| notmyname | wasmum isn't here anymore :-( | 16:25 |
| notmyname | portante: good morning. around? | 16:28 |
| *** mmcardle has quit IRC | 16:30 | |
| *** ams0__ has quit IRC | 16:31 | |
| openstackgerrit | David Goetz proposed a change to openstack/swift: xLO bug with auth tokens expiring during download. https://review.openstack.org/92165 | 16:39 |
| *** mlipchuk has quit IRC | 16:41 | |
| notmyname | dfg: nice | 16:45 |
| dfg | notmyname: well- wait for comment :) | 16:46 |
| *** piousbox_ has joined #openstack-swift | 16:46 | |
| *** zaitcev has joined #openstack-swift | 16:46 | |
| *** ChanServ sets mode: +v zaitcev | 16:46 | |
| notmyname | oh I haven't looked at the code yet, but the commit message sounds good! | 16:46 |
| notmyname | (+2 ship it!) | 16:46 |
| *** xiaowei-swift has joined #openstack-swift | 16:58 | |
| *** nacim has quit IRC | 16:59 | |
| *** tdasilva has left #openstack-swift | 17:01 | |
| *** xiaowei-swift has left #openstack-swift | 17:03 | |
| *** shri has joined #openstack-swift | 17:14 | |
| *** wkelly has quit IRC | 17:20 | |
| *** openstackgerrit has quit IRC | 17:20 | |
| *** wkelly has joined #openstack-swift | 17:22 | |
| notmyname | if you are leading a summit session or giving a swift-related conference talk next week, please come to today's team meeting at 1900UTC | 17:23 |
| notmyname | https://wiki.openstack.org/wiki/Meetings/Swift | 17:23 |
| notmyname | I want to make sure you have everything you need and are ready for next week | 17:24 |
| *** openstackgerrit has joined #openstack-swift | 17:24 | |
| *** zul has quit IRC | 17:24 | |
| *** nshaikh has quit IRC | 17:24 | |
| portante | notmyname: now I am! | 17:24 |
| notmyname | err. not today's meeting. this week's meeting | 17:24 |
| * notmyname normally does prep for the wednesday meeting on wednesday. not monday | 17:24 | |
| portante | notmyname: phew, that I was going to miss today's meeting | 17:25 |
| portante | I plan on attending Wednesday, FWIW | 17:25 |
| notmyname | portante: not related to the meeting, but what's up with the in-process func tests in the gate? I haven't looked at the logs, but I saw a lot of emails saying that it failed | 17:25 |
| *** mlanner has quit IRC | 17:25 | |
| *** zul has joined #openstack-swift | 17:25 | |
| *** mlanner has joined #openstack-swift | 17:28 | |
| portante | yes, the underlying mechanism is using run-unittests.sh, which does not actually support what we want to do, but I am told there is a way to do what we want using run-tox.sh instead | 17:28 |
| portante | however, I have not made further progress on that. :( Sorry | 17:28 |
| portante | notmyname: | 17:28 |
| notmyname | no worries (in general). is there any time-frame for getting it working (either from your end or from -infra--I'm not sure who would do a "fix")? | 17:29 |
| notmyname | portante: will it be this week or after the summit is the biggest question | 17:29 |
| notmyname | portante: IMO, sooner is better because I don't want to train ourselves to ignore the results | 17:30 |
| portante | notmyname: this week, at least that is what I think can be done | 17:31 |
| portante | but I thought it was doable last week, but not knowing how the -infra code actually works, that did not pan out | 17:31 |
| *** zul has quit IRC | 17:31 | |
| notmyname | :-) | 17:32 |
| notmyname | portante: I'm pretty busy with summit prep this week, but let me know if there is anything you need from me | 17:32 |
| portante | will do, thanks | 17:32 |
| *** klrmn has joined #openstack-swift | 17:36 | |
| klrmn | chmouel: swifterdarrell suggested you might be able to loan me a clue about swift + keystone. i got x-container-read to work with '_member_', but x-container-write doesn't appear to be working | 17:38 |
| *** lpabon has joined #openstack-swift | 17:39 | |
| *** cheri has quit IRC | 17:47 | |
| * creiht shakes his fist at pbr | 18:00 | |
| creiht | again | 18:00 |
| creiht | heh | 18:00 |
| *** byeager has quit IRC | 18:01 | |
| *** byeager has joined #openstack-swift | 18:01 | |
| portante | notmyname: I have filed a new commit to try to get things going for the in-process functional tests, but I feel it is just a stab in the dark, so hopefully the -infra team will help out | 18:02 |
| notmyname | ok | 18:02 |
| notmyname | creiht: what, specifically? | 18:02 |
| creiht | just a little thing | 18:02 |
| portante | notmyname: https://review.openstack.org/92174 | 18:02 |
| portante | patchbot p 92174 | 18:02 |
| patchbot | portante: https://review.openstack.org/#/c/92174/ | 18:02 |
| *** byeager has quit IRC | 18:02 | |
| *** byeager_ has joined #openstack-swift | 18:02 | |
| creiht | notmyname: where does the rc1 come from in the version number when we are in release candidate phase? | 18:03 |
| creiht | we last built packages during rc phase | 18:04 |
| creiht | so version number was 1.13.1.rc1.25 | 18:04 |
| creiht | current version is 1.13.1.62 | 18:04 |
| creiht | rc1 > 62 for versions | 18:04 |
| * notmyname joins creiht in fist-shaking | 18:05 | |
| creiht | lol | 18:05 |
| notmyname | so, ya, that looks really bad. and probably a bug. but I'm not sure if it's in pbr or in the process (probably process) | 18:05 |
| notmyname | creiht: pbr looks for the first signed tag that is an ancestor of the current branch | 18:06 |
| creiht | yeah I was thinking the same, but I like blaming everything on pbr ;) | 18:06 |
| notmyname | heh | 18:06 |
| notmyname | so what happens is that at the openstack release time (six month time), the milestone-proposed branch is created and the rc1 signed tag is created | 18:07 |
| notmyname | since there are no commits on m-p, then the rc1 tag is an ancestor of any subsequent commits that land on master | 18:07 |
| notmyname | so... | 18:08 |
| notmyname | once the final version is approved, the m-p branch is tagged with the final release number and then merged to master | 18:08 |
| notmyname | when then makes any patches landing on master after the merge commit use the final tag | 18:09 |
| notmyname | this happens in every swift release, but in the openstack cycle releases this takes a lot longer (ie weeks) | 18:09 |
| creiht | notmyname: can you just push a 1.13.2 release? ;) | 18:09 |
| notmyname | (one) downside of this process is that the "master" version is wrong until the m-p branch is merged | 18:10 |
| notmyname | but it also seems that the rc version itself is a major problem since a later tag is getting released with an older version number | 18:11 |
| notmyname | creiht: ya, I think that a 1.13.2 tag is probably the only way to fix this | 18:11 |
| notmyname | creiht: what is the version compare tool that shows how they compare? | 18:11 |
| notmyname | the debian package thing | 18:11 |
| creiht | dch | 18:11 |
| creiht | I think? | 18:11 |
| notmyname | $ sudo apt-get install dch | 18:13 |
| notmyname | Reading package lists... Done | 18:13 |
| notmyname | Building dependency tree | 18:13 |
| notmyname | Reading state information... Done | 18:13 |
| notmyname | E: Unable to locate package dch | 18:13 |
| notmyname | pandemicsyn: ? swifterdarrell: ? | 18:13 |
| creiht | devscripts | 18:13 |
| creiht | or something like that is the pkg | 18:13 |
| creiht | if you type dch in the command line it will tell you | 18:13 |
| notmyname | dpkg --compare-versions | 18:14 |
| creiht | but dch may not be the right thing | 18:14 |
| *** bvandenh has joined #openstack-swift | 18:15 | |
| notmyname | creiht: also, I think you should probably always be using the env vars to set the version number when you build packages. that may have avoided the issue. (the issue here being that the version tagged for rc is clearly wrong) | 18:18 |
| creiht | heh... well we don't use pbr when building packages | 18:19 |
| creiht | other than to determine what the current verion is | 18:19 |
| notmyname | creiht: so in the mean time, you should use the env vars. I doubt I can get ttx to tag something new by the end of the day today (he may still be on vacation). I'll see what I can do | 18:20 |
| creiht | yeah no worries, we will figure something else out short term | 18:21 |
| notmyname | creiht: http://docs.openstack.org/developer/pbr/packagers.html <-- use pbr (ie don't fork setup.py) but don't use pbr's versions | 18:21 |
| creiht | except pbr will hose our build system | 18:21 |
| *** bach has joined #openstack-swift | 18:22 | |
| notmyname | creiht: please write that down in as much detail that you can and email it to me. | 18:22 |
| creiht | heh | 18:22 |
| creiht | it has been a while | 18:22 |
| creiht | but we have been through that before with those that should not be named | 18:23 |
| creiht | :) | 18:23 |
| creiht | last time it boiled down to "We don't support Lucid" | 18:26 |
| notmyname | creiht: I know that it's a sore spot. please ignore "those that should not be named" and give me the tools to help. I've seen both arguments presented, and frankly both side have done a _terrible_ job at giving clear arguments | 18:26 |
| *** zul has joined #openstack-swift | 18:26 | |
| creiht | I'm just kinda tired of it | 18:29 |
| creiht | We have our workarounds | 18:29 |
| creiht | we can live with it | 18:29 |
| notmyname | as do we at swiftstack (using the env vars) | 18:29 |
| creiht | at the end of the day it also boils down to a difference in philosophy | 18:29 |
| notmyname | creiht: I'm trying to talk in-person next week to some -infra team members, and this would be a great topic. but unfortunately there are more than technical reasons for pbr (I know you despise that). and right now, both sides have pretty much said "well I don't agree with the other guy because he doesn't have my use case". | 18:29 |
| creiht | and you can never "win" those discussions | 18:30 |
| notmyname | it's not about winning | 18:30 |
| notmyname | unless "winning" is something other than "I will never talk to those people again" | 18:31 |
| creiht | well I use the term win loosely | 18:31 |
| notmyname | which is where I think things are now. which is terrible | 18:31 |
| creiht | meh | 18:31 |
| creiht | I've resided to the fact that none of that stuff is ever going to get better | 18:31 |
| notmyname | and you're right that there are philosophical differences. which is also why IRC is a terrible medium for discussing it | 18:31 |
| creiht | and try to focus on trying to make things good where I can | 18:32 |
| creiht | that add more value | 18:32 |
| creiht | as long as there are very conflicting philosophies I doubt you will ever come to a happy medium | 18:32 |
| creiht | if you can awesome | 18:33 |
| creiht | but I've decided that it isn't work the effort | 18:33 |
| creiht | for me at least | 18:33 |
| clayg | well it just sucks, because now that trusty is out we're going to be having the same fights for precise - they can't just keep saying "oh that's two or three years old we don't support that anymore" | 18:34 |
| creiht | heh | 18:34 |
| creiht | well and lucid is still good for another year | 18:35 |
| notmyname | actually that's why I'd love to get very specific details from you. so that you don't have to worry about it and I can make arguments beyond "well that tool doesn't work for deployer X" | 18:35 |
| notmyname | clayg: ++ | 18:35 |
| *** sunglasses has quit IRC | 18:35 | |
| creiht | notmyname: but that is the problem... the last time it boiled down to us telling them to try installing it on lucid, and getting told it wasn't supported | 18:36 |
| creiht | not sure how it can be any more clear | 18:36 |
| *** sunglasses has joined #openstack-swift | 18:36 | |
| notmyname | creiht: so dump a screen session of installing it on lucid and email it to me. rather than "just go try it an you'll see it's hard". you've already got that I think, since you've run in to the problems. basically a bug report. not "it's busted" | 18:37 |
| openstackgerrit | Clay Gerrard proposed a change to openstack/swift: Make PBR based setup completely optional https://review.openstack.org/77091 | 18:37 |
| clayg | ^ still think thats pretty reasonable | 18:38 |
| creiht | notmyname: but that is the problem... go through the effort to get it marked as invalid | 18:38 |
| creiht | because it isn't "supported" | 18:38 |
| sunglasses | Hi everyone! I was wondering what does Swift do in the background when someone requests to read data and while the client is reading the object, another client starts sending request to modify the same object. The proxy server does not spool any data. From my tests, I have not had any issue w/ it. Just curious if this is always guaranteed. | 18:38 |
| notmyname | creiht: also, to be clear, the arguments for keeping pbr are very bad too (basically, "it makes installing from source easier for -infra") and the resistance to something other than a one-size-fits-all solution is huge | 18:39 |
| sunglasses | Also, what happens in a similar situation when the reading client flags the query to return the latest ojbect. | 18:39 |
| clayg | sunglasses: that's just how file systems work? get a open file handle and then delete the file - see what you get when you read off the open file handle | 18:39 |
| clayg | sunglasses: well X-newest isn't really as atomic as you might imagine, but whatever is available when the request starts is what's going to come out... | 18:40 |
| sunglasses | Excellent. That's what I had assumed. I just wanted to make sure. | 18:40 |
| clayg | sunglasses: then you should test it - don't believe some random johnny on the internet - unless it's creiht... he's pretty smart. you can trust him. | 18:41 |
| creiht | hah | 18:41 |
| creiht | sunglasses: well the new upload spools to a tmp dir until it is complete and then moved | 18:41 |
| creiht | then unlinks the older versions | 18:41 |
| creiht | so depends on what the os does with the read to the unlinked file descriptor | 18:42 |
| creiht | which I'm not really sure off the top of my head :) | 18:42 |
| clayg | ^ sunglasses nm - ask redbo... he's pretty smart. you can trust him. | 18:43 |
| clayg | oh oh - or portante! | 18:43 |
| creiht | hehe | 18:43 |
| clayg | ;) | 18:43 |
| redbo | It's what you said. We dpmIf you have an open filehandle and you're reading from it, then someone else unlinks the file, you can keep reading from it until you close your file handle. | 18:45 |
| redbo | er.. hit enter while editing. We don't modify files in-place. | 18:45 |
| * clayg closes google.com?dpmIf tab | 18:46 | |
| *** rturk-away is now known as rturk | 18:58 | |
| *** rturk has quit IRC | 18:59 | |
| pandemicsyn | clayg: heh google tried to autocorrect that for me to something very nsfw | 19:00 |
| redbo | I think it takes search history into account when autocorrecting | 19:01 |
| glange | I thought "google.com?dpmIf" was the german word for mother? | 19:02 |
| creiht | this isn't going anywhere good | 19:03 |
| *** bvandenh has quit IRC | 19:05 | |
| *** scotticus has joined #openstack-swift | 19:36 | |
| sunglasses | okay thhanks | 19:38 |
| *** mkollaro has quit IRC | 19:39 | |
| *** mkollaro has joined #openstack-swift | 19:43 | |
| *** ams0 has joined #openstack-swift | 20:01 | |
| *** wer has quit IRC | 20:10 | |
| *** ams0 has quit IRC | 20:11 | |
| *** wer has joined #openstack-swift | 20:13 | |
| *** fbo_away has joined #openstack-swift | 20:15 | |
| *** physcx has quit IRC | 20:16 | |
| *** tdasilva has joined #openstack-swift | 20:22 | |
| *** tnewsome has joined #openstack-swift | 20:22 | |
| *** occup4nt is now known as occupant | 20:22 | |
| *** wer has quit IRC | 20:23 | |
| tnewsome | How do I list all the objects in a container? | 20:23 |
| tnewsome | Eg. 'swift list' only gives me 10,000 of them. | 20:24 |
| *** wer has joined #openstack-swift | 20:24 | |
| clayg | i thought the cli would do the marker queries? | 20:24 |
| tnewsome | Maybe my CLI is too old. | 20:25 |
| clayg | hrmm... code sure does look like it wants to... while True and all that... | 20:25 |
| tnewsome | Is there a doc about the kind of queries I need to do? | 20:25 |
| tnewsome | I don't actually want to use the CLI. I just used it as a convenient example. | 20:25 |
| clayg | you just make the GET /v1/account/container?marker=<last-object-in-the-previous-GET> | 20:25 |
| *** lpabon has quit IRC | 20:25 | |
| tnewsome | clayg: Will do. Thank you. | 20:26 |
| tnewsome | clayg: Is there any indication in a response that more queries are required? | 20:27 |
| *** ams0 has joined #openstack-swift | 20:28 | |
| clayg | nope, when doing container listings you always wanna make another request with the last object as the marker until you get back an empty list | 20:28 |
| clayg | then you only have to worry about objects that were uploaded while you were doing the listings ;) | 20:28 |
| tnewsome | I'm not worried about those. | 20:29 |
| *** bach has quit IRC | 20:29 | |
| *** bach has joined #openstack-swift | 20:29 | |
| tnewsome | clayg: Works perfectly. | 20:39 |
| *** physcx has joined #openstack-swift | 20:43 | |
| openstackgerrit | Clay Gerrard proposed a change to openstack/swift: cleanup policy parsing a bit more https://review.openstack.org/87078 | 20:47 |
| *** ams0 has quit IRC | 20:54 | |
| *** ams0 has joined #openstack-swift | 21:13 | |
| *** ams0 has quit IRC | 21:24 | |
| openstackgerrit | A change was merged to openstack/swift: Add Storage Policy Support to Recon Middleware https://review.openstack.org/87387 | 21:24 |
| *** ams0 has joined #openstack-swift | 21:24 | |
| *** ams0_ has joined #openstack-swift | 21:30 | |
| *** ams0 has quit IRC | 21:30 | |
| *** ams0_ has quit IRC | 21:35 | |
| peluse_ | ^ that had to be a record on rechecks at least recently anyway | 21:35 |
| *** tdasilva has left #openstack-swift | 21:37 | |
| tnewsome | What's the best way to delete 100k objects? Multiple threads sending requests? | 21:40 |
| physcx | are you deleting a specific 100k objects or are you removing all objects stored? | 21:41 |
| tnewsome | Specific objects. | 21:42 |
| tnewsome | Basically delete-at isn't flexible enough for my application, and I'm expiring based on a different database. | 21:42 |
| tnewsome | With my naive implementation I'm deleting about 260 objects/minute. | 21:42 |
| tnewsome | I suspect I could get to at least 1000 with multi-threading, but I haven't tested it. | 21:43 |
| tnewsome | (I'm not using http keep-alive because I'm limited to an ancient requests library currently.) | 21:43 |
| physcx | sorry I don't have any advice for that, I'm doing performance testing on some changes and routinely need to clear out all data so depending on how much is in there we either rm -rf on the disk or reformat as fastest way but thats not gonna help you lol | 21:44 |
| shri | doesn't bulk delete have ability to delete 1000 objects with just one request? | 21:44 |
| shri | reduces rtts | 21:44 |
| tnewsome | shri: That sounds promising. | 21:45 |
| notmyname | tnewsome: ya, bunk delete + concurrency + time | 21:46 |
| tnewsome | notmyname: Is there any point to concurrency if bulk delete can do 1k at once? | 21:47 |
| notmyname | yes, if you want to do 2k at once ;-) | 21:47 |
| tnewsome | notmyname: I assume that it would keep the db locked while processing. | 21:47 |
| notmyname | sorry, not meaning to be snarky :-) | 21:47 |
| notmyname | concurrency is to make sure your network stays saturated | 21:47 |
| tnewsome | notmyname: Snark is fine. :-) | 21:48 |
| tnewsome | notmyname: I just figured my network would become relatively insignificant if the bulk is large enough. | 21:48 |
| notmyname | tnewsome: ya, you've obviously got to balance the network saturation with your cpu saturation and your internal network congestion | 21:48 |
| notmyname | tnewsome: so I guess it's actually bulk delete + time [+ concurrency] | 21:50 |
| tnewsome | The docstring for bulk.py is far more insightful than the manual online. | 21:51 |
| tnewsome | And it does sound awesome. :-) | 21:51 |
| notmyname | ya, I was just reminded that we need better client docs for the bulk middleware | 21:52 |
| clayg | notmyname: can we just give out dfg's cell phone number? | 21:52 |
| notmyname | tnewsome: the docstring is automatically at http://docs.openstack.org/developer/swift/middleware.html#module-swift.common.middleware.bulk | 21:52 |
| notmyname | clayg: :-) | 21:53 |
| tnewsome | notmyname: Google sent me to http://docs.openstack.org/trunk/config-reference/content/object-storage-bulk-delete.html | 21:55 |
| *** ams0 has joined #openstack-swift | 21:55 | |
| *** ams0 has quit IRC | 21:58 | |
| *** csd has joined #openstack-swift | 21:58 | |
| *** ams0 has joined #openstack-swift | 21:58 | |
| *** ams0 has quit IRC | 22:03 | |
| *** zul has quit IRC | 22:03 | |
| *** krtaylor has quit IRC | 22:10 | |
| portante | notmyname: not sure I am actually doing anything useful with this in-process functional test stuff, perhaps we can/should ask chmouel to help me here? | 22:11 |
| portante | See https://review.openstack.org/92174 | 22:12 |
| openstackgerrit | A change was merged to openstack/swift: Adding functional tests for Storage Policy https://review.openstack.org/87265 | 22:21 |
| notmyname | portante: I see it, but I don't think I have any more insights that you on it | 22:21 |
| portante | notmyname: k thanks | 22:21 |
| *** zul has joined #openstack-swift | 22:24 | |
| * notmyname was hoping portante would have it all figured out soon so notmyname could learn how to set up the community qa cluster stuff from him | 22:24 | |
| portante | ;) | 22:24 |
| portante | good luck with that | 22:24 |
| *** ams0 has joined #openstack-swift | 22:26 | |
| notmyname | I've had that on my todo list for a while now, but other things have kept me busy. | 22:27 |
| *** zul has quit IRC | 22:39 | |
| *** fbo_away is now known as fbo | 22:46 | |
| *** byeager has joined #openstack-swift | 22:47 | |
| *** byeager_ has quit IRC | 22:48 | |
| openstackgerrit | Clay Gerrard proposed a change to openstack/swift: cleanup policy parsing a bit more https://review.openstack.org/87078 | 22:55 |
| *** byeager_ has joined #openstack-swift | 23:01 | |
| *** byeager has quit IRC | 23:01 | |
| *** byeager_ has quit IRC | 23:01 | |
| *** mkollaro has quit IRC | 23:10 | |
| *** ams0_ has joined #openstack-swift | 23:13 | |
| *** ams0_ has quit IRC | 23:15 | |
| *** ams0_ has joined #openstack-swift | 23:15 | |
| portante | notmyname: thanks for jumpin there | 23:17 |
| *** ams0 has quit IRC | 23:17 | |
| notmyname | portante: looks like clarkb may have fixed it? | 23:17 |
| portante | cool | 23:19 |
| *** ams0_ has quit IRC | 23:20 | |
| openstackgerrit | A change was merged to openstack/swift: Fix X-Backend-Storage-Policy-Index requests https://review.openstack.org/90929 | 23:24 |
| openstackgerrit | A change was merged to openstack/swift: Fix recursive Policy patching https://review.openstack.org/90930 | 23:24 |
| openstackgerrit | A change was merged to openstack/swift: Cleanup container update headers in obj server tests https://review.openstack.org/90931 | 23:24 |
| clayg | peluse_: oh yeah, the container_sync test... i'm trying fix that up now - I think it was just a merge race. | 23:37 |
| *** nmap911 has joined #openstack-swift | 23:41 | |
| nmap911 | Evening | 23:41 |
| nmap911 | Is there a swift mailing list? | 23:42 |
| clayg | folks mostly just use openstack-dev put [swift] in the subject or w/e | 23:42 |
| nmap911 | Is that mailing list ment mainly for code/development discussion? I'd like to get some feedback on performance results im getting out of my dev deployment, to confirm whether its whats expected | 23:44 |
| *** bill_az has joined #openstack-swift | 23:46 | |
| openstackgerrit | A change was merged to openstack/swift: Fix coverage report for newer versions of coverage https://review.openstack.org/90144 | 23:46 |
| nmap911 | I just found ask.openstack - i think i'll stick it up there. thanks for your time | 23:47 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!