| *** tonanhngo has quit IRC | 00:01 | |
| *** tqtran has quit IRC | 00:04 | |
| *** tonanhngo has joined #openstack-swift | 00:15 | |
| *** dmorita has quit IRC | 00:17 | |
| *** dmorita has joined #openstack-swift | 00:19 | |
| *** tonanhngo has quit IRC | 00:19 | |
| *** winggundamth_ has joined #openstack-swift | 00:39 | |
| *** dmorita has quit IRC | 00:53 | |
| *** winggundamth_ has quit IRC | 00:54 | |
| *** dmorita has joined #openstack-swift | 00:54 | |
| *** tonanhngo has joined #openstack-swift | 00:57 | |
| *** tonanhngo has quit IRC | 01:02 | |
| *** sams-gleb has joined #openstack-swift | 01:02 | |
| *** openstackgerrit has quit IRC | 01:02 | |
| *** sams-gleb has quit IRC | 01:07 | |
| *** dmorita has quit IRC | 01:17 | |
| *** dmorita has joined #openstack-swift | 01:18 | |
| *** tonanhngo has joined #openstack-swift | 01:49 | |
| *** JimCheung has quit IRC | 01:50 | |
| *** tonanhngo has quit IRC | 01:50 | |
| *** tonanhngo has joined #openstack-swift | 01:50 | |
| *** tanee is now known as tanee_away | 02:08 | |
| *** Jeffrey4l_ has quit IRC | 02:10 | |
| *** tanee_away is now known as tanee | 02:10 | |
| *** klrmn has quit IRC | 02:12 | |
| *** cdelatte has quit IRC | 02:19 | |
| *** SkyRocknRoll has joined #openstack-swift | 02:23 | |
| *** JimCheung has joined #openstack-swift | 02:31 | |
| *** JimCheung has quit IRC | 02:39 | |
| clayg | isn't there a per-policy proxy sorting option patch floating around? | 02:47 |
|---|---|---|
| clayg | patch 382766 | 02:47 |
| patchbot | https://review.openstack.org/#/c/382766/ - swift - Support Read/Write Affinity Per Policy | 02:47 |
| *** bkopilov has quit IRC | 03:02 | |
| *** openstackgerrit has joined #openstack-swift | 03:15 | |
| openstackgerrit | Clay Gerrard proposed openstack/swift master: WIP: trying to understand global-ec https://review.openstack.org/427994 | 03:15 |
| *** dmorita has quit IRC | 03:25 | |
| *** SkyRocknRoll has quit IRC | 03:30 | |
| openstackgerrit | Merged openstack/swift master: Rebalance with min_part_seconds_left > 0 https://review.openstack.org/326967 | 03:33 |
| *** Jeffrey4l has joined #openstack-swift | 03:41 | |
| *** dmorita has joined #openstack-swift | 03:55 | |
| *** dmorita has quit IRC | 03:59 | |
| timburke | :-( all this reconstructor stuff is prompting a lot of conflicts... | 04:01 |
| mahatic | good morning | 04:15 |
| mahatic | clayg: I like the idea of squashing patch 426336, especially coz the other reviewers don't have to look at another change while reviewing patch 419787 | 04:18 |
| *** psachin has joined #openstack-swift | 04:18 | |
| patchbot | https://review.openstack.org/#/c/426336/ - swift - Fix race when consolidating new partition | 04:18 |
| patchbot | https://review.openstack.org/#/c/419787/ - swift - Better optimistic lock in get_hashes | 04:18 |
| mahatic | but then both just need a +A atm | 04:18 |
| *** bkopilov has joined #openstack-swift | 04:19 | |
| mahatic | if you're gonna have another patchset for 419787 though, squashing makes sense to me | 04:21 |
| *** silor has joined #openstack-swift | 04:37 | |
| *** zaitcev has quit IRC | 04:50 | |
| *** silor has quit IRC | 04:51 | |
| *** tonanhngo has quit IRC | 04:54 | |
| *** sams-gleb has joined #openstack-swift | 05:08 | |
| *** sams-gleb has quit IRC | 05:13 | |
| *** bkopilov_ has joined #openstack-swift | 05:25 | |
| *** ppai has joined #openstack-swift | 05:27 | |
| *** m_kazuhiro has joined #openstack-swift | 05:41 | |
| *** SkyRocknRoll has joined #openstack-swift | 05:45 | |
| m_kazuhiro | good morning | 05:46 |
| *** SkyRocknRoll has quit IRC | 05:49 | |
| *** tonanhngo has joined #openstack-swift | 05:56 | |
| *** gmmaha has quit IRC | 06:11 | |
| *** cschwede has quit IRC | 06:12 | |
| *** kota_ has quit IRC | 06:12 | |
| *** kota_ has joined #openstack-swift | 06:13 | |
| *** ChanServ sets mode: +v kota_ | 06:13 | |
| *** remix_tj has quit IRC | 06:13 | |
| *** swifterdarrell has quit IRC | 06:13 | |
| *** gmmaha has joined #openstack-swift | 06:13 | |
| *** remix_tj has joined #openstack-swift | 06:16 | |
| *** swifterdarrell has joined #openstack-swift | 06:18 | |
| *** ChanServ sets mode: +v swifterdarrell | 06:18 | |
| *** cschwede has joined #openstack-swift | 06:21 | |
| *** Jeffrey4l has quit IRC | 06:25 | |
| *** Jeffrey4l has joined #openstack-swift | 06:26 | |
| *** McMurlock1 has joined #openstack-swift | 06:35 | |
| *** psachin has quit IRC | 06:42 | |
| *** tonanhngo has quit IRC | 06:45 | |
| *** tonanhngo has joined #openstack-swift | 06:45 | |
| *** tonanhngo has quit IRC | 06:46 | |
| *** sams-gleb has joined #openstack-swift | 06:51 | |
| kota_ | clayg: thanks for a lot of comments, will check out! | 06:56 |
| *** psachin has joined #openstack-swift | 06:58 | |
| *** rcernin has joined #openstack-swift | 07:06 | |
| *** wer has quit IRC | 07:13 | |
| *** tesseract has joined #openstack-swift | 07:16 | |
| *** ChubYann has quit IRC | 07:17 | |
| *** psachin has quit IRC | 07:30 | |
| *** m_kazuhiro has quit IRC | 07:36 | |
| *** pcaruana has joined #openstack-swift | 07:39 | |
| *** pcaruana has quit IRC | 07:42 | |
| *** psachin has joined #openstack-swift | 07:42 | |
| *** pcaruana has joined #openstack-swift | 07:43 | |
| mathiasb | good morning! | 07:47 |
| *** wer has joined #openstack-swift | 07:49 | |
| *** oshritf has joined #openstack-swift | 07:49 | |
| *** m_kazuhiro has joined #openstack-swift | 07:57 | |
| *** tonanhngo has joined #openstack-swift | 08:02 | |
| *** tonanhngo has quit IRC | 08:04 | |
| *** rledisez has joined #openstack-swift | 08:08 | |
| *** ma91 has joined #openstack-swift | 08:27 | |
| *** geaaru has joined #openstack-swift | 08:27 | |
| ma91 | Hi, does anybody know, I would need a REST API to transfer data between Swift and a POSIX filesystem , does this exist already? | 08:27 |
| *** rcernin has quit IRC | 08:28 | |
| ma91 | so the REST API would allow listing the POSIX fs, listing the Swift objects and do (asynchronous) transfers from one to the other | 08:28 |
| *** oshritf has quit IRC | 08:31 | |
| *** oshritf has joined #openstack-swift | 08:32 | |
| *** cbartz has joined #openstack-swift | 08:46 | |
| *** rcernin has joined #openstack-swift | 08:52 | |
| *** cbartz has joined #openstack-swift | 08:53 | |
| *** remix_tj has quit IRC | 09:13 | |
| *** oshritf has quit IRC | 09:15 | |
| *** remix_tj has joined #openstack-swift | 09:16 | |
| *** Shashikant86 has joined #openstack-swift | 09:16 | |
| *** jordanP has joined #openstack-swift | 09:22 | |
| *** Jeffrey4l has quit IRC | 09:30 | |
| *** Shashikant86 has quit IRC | 09:30 | |
| *** Jeffrey4l has joined #openstack-swift | 09:31 | |
| cschwede | cbartz: Hello! You can easily check which other middlewares are in the pipeline *after* your middleware (because these got loaded earlier) | 09:45 |
| cschwede | cbartz: another option is to parse the config again in your middleware | 09:46 |
| *** Shashikant86 has joined #openstack-swift | 09:47 | |
| cschwede | cbartz: the conf dictionary has an attribute __file__ that returns the current proxy server config filename | 09:48 |
| cschwede | cbartz: however, if the file changed between start of the proxy and parsing it in the middleware, it's outdated (granted, it's just a short timeframe, but nevertheless might be a reason for a race condition) | 09:49 |
| *** m_kazuhiro has quit IRC | 10:20 | |
| *** Shashikant86 has quit IRC | 10:32 | |
| *** Shashikant86 has joined #openstack-swift | 10:36 | |
| *** sams-gleb has quit IRC | 10:45 | |
| *** sams-gleb has joined #openstack-swift | 10:46 | |
| *** sams-gleb has quit IRC | 10:51 | |
| *** dmorita has joined #openstack-swift | 10:57 | |
| *** lxsli has quit IRC | 10:58 | |
| *** dmorita has quit IRC | 11:02 | |
| *** vinsh has quit IRC | 11:04 | |
| *** kei_yama has quit IRC | 11:07 | |
| *** NM has joined #openstack-swift | 11:09 | |
| *** sams-gleb has joined #openstack-swift | 11:31 | |
| *** psachin has quit IRC | 11:43 | |
| *** Tahvok has joined #openstack-swift | 11:45 | |
| Tahvok | Hey guys! | 11:45 |
| Tahvok | I'm trying to understand some inconsistency between the guide and a conf file in mitaka | 11:46 |
| Tahvok | 'filter:authtoken' section at the bottom here: http://docs.openstack.org/mitaka/install-guide-ubuntu/swift-controller-install.html | 11:46 |
| *** catintheroof has joined #openstack-swift | 11:47 | |
| Tahvok | It says to configure 'memcached_servers', 'auth_type', 'project_domain_name' and 'user_domain_name'. However memcached config does not exist, and the others have a different name: auth_plugin, project_domain_id and user_domain_id | 11:49 |
| Tahvok | So what is the correct name for this configurations? From the guide or from the config file? | 11:50 |
| *** psachin has joined #openstack-swift | 11:55 | |
| *** bkopilov_ has quit IRC | 11:57 | |
| *** bkopilov has quit IRC | 11:58 | |
| Tahvok | I checked, and it's the same issue for newton: http://docs.openstack.org/project-install-guide/object-storage/draft/controller-install-ubuntu.html | 11:59 |
| *** openstackgerrit has quit IRC | 12:03 | |
| *** tonanhngo has joined #openstack-swift | 12:04 | |
| *** tonanhngo has quit IRC | 12:07 | |
| mahatic | Tahvok: Not sure if this is helpful, but Keystone api docs say both are acceptable parameters - http://docs.openstack.org/developer/python-keystoneclient/api/keystoneclient.auth.identity.v3.html | 12:18 |
| *** geaaru has quit IRC | 12:23 | |
| *** Shashikant86 has quit IRC | 12:24 | |
| cbartz | cschwede: Thank you. Since it is possible to know all middleware *after*, it is obvious if a required middleware (like copy) comes before or not. A question to the technical side: To retrieve the middleware after, you would use the instance variable most middlewares I have seen name *app*? All middleware shipped by swift use this name, right? | 12:24 |
| *** Shashikant86 has joined #openstack-swift | 12:33 | |
| *** geaaru has joined #openstack-swift | 12:36 | |
| *** openstackgerrit has joined #openstack-swift | 12:40 | |
| openstackgerrit | Christian Schwede proposed openstack/swift master: Add support to increase object ring partition power https://review.openstack.org/337297 | 12:40 |
| *** vint_bra has joined #openstack-swift | 12:44 | |
| *** ppai has quit IRC | 12:48 | |
| *** mvk has quit IRC | 12:55 | |
| *** ppai has joined #openstack-swift | 13:05 | |
| cschwede | cbartz: yes, you'd use the app parameter. I think all included middlewares use the name "app", but I didn't checked that - though I vague remember that this is a WSGI requirement | 13:23 |
| cschwede | cbartz: well, it seems all middlewares use "app": http://paste.openstack.org/show/597353/ | 13:25 |
| *** klamath has joined #openstack-swift | 13:28 | |
| *** klamath has quit IRC | 13:28 | |
| *** klamath has joined #openstack-swift | 13:29 | |
| *** mvk has joined #openstack-swift | 13:29 | |
| cbartz | cschwede: Thank you very much! | 13:32 |
| cschwede | cbartz: you're welcome :) | 13:33 |
| *** ma91 has left #openstack-swift | 13:36 | |
| *** balu_23 has joined #openstack-swift | 13:40 | |
| *** silor has joined #openstack-swift | 13:42 | |
| openstackgerrit | Christian Schwede proposed openstack/swift master: Add support to increase object ring partition power https://review.openstack.org/337297 | 13:47 |
| *** silor1 has joined #openstack-swift | 13:50 | |
| *** silor has quit IRC | 13:52 | |
| *** silor1 is now known as silor | 13:52 | |
| *** psachin has quit IRC | 13:55 | |
| *** dmorita has joined #openstack-swift | 14:18 | |
| *** dmorita has quit IRC | 14:18 | |
| *** dmorita has joined #openstack-swift | 14:19 | |
| *** bkopilov_ has joined #openstack-swift | 14:19 | |
| *** bkopilov has joined #openstack-swift | 14:19 | |
| *** Jeffrey4l has quit IRC | 14:24 | |
| *** psachin has joined #openstack-swift | 14:26 | |
| openstackgerrit | Gábor Antal proposed openstack/swift master: Handle log message interpolation by the logger https://review.openstack.org/428159 | 14:27 |
| *** zaitcev has joined #openstack-swift | 14:30 | |
| *** ChanServ sets mode: +v zaitcev | 14:30 | |
| *** Jeffrey4l has joined #openstack-swift | 14:36 | |
| *** psachin has quit IRC | 14:37 | |
| *** Jeffrey4l has quit IRC | 14:42 | |
| *** Jeffrey4l has joined #openstack-swift | 14:44 | |
| *** ppai has quit IRC | 14:49 | |
| *** tonanhngo has joined #openstack-swift | 14:49 | |
| *** balu_23 has quit IRC | 14:54 | |
| *** _JZ_ has joined #openstack-swift | 15:01 | |
| *** sams-gleb has quit IRC | 15:04 | |
| *** sams-gleb has joined #openstack-swift | 15:05 | |
| *** dmorita has quit IRC | 15:08 | |
| *** balu_23 has joined #openstack-swift | 15:08 | |
| *** balu_23 has quit IRC | 15:08 | |
| *** sams-gleb has quit IRC | 15:09 | |
| *** dmorita has joined #openstack-swift | 15:16 | |
| *** dmorita has quit IRC | 15:22 | |
| *** sams-gleb has joined #openstack-swift | 15:23 | |
| *** gabor_antal_ has joined #openstack-swift | 15:32 | |
| *** gabor_antal has quit IRC | 15:32 | |
| *** vinsh has joined #openstack-swift | 16:00 | |
| notmyname | hello | 16:05 |
| *** oshritf has joined #openstack-swift | 16:27 | |
| *** vinsh has quit IRC | 16:28 | |
| *** rcernin has quit IRC | 16:29 | |
| *** tesseract has quit IRC | 16:41 | |
| *** oshritf has quit IRC | 16:41 | |
| *** klrmn has joined #openstack-swift | 16:43 | |
| *** chsc has joined #openstack-swift | 16:47 | |
| *** chsc has joined #openstack-swift | 16:47 | |
| *** vinsh has joined #openstack-swift | 16:51 | |
| *** McMurlock1 has quit IRC | 16:54 | |
| *** JimCheung has joined #openstack-swift | 16:54 | |
| *** cbartz has quit IRC | 16:59 | |
| jrichli | Tahvok: when I grep for memcached_server over the repo, I only see it in the install guide, but not in the code being used. I see that cache = swift.cache is being used, though. | 16:59 |
| jrichli | so I'd go with the http://docs.openstack.org/developer/swift/overview_auth.html config example wrt cache | 17:00 |
| *** hoonetorg has quit IRC | 17:03 | |
| *** hoonetorg has joined #openstack-swift | 17:04 | |
| Tahvok | From what I see under bin/swift-proxy-server | 17:11 |
| Tahvok | It doesn't care of the options, and passes them to 'proxy-server' - what is that proxy-server I'm trying to understand now | 17:11 |
| *** tonanhngo has quit IRC | 17:13 | |
| jrichli | Tahvok: the proxy-server is the process on your proxy node. What type of cluster are you running / building? | 17:17 |
| *** JimCheung has quit IRC | 17:18 | |
| jrichli | Tahvok: fwiw, in authtoken, using ids are preferred over using names because they are unique. | 17:19 |
| *** JimCheung has joined #openstack-swift | 17:19 | |
| * jrichli is off to lunch | 17:21 | |
| *** JimCheung has quit IRC | 17:21 | |
| Tahvok | I'll continue to investigate later.. | 17:22 |
| *** dmorita has joined #openstack-swift | 17:22 | |
| Tahvok | Hopefully some swift dev would be able to help me to solve the mistery | 17:22 |
| notmyname | Tahvok: I didn't see what the original mystery is that you're trying to solve | 17:23 |
| notmyname | what's up? | 17:23 |
| Tahvok | 'filter:authtoken' section at the bottom here: http://docs.openstack.org/mitaka/install-guide-ubuntu/swift-controller-install.html | 17:25 |
| Tahvok | It says to configure 'memcached_servers', 'auth_type', 'project_domain_name' and 'user_domain_name'. However memcached config does not exist, and the others have a different name: auth_plugin, project_domain_id and user_domain_id | 17:25 |
| Tahvok | I checked, and it's the same issue for newton: http://docs.openstack.org/project-install-guide/object-storage/draft/controller-install-ubuntu.html | 17:25 |
| notmyname | Tahvok: the memcache settings are in a different config section. see the 'filter:cache' section instead of 'filter:authoken' | 17:26 |
| *** Shashikant86 has quit IRC | 17:27 | |
| Tahvok | notmyname: you're speaking of memcache_servers setting. I'm speaking of memcached_servers settings as per the guide: http://docs.openstack.org/mitaka/install-guide-ubuntu/swift-controller-install.html | 17:27 |
| Tahvok | As I said in pm to jrichli, even the developer docs contradict each other | 17:27 |
| *** dmorita has quit IRC | 17:27 | |
| Tahvok | Here it says 'user_domain_id': http://docs.openstack.org/developer/swift/overview_auth.html | 17:27 |
| Tahvok | And here 'user_domain_name': http://docs.openstack.org/developer/keystonemiddleware/middlewarearchitecture.html#configuration | 17:28 |
| notmyname | Tahvok: we've got some well-commented sample config files that may help https://github.com/openstack/swift/blob/master/etc/proxy-server.conf-sample | 17:31 |
| Tahvok | And I'm speaking of this well-commented file. | 17:32 |
| Tahvok | It has 1 missing config, and 3 different than what is written on the official install guide. | 17:33 |
| *** JimCheung has joined #openstack-swift | 17:33 | |
| notmyname | Tahvok: AFAIK, both id and name will work. the authtoken config values will be passed to the keystone code, so when there's disagreement, I'd prefer the one written by keystone people | 17:33 |
| notmyname | the ones that are in the swift docs were tested to work, and I believe there's a comment to that effect in the sample config file | 17:34 |
| Tahvok | notmyname: nothing in the sample file, or I wouldn't have asked the question here. Either way, I understand that it should work, but maybe update the sample file to be similar to the official doc? | 17:35 |
| notmyname | of the different things you're looking at, there is not any one that is more official than any other | 17:36 |
| notmyname | it's just words written by different groups of people | 17:36 |
| notmyname | specifically, I'm referring to https://github.com/openstack/swift/blob/master/etc/proxy-server.conf-sample#L320 | 17:36 |
| notmyname | where it says these were tested to work, but defers to keystone's docs for any conflicts | 17:37 |
| Tahvok | I see | 17:37 |
| *** hoonetorg has quit IRC | 17:38 | |
| *** jordanP has quit IRC | 17:38 | |
| timburke | good morning | 17:39 |
| *** chlong has joined #openstack-swift | 17:39 | |
| *** hoonetorg has joined #openstack-swift | 17:40 | |
| *** dmorita has joined #openstack-swift | 17:46 | |
| *** hoonetorg has quit IRC | 17:47 | |
| *** dmorita has quit IRC | 17:48 | |
| *** dmorita has joined #openstack-swift | 17:48 | |
| *** hoonetorg has joined #openstack-swift | 17:58 | |
| *** McMurlock1 has joined #openstack-swift | 18:00 | |
| *** tonanhngo has joined #openstack-swift | 18:03 | |
| *** rledisez has quit IRC | 18:06 | |
| *** tqtran has joined #openstack-swift | 18:19 | |
| *** JimCheung has quit IRC | 18:22 | |
| *** geaaru has quit IRC | 18:22 | |
| *** JimCheung has joined #openstack-swift | 18:23 | |
| *** JimCheung has quit IRC | 18:25 | |
| *** mvk has quit IRC | 18:35 | |
| notmyname | anyone else following the "openstack projects using more memory now" ML thread? | 18:38 |
| notmyname | i wonder if anyone is/has tracked swift's memory usage in prod and seen significant changes | 18:38 |
| si1ver | I have graphs from before and after a liberty to mitaka upgrade. | 18:47 |
| si1ver | Or are we talking about newton? | 18:47 |
| notmyname | mitaka->newton->ocata is shown on http://lists.openstack.org/pipermail/openstack-dev/2017-February/111413.html | 18:49 |
| notmyname | but I'm curious about what your graphs may show | 18:49 |
| si1ver | Ah so my numbers won't help. I need to get some ubuntu 16.04 machines built before I can upgrade further. | 18:49 |
| notmyname | no, I'm not as much interested in the context of that particular ML thread. it just raised the question, in general, for me | 18:50 |
| notmyname | what's the blocker by not having 16.04? | 18:51 |
| si1ver | I believe the ubuntu-cloud repos for 14.04 only go up to mitaka, and we are tracking with those. | 18:51 |
| si1ver | https://wiki.ubuntu.com/OpenStack/CloudArchive | 18:52 |
| notmyname | oh I see. because you're using canonical's packaging for it | 18:53 |
| si1ver | Yes that's right. | 18:53 |
| si1ver | Mostly that is due to the maturity of the ubuntu openstack install guides on openstack.org. | 18:53 |
| si1ver | I do see our average memory use go up by 2GB or so around the mitake upgrade. | 18:57 |
| notmyname | same number of workers? same number of storage policies or nodes in a ring? | 19:01 |
| notmyname | I mean, you didn't grow your capacity by 10x or something | 19:01 |
| si1ver | I'm grabbing a graph screenshot for you from when I deployed this commit. | 19:02 |
| si1ver | http://i.imgur.com/nmK7n3s.png | 19:08 |
| si1ver | Not a clear smoking gun there, but upgrading to mitaka was the only change around that time. | 19:09 |
| notmyname | yeah, interesting | 19:12 |
| clayg | cschwede: you didn't get a chance to look at the get hashes! | 19:12 |
| clayg | :\ | 19:12 |
| clayg | for *me* fixing this mess in get_hashes is a higher priority some of the other patches on the list - and not because I worked on it - I worked on it - *BECAUSE* fixing bugs which are increasing our durability risk and optimizing rebalance are so important! | 19:15 |
| * clayg steps off soapbox | 19:15 | |
| *** McMurlock1 has quit IRC | 19:28 | |
| timburke | clayg: you never put it on https://wiki.openstack.org/wiki/Swift/PriorityReviews :-( | 19:29 |
| *** eranrom has joined #openstack-swift | 19:29 | |
| clayg | timburke: I guess? we've been talking about it in every weekly meeting for three weeks? | 19:32 |
| clayg | I don't normally add things to that page - but that's probably mostly just notmyname picking up my slack | 19:32 |
| timburke | i know. and it seems like there's already a lot of support behind it. but this week, that's been my list of things to run down | 19:32 |
| timburke | (that and i hate trying to spot concurrency issues) | 19:33 |
| clayg | timburke: it's ok - if I had to guess; I would think cschwede just got excited about your excellent review of part-power and wanted to get you a fresh rev and ran out of time. | 19:34 |
| clayg | it's NBD - either we'll merge it today or this evening or I'll package a patched version of the release so we can start QA | 19:34 |
| timburke | it's part of why i included a patch with "here's everything we need for me to +2" :-) | 19:35 |
| clayg | yeah reviewing that patch is mess | 19:35 |
| clayg | the *only* think it has going for it is that it's been through a *bunch* of iterations | 19:35 |
| clayg | it's unlikely in my opinion someone coming to it fresh is going to find something *new* that hasn't been considered or debated to some extent - it's mostly just catching up to snuff and understanding the impact of the change - it's shouldn't be like acoles_ clayg mahatic and Pavel are *only* people that know how suffix hashing works anymore | 19:37 |
| *** dmorita has quit IRC | 19:37 | |
| *** dmorita has joined #openstack-swift | 19:38 | |
| clayg | timburke: although it does look cschwede got part-power-increase square for you! that's *huge* kudos! | 19:40 |
| *** dmorita has quit IRC | 19:49 | |
| openstackgerrit | Merged openstack/swift master: Handle log message interpolation by the logger https://review.openstack.org/428159 | 19:54 |
| notmyname | for your consideration, here is the draft logo from the foundation for swift http://d.not.mn/swift_draft_logo.png | 19:58 |
| notmyname | (I'm told the hairline white triangles aren't there in the final) | 19:58 |
| *** dmorita has joined #openstack-swift | 19:59 | |
| notmyname | please leave comments in here or pm me or email me if you have issues with it | 19:59 |
| notmyname | (and with that, i need to step away briefly) | 19:59 |
| *** ChubYann has joined #openstack-swift | 20:00 | |
| *** tqtran has quit IRC | 20:05 | |
| *** dmorita has quit IRC | 20:05 | |
| *** dmorita has joined #openstack-swift | 20:09 | |
| *** lespaul has joined #openstack-swift | 20:14 | |
| *** lespaul has quit IRC | 20:16 | |
| openstackgerrit | Clay Gerrard proposed openstack/swift master: Optimize reconstructor handoffs_first https://review.openstack.org/428408 | 20:30 |
| zaitcev | Aww Tim, you cannot do line veto :-) +2 is +2 | 20:31 |
| openstackgerrit | Clay Gerrard proposed openstack/swift master: Optimize reconstructor handoffs_first https://review.openstack.org/428408 | 20:33 |
| clayg | line veto sounds awesome - I remember we used to talk about that alot when we did specs "find the parts for which there's strong agreement and merge those!" | 20:34 |
| clayg | I wish legislation followed the same practice - good engineering encourages the avoidance of riders - stick to what's minimally nessecary - then expand/enhance in followups - refactor before building out | 20:35 |
| *** joeljwright has joined #openstack-swift | 20:35 | |
| *** ChanServ sets mode: +v joeljwright | 20:35 | |
| clayg | it's hard tho :\ | 20:36 |
| clayg | i think it's literally more work (for the author) to land a bunch of small patches than one big one that just fixes a bunch of stuff | 20:36 |
| *** mvk has joined #openstack-swift | 20:38 | |
| timburke | zaitcev: it's a *good* patch! +2 is +2! just throwing out that there may be some further improvements to be made, but i'm happy to see it as future work | 20:40 |
| timburke | *if* we want to address them now, so much the better, i'll +2 that patchset too! +2 all the things! | 20:41 |
| timburke | notmyname: still multiple colors? i'm not really opposed to blue instead of brown, but the white underside is closer to a barn swallow | 20:45 |
| *** joeljwright has quit IRC | 20:50 | |
| clayg | kota_: look forward to your thoughts on patch 428408 and response on patch 425493 - fixing the broken handoffs_first mode in the reconstructor is a high priority for me | 20:51 |
| patchbot | https://review.openstack.org/#/c/428408/ - swift - Optimize reconstructor handoffs_first | 20:51 |
| patchbot | https://review.openstack.org/#/c/425493/ - swift - Make the reconstructor handoffs_first work (and us... | 20:51 |
| * clayg is back to global-ec | 20:51 | |
| clayg | timburke: I think we should change the logo to like a raven - something to scare everyone away | 20:52 |
| notmyname | timburke: they did a much better job on this one. first version was clearly a swallow. this one is more like a swift (but kinda like a house martin) http://www.bbc.co.uk/nature/22527420 | 20:52 |
| notmyname | timburke: however, I'm not sure that we'll get orthothographical purity from the foundation logo | 20:53 |
| clayg | we're all learning way too much about birds in this process | 20:54 |
| timburke | would it really be so hard to just do something like http://imgur.com/a/62naP ? | 20:59 |
| notmyname | timburke: nope. I don't imagine that would be hard at all | 21:00 |
| clayg | omg - now we're bird *and* photoshop experts!? | 21:01 |
| timburke | notmyname: i suppose the other question is, would you like it more? | 21:01 |
| notmyname | timburke: I might | 21:01 |
| notmyname | clayg: you can call us "full-stack" open source contributors | 21:01 |
| timburke | clayg: if i were still using windows much at all, i would've done that in paint | 21:01 |
| clayg | timburke: i've run out of ^&*#s | 21:01 |
| clayg | timburke: notmyname's sister does the official Swift logos - OpenStack can do whatever they want and call it whatever they want | 21:02 |
| notmyname | clayg is like, "uhhh, I don't understand. this bird-like picture doesn't help me rebalance my rings *at all*" | 21:02 |
| timburke | clayg: i feel like that general attitude has led to some of our recent tensions... | 21:02 |
| timburke | clayg: if we don't kinda *try* to help, it's all going to devolve until we get something like https://commons.wikimedia.org/wiki/File:Chimera_Apulia_Louvre_K362.jpg | 21:04 |
| notmyname | I think the marketing team at the foundation had an idea to make logos for everyone, thought it would be a fun little thing, and completely didn't realize either the level of effort or the reactions they'd get from devs | 21:07 |
| clayg | rofl | 21:08 |
| *** vint_bra has quit IRC | 21:08 | |
| clayg | you two are both *awesome* - it really does take a village - keep up the good fight! | 21:08 |
| notmyname | I still have no idea how "let's make a logo for everyone" squares with "we don't promote individual project brands" | 21:08 |
| *** NM has quit IRC | 21:09 | |
| notmyname | eg ttx's email here http://lists.openstack.org/pipermail/openstack-dev/2017-January/109936.html | 21:09 |
| *** catintheroof has quit IRC | 21:14 | |
| clayg | notmyname: mattoliverau: I need the review for container sharding - anyone have it starred? | 21:14 |
| clayg | patch 423906 | 21:15 |
| patchbot | https://review.openstack.org/#/c/423906/ - swift - Add container sharding to Swift containers | 21:15 |
| *** catintheroof has joined #openstack-swift | 21:15 | |
| clayg | sorry, turns out I *do* know how to use gerrit search! | 21:15 |
| notmyname | was nearly there :-) | 21:15 |
| *** catintheroof has quit IRC | 21:19 | |
| notmyname | just as something to keep in the back of your minds... | 21:19 |
| notmyname | there's been a bunch of complaining in openstack about using eventlet. (just saw it again on the mailing list) | 21:20 |
| timburke | i've noticed that too... | 21:20 |
| notmyname | I wouldn't be surprised if there is a larger effort to remove eventlet or something | 21:20 |
| clayg | yeah transparent cooperative concurrency in general can be pretty annoying :\ | 21:21 |
| clayg | what's the thread? what's the alternative suggested? | 21:21 |
| *** silor has quit IRC | 21:21 | |
| notmyname | there's the pike goal of "everything as a wsgi app" by which they mean "run it under apache/mod_wsgi" | 21:22 |
| timburke | clayg: http://lists.openstack.org/pipermail/openstack-dev/2017-February/111516.html | 21:22 |
| notmyname | clayg: no. the two things I've seen recently are "every time eventlet upgrades, everything breaks" and in the current oom-killer thread talking about "use a 'real' web server instead of eventlet" | 21:22 |
| clayg | seems orthogonal? Maybe their conflating their gripes? I used to do that - anytime anything with within two modules of setup.py/disutils/pkg_resources caused me problems I blamed pbr | 21:22 |
| clayg | how come horizon get's to be so small! | 21:24 |
| notmyname | yeah, maybe. but it doesn't matter? my point is that eventlet keeps getting brought up as a punching bag | 21:24 |
| clayg | yeah let's punch it! | 21:24 |
| clayg | i got no love for eventlet - it's like any good drug - it feels fine at first but the more you use it the more you hate yourself :'( | 21:25 |
| clayg | ... doesn't mean i'm going to stop | 21:25 |
| clayg | I NEED my coffee | 21:25 |
| notmyname | have you used *any* libray in any language in your career that you still love after using it for years? | 21:25 |
| notmyname | besides swift, of course ;-) | 21:25 |
| timburke | notmyname: as a library? i think most people here would agree, swiftclient *still* sucks :P | 21:26 |
| notmyname | lol | 21:26 |
| notmyname | programming: everything's terrible, but perhaps we can make it slightly less terrible | 21:26 |
| timburke | notmyname: ...but then clayg tells you, "no, now it *has* to be terrible, because clients *expect* it to be terrible" | 21:28 |
| tdasilva | rofl | 21:31 |
| tdasilva | ouch | 21:31 |
| *** Jeffrey4l has quit IRC | 21:35 | |
| *** Jeffrey4l has joined #openstack-swift | 21:38 | |
| jrichli | does anyone have a feel for the prevalence of older swift versions? as in, you can be fairly confident that people are running with at least version x? | 21:57 |
| si1ver | I see a lot of icehouse running around, and I have a havana cluster up and in prod. | 21:58 |
| jrichli | si1ver: thank you for the feedback. | 21:58 |
| si1ver | Looks like ubuntu still supports 12.04, and that shipped with essex. FWIW. | 21:59 |
| jrichli | ok | 21:59 |
| openstackgerrit | Clay Gerrard proposed openstack/swift master: Optimize reconstructor handoffs_first https://review.openstack.org/428408 | 21:59 |
| clayg | ^ oops! one of those days | 21:59 |
| clayg | jrichli: IME people can be very slow to upgrade - we have some people running basically every version of swift we've ever released - which is ~1.7 - although recently we've made some headway bring people up to ~1.13 which was our last release before storage policies | 22:02 |
| timburke | yowza! essex would put it back at 1.4.8 -- havana was 1.10.0, icehouse 1.13.1 | 22:02 |
| clayg | I don't remember exactly how old that is or what OS release it ties back to | 22:02 |
| clayg | was icehouse a long time ago? | 22:02 |
| clayg | Apr 2014? seems like awhile ago :'( | 22:03 |
| jrichli | clayg: ok, thanks for feedback | 22:04 |
| clayg | I think storage infrastructure folks are used to upgrades every 3-5 years and don't normally think much about software updates to the storage that was a box - how often do you upgrade your refridgerator? you don't - you stick stuff in it - eventually you buy one that makes ice | 22:04 |
| clayg | it's a brave new world | 22:05 |
| clayg | jrichli: fwiw lots of people upgrade everytime we release! lots of other folks upgrade to get some feature and then go back to the old anit-upgrade party. I'm not sure how representative swiftstack deployments are in terms of upgrade patterns for the broader devops oriented teams. | 22:06 |
| clayg | if the question is about what features should apps code to - it depends on if they need to support private or public installations - and I think our customers would upgrade to get a feature that was required by an application as long as it used info correct and told them they need a newer swift? | 22:07 |
| clayg | "dood i'm the bomb app, you know you want me - your swift is 3 years old, upgrade already" | 22:08 |
| jrichli | clayg: the question came from a group that I didn't even know about (darn IBM is so big!). but I know the question was more about would it be ok to possible drop support for some older versions | 22:10 |
| clayg | drop support in... a client or like deployment tooling? No one goes out of their way to *deploy* old versions. If it's just management tools as long as they say "you need to upgrade your swift before you upgrade to management suite XYZ" that should be easy enough to deprecate support and then eventually do a checkpoint release? | 22:11 |
| *** tqtran has joined #openstack-swift | 22:11 | |
| clayg | or just wing it and see what happens | 22:11 |
| jrichli | right, makes sense. i really dont know about what this group is providing at all. I am going to get more info. but this helps me respond with something. thanks! | 22:13 |
| clayg | sorry - I'll shut up now | 22:13 |
| *** dmorita has quit IRC | 22:24 | |
| *** dmorita has joined #openstack-swift | 22:25 | |
| *** dmorita has quit IRC | 22:25 | |
| *** dmorita has joined #openstack-swift | 22:25 | |
| *** NM has joined #openstack-swift | 22:28 | |
| *** NM has quit IRC | 22:35 | |
| jrichli | mattoliverau, hurricanerix, dfg_, redbo_ : what version of swift is rackspace running with? (clayg - you happen to know?) | 22:44 |
| *** sams-gleb has quit IRC | 22:46 | |
| *** chsc has quit IRC | 22:48 | |
| jrichli | clayg: turns out this system acts as a client for possibly any swift cluster / probably the more commonly used like public clusters | 22:52 |
| -openstackstatus- NOTICE: Restarting gerrit due to performance problems | 22:52 | |
| *** NM has joined #openstack-swift | 22:56 | |
| *** NM has quit IRC | 22:57 | |
| *** klamath has quit IRC | 23:22 | |
| openstackgerrit | Clay Gerrard proposed openstack/swift master: Optimize reconstructor handoffs_first https://review.openstack.org/428408 | 23:28 |
| notmyname | jrichli: according to /info RAX is running 2.2.0 | 23:28 |
| *** tonanhngo has quit IRC | 23:29 | |
| clayg | hey that's pretty good! | 23:29 |
| notmyname | that was released dec 2014 | 23:29 |
| jrichli | notmyname: oh, thanks! | 23:29 |
| *** kei_yama has joined #openstack-swift | 23:29 | |
| notmyname | personally I don't like it because it doesn't support container tempurls, and I wanted to use that | 23:29 |
| *** tonanhngo has joined #openstack-swift | 23:34 | |
| clayg | i swear I've checked out patch 219165 like 3 times - if tests fail again on patch 419787 i give up - computers are too hard - i'm going home | 23:38 |
| patchbot | https://review.openstack.org/#/c/219165/ - swift - EC Fragment Duplication - Foundational Global EC C... | 23:38 |
| patchbot | https://review.openstack.org/#/c/419787/ - swift - Better optimistic lock in get_hashes | 23:38 |
| clayg | oh.. no not that one - better get hashes is awesome - stupid this patch 428408 | 23:38 |
| patchbot | https://review.openstack.org/#/c/428408/ - swift - Optimize reconstructor handoffs_first | 23:38 |
| *** tonanhngo has quit IRC | 23:39 | |
| * clayg cries - thanks timburke I agree the stub part_info dicts should have representative keys if possible - good call | 23:40 | |
| timburke | clayg: just further proof that there can be no greater time suck than writing a patch whose commit message starts "optimize..." :-( | 23:42 |
| timburke | it's just, like, destined to take 5x longer than you thought it would | 23:42 |
| clayg | timburke: rofl!! | 23:44 |
| *** sams-gleb has joined #openstack-swift | 23:47 | |
| openstackgerrit | Clay Gerrard proposed openstack/swift master: Optimize reconstructor handoffs_first https://review.openstack.org/428408 | 23:49 |
| *** sams-gleb has quit IRC | 23:52 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
