| *** pberis has joined #openstack-swift | 00:00 | |
| *** haomaiw__ has quit IRC | 00:04 | |
| *** wer_ has joined #openstack-swift | 00:04 | |
| *** haomaiwang has joined #openstack-swift | 00:05 | |
| *** thurloat_isgone is now known as thurloat | 00:08 | |
| *** pberis has quit IRC | 00:11 | |
| *** delatte has quit IRC | 00:15 | |
| *** delattec has quit IRC | 00:15 | |
| *** doxavore has quit IRC | 00:16 | |
| torgomatic | notmyname: so it looks like oslo.policy has a hard dependency on oslo.config; not so much for configuration, but just in terms of import lines | 00:17 |
|---|---|---|
| torgomatic | however, based on a quick reading of the code, it seems like there's just some missing plumbing for the policy path in Enforcer.__init__ | 00:18 |
| *** gyee has quit IRC | 00:23 | |
| *** kota_ has joined #openstack-swift | 00:26 | |
| *** ChanServ sets mode: +v kota_ | 00:26 | |
| *** haomaiwang has quit IRC | 00:29 | |
| *** MooingLemur has quit IRC | 00:30 | |
| *** zhill has quit IRC | 00:31 | |
| *** thurloat is now known as thurloat_isgone | 00:31 | |
| *** dmorita_ has joined #openstack-swift | 00:34 | |
| *** jamielennox|away is now known as jamielennox | 00:35 | |
| *** jasondotstar has quit IRC | 00:39 | |
| *** cdelatte has joined #openstack-swift | 00:39 | |
| *** delattec has joined #openstack-swift | 00:39 | |
| tdasilva | redbo: re current expiring object container getting hammered. doesn't that patch to shard containers help with that? | 00:46 |
| *** cdelatte has quit IRC | 00:49 | |
| *** delattec has quit IRC | 00:50 | |
| *** cdelatte has joined #openstack-swift | 00:57 | |
| *** delattec has joined #openstack-swift | 00:57 | |
| *** jamielennox is now known as jamielennox|away | 01:00 | |
| *** lastops has joined #openstack-swift | 01:05 | |
| *** jamielennox|away is now known as jamielennox | 01:08 | |
| *** haomaiw__ has joined #openstack-swift | 01:09 | |
| *** lastops has quit IRC | 01:10 | |
| *** pgbridge has joined #openstack-swift | 01:13 | |
| *** haomaiw__ has quit IRC | 01:15 | |
| *** janonymous__ has quit IRC | 01:18 | |
| *** proteusguy has joined #openstack-swift | 01:21 | |
| *** Kennan has quit IRC | 01:22 | |
| *** Kennan has joined #openstack-swift | 01:23 | |
| *** jrichli has joined #openstack-swift | 01:54 | |
| redbo | I guess you could try to balance your expiring objects container interval against what size containers are sharded at and how fast a container can process DELETEs, but we could probably just redesign it to be less goofy and centralized. | 01:56 |
| redbo | like if a container can process a million deletes an hour and they het sharded at 5 million you have to set the interval to 20 minutes | 02:10 |
| redbo | s/het/get | 02:11 |
| *** kcaj has joined #openstack-swift | 02:13 | |
| *** asettle is now known as asettle-afk | 02:17 | |
| *** haomaiwang has joined #openstack-swift | 02:17 | |
| zaitcev | probably | 02:19 |
| *** jamielennox is now known as jamielennox|away | 02:20 | |
| *** zaitcev has quit IRC | 02:20 | |
| *** bkopilov has quit IRC | 02:25 | |
| *** jamielennox|away is now known as jamielennox | 03:03 | |
| *** pgbridge has quit IRC | 03:13 | |
| *** silor has joined #openstack-swift | 03:14 | |
| *** jamielennox is now known as jamielennox|away | 03:15 | |
| *** jamielennox|away is now known as jamielennox | 03:25 | |
| *** asettle-afk is now known as asettle | 03:30 | |
| jamielennox | does swiftclient support keystone v3 authentication? | 03:32 |
| jamielennox | and i mean using the Connection object - not the CLI | 03:32 |
| *** delattec has quit IRC | 03:37 | |
| *** cdelatte has quit IRC | 03:37 | |
| clayg | who was working on lp bug #1452553 | 03:37 |
| openstack | Launchpad bug 1452553 in OpenStack Object Storage (swift) "don't rebuild existing fragments" [Undecided,New] https://launchpad.net/bugs/1452553 - Assigned to Minwoo Bae (minwoob) | 03:37 |
| clayg | minwoo - what's his handle? | 03:37 |
| clayg | peluse: you're gunna love this - i was telling him he should look at the patch I attached to the lp bug - and guess what - the patch had a bug in it | 03:39 |
| clayg | i'm like 0/2 | 03:40 |
| clayg | jamielennox: i think it does - but idk what all you have to fill into the os_options dict to make it work | 03:43 |
| jrichli | clayg: minwoob | 03:43 |
| clayg | jrichli: thanks - but he's not in chan :\ | 03:43 |
| clayg | anyway - if anyone seems minwoob - can you tell him that patch I gave him was crap and I updated the bug :P | 03:44 |
| jrichli | I can tell him tomorrow morning :-) | 03:44 |
| jrichli | he works with me | 03:44 |
| jamielennox | clayg: it seems (and i say seems because i'm still trying to figure out the path from shell -> connection) like passing an os_options dict might work, but that whole pattern would indicate it's really only used by the CLI | 03:44 |
| clayg | jamielennox: well the tricky bit from shell -> connection goes through service | 03:46 |
| jamielennox | clayg: yep, is that expected to be something that other services would use? | 03:46 |
| jamielennox | or a CLI only thing? | 03:46 |
| jamielennox | for context i need to look at how the swift backend for glance can be made to use v3 authentication | 03:48 |
| jamielennox | and at the moment i'm tempted to just use requests directly :) | 03:48 |
| clayg | jamielennox: but i'm not sure that's relevant unless you want a swiftservice instance - if you just want a connection it seems like os_options and auth_version should do the trick | 03:48 |
| clayg | hey using requests directly is a great way to do it! | 03:49 |
| clayg | the SwiftServices thing is great if you can deal with it's opionated mannor - it's going to spin up some worker threads and manaage a bunch of stuff like outputting on it's own | 03:49 |
| jamielennox | clayg: so i'm not on the swift or glance teams - is that a "swift official position" ? | 03:49 |
| *** bkopilov has joined #openstack-swift | 03:49 | |
| jamielennox | like i can do all this with requests and tell glance that this was recommended? | 03:50 |
| clayg | you can probably *override* the bits you want - but most of the time I think folks need something less opinonated - either the Connection class - or just roll your own over your favorite http library | 03:50 |
| jamielennox | yea, i don't think i want to be dealing with the multithreaded aspects of the SwiftService from another service | 03:51 |
| clayg | yeah so just use a connection - i mean it'll buy you the retry handling and stuff at least | 03:51 |
| clayg | but the keystone v3 stuff should be a non-issue - the connection class will just proxy your kwargs to keystoneclient | 03:52 |
| jamielennox | ok, just figure out the mix of where things need to go between regular params and os_options | 03:52 |
| clayg | and while using a simple lowlevel http lib to talk *to swift* is probably reasonable - I doubt you attempt something like that for keystone | 03:52 |
| jamielennox | clayg: well keystoneclient sessions buy me everything i need to talk to keystone already | 03:53 |
| clayg | seems to me it's all in swiftclient.client | 03:53 |
| jamielennox | which would leave me with pretty much an authenticated requests.Session i could use for swift | 03:53 |
| jamielennox | incl. some retry handling | 03:53 |
| clayg | jamielennox: so can you just use preauth token and preauth url? | 03:53 |
| clayg | or subclass connection and swap out get_auth | 03:54 |
| jamielennox | i could, it would mean i'd have to handle all the cacerts and stuff twice but that's not the end of the world | 03:55 |
| jamielennox | subclassing is not a bad idea, i'll need to look at that one | 03:55 |
| openstackgerrit | Kuo-tung Kao proposed openstack/swift: Use testtools for base test class https://review.openstack.org/192960 | 03:56 |
| jamielennox | clayg: ok, that's a few leads for me to go on, thanks | 03:56 |
| *** jrichli has quit IRC | 03:57 | |
| clayg | gl | 03:57 |
| clayg | acoles_away: the 'project' thing is crazy - it's acctually an *attribute* of CONF objects - which also happen to override __getattr__ for missing attributes | 04:03 |
| clayg | acoles_away: so if you don't initalize the conf object the attribute is missing and it will fall through to picking it up from the conf file - so crazy - total hack | 04:04 |
| *** ppai has joined #openstack-swift | 04:35 | |
| *** nadeem has joined #openstack-swift | 04:36 | |
| *** nadeem has quit IRC | 04:37 | |
| *** nadeem has joined #openstack-swift | 04:37 | |
| *** SkyRocknRoll has joined #openstack-swift | 04:55 | |
| *** SkyRocknRoll has quit IRC | 05:00 | |
| *** SkyRocknRoll has joined #openstack-swift | 05:01 | |
| *** Kennan has quit IRC | 05:09 | |
| *** Kennan has joined #openstack-swift | 05:10 | |
| *** kota_ has quit IRC | 05:23 | |
| *** jamielennox is now known as jamielennox|away | 05:23 | |
| *** jamielennox|away is now known as jamielennox | 05:26 | |
| *** mmcardle has joined #openstack-swift | 05:31 | |
| *** mmcardle has quit IRC | 05:37 | |
| *** chlong has quit IRC | 05:38 | |
| *** ianbrown has quit IRC | 05:39 | |
| openstackgerrit | Kuo-tung Kao proposed openstack/swift: Use testtools for base test class https://review.openstack.org/192960 | 05:44 |
| *** SkyRocknRoll has quit IRC | 05:46 | |
| *** SkyRocknRoll has joined #openstack-swift | 05:48 | |
| *** ronenkat has quit IRC | 05:49 | |
| *** chlong has joined #openstack-swift | 05:53 | |
| *** Kennan has quit IRC | 05:58 | |
| *** Kennan has joined #openstack-swift | 06:01 | |
| *** Kennan has quit IRC | 06:06 | |
| *** Kennan has joined #openstack-swift | 06:06 | |
| *** pgbridge has joined #openstack-swift | 06:08 | |
| *** nadeem has quit IRC | 06:39 | |
| *** ianbrown has joined #openstack-swift | 06:46 | |
| *** krykowski has joined #openstack-swift | 06:59 | |
| openstackgerrit | Kuo-tung Kao proposed openstack/swift: Use testtools for base test class https://review.openstack.org/192960 | 07:04 |
| *** ianbrown has quit IRC | 07:06 | |
| *** ppai has quit IRC | 07:09 | |
| *** ppai has joined #openstack-swift | 07:14 | |
| *** ianbrown has joined #openstack-swift | 07:15 | |
| *** slavisa has joined #openstack-swift | 07:18 | |
| *** chlong has quit IRC | 07:18 | |
| *** jordanP has joined #openstack-swift | 07:34 | |
| *** proteusguy has quit IRC | 07:44 | |
| *** acoles_away is now known as acoles | 07:50 | |
| acoles | good morning | 07:53 |
| *** joeljwright has joined #openstack-swift | 07:55 | |
| *** ChanServ sets mode: +v joeljwright | 07:55 | |
| *** ianbrown has quit IRC | 07:55 | |
| *** joeljwright has quit IRC | 07:57 | |
| *** geaaru has joined #openstack-swift | 07:58 | |
| acoles | clayg: huh. so sounds like project could be whatever you put in conf file if you fail to init the object. if i understand you. | 07:59 |
| *** joeljwright has joined #openstack-swift | 08:02 | |
| *** ChanServ sets mode: +v joeljwright | 08:02 | |
| *** joeljwright has quit IRC | 08:02 | |
| *** proteusguy has joined #openstack-swift | 08:06 | |
| ho | acoles: morning! | 08:25 |
| *** joeljwright has joined #openstack-swift | 08:37 | |
| *** ChanServ sets mode: +v joeljwright | 08:37 | |
| *** mmcardle has joined #openstack-swift | 08:39 | |
| acoles | ho: hello | 08:40 |
| cschwede | Good Morning! | 08:41 |
| acoles | joeljwright: i have a feeling that figuring out how to support keystone sessions is going to be significant for swiftclient going forwards http://lists.openstack.org/pipermail/openstack-dev/2015-June/067381.html | 08:41 |
| *** mmcardle has quit IRC | 08:42 | |
| ho | cschwede: morning! | 08:49 |
| *** mmcardle has joined #openstack-swift | 08:49 | |
| cschwede | ho: hello! | 08:50 |
| *** mmcardle has quit IRC | 09:12 | |
| *** mmcardle has joined #openstack-swift | 09:23 | |
| *** haomaiwang has quit IRC | 09:29 | |
| *** joeljwright has quit IRC | 09:34 | |
| *** slavisa has quit IRC | 09:34 | |
| *** aix has joined #openstack-swift | 09:37 | |
| *** joeljwright has joined #openstack-swift | 09:38 | |
| *** ChanServ sets mode: +v joeljwright | 09:38 | |
| *** jasondotstar has joined #openstack-swift | 09:52 | |
| *** mmcardle has quit IRC | 09:54 | |
| *** dmorita_ has quit IRC | 09:55 | |
| *** mmcardle has joined #openstack-swift | 10:00 | |
| joeljwright | acoles: Yeah, keystone sessions are on our list of things to looks at | 10:01 |
| joeljwright | acoles: thanks for the link, will take a read of that later today | 10:02 |
| acoles | joeljwright: i need to get up to speed on them. the historical impediment with swiftclient has been needing to support v1 auth and not wanting a hard dependency on keystoneclient | 10:03 |
| acoles | i had this hand-waving idea that maybe v1 auth could be encapsulated in a 'fake' session-like class so client then becomes session compatible | 10:04 |
| * acoles stops waving hands | 10:04 | |
| joeljwright | acoles: I spent a bit of time talking with Tim in Vancouver about something like a fake session object (it might have been triggered by a comment made by you) | 10:05 |
| acoles | joeljwright: did you trash the idea: :P | 10:05 |
| joeljwright | I'm crazy busy with a software release until the end of June | 10:05 |
| acoles | ? | 10:05 |
| joeljwright | but I want to look at that asap afterwards | 10:05 |
| acoles | me too. :) back to the day job... | 10:06 |
| joeljwright | yeah me too... | 10:06 |
| *** SkyRocknRoll has quit IRC | 10:08 | |
| *** aix has quit IRC | 10:21 | |
| openstackgerrit | Hisashi Osanai proposed openstack/swift-specs: Oslo config support in Swift https://review.openstack.org/192094 | 10:25 |
| *** haomaiwa_ has joined #openstack-swift | 10:30 | |
| *** mmcardle has quit IRC | 10:33 | |
| *** joeljwright has quit IRC | 10:33 | |
| *** mmcardle has joined #openstack-swift | 10:37 | |
| *** ho has quit IRC | 10:38 | |
| *** joeljwright has joined #openstack-swift | 10:39 | |
| *** ChanServ sets mode: +v joeljwright | 10:39 | |
| *** CR7 has joined #openstack-swift | 10:43 | |
| *** SkyRocknRoll has joined #openstack-swift | 10:57 | |
| *** SkyRocknRoll has joined #openstack-swift | 10:57 | |
| *** silor has quit IRC | 10:59 | |
| *** mmcardle has quit IRC | 11:06 | |
| *** joeljwright has quit IRC | 11:06 | |
| *** ronenkat has joined #openstack-swift | 11:17 | |
| *** aix has joined #openstack-swift | 11:32 | |
| *** bkopilov has quit IRC | 11:34 | |
| *** jasondotstar has quit IRC | 11:46 | |
| *** SkyRocknRoll has quit IRC | 11:53 | |
| *** mmcardle has joined #openstack-swift | 12:00 | |
| *** joeljwright has joined #openstack-swift | 12:00 | |
| *** ChanServ sets mode: +v joeljwright | 12:00 | |
| *** csmart has joined #openstack-swift | 12:01 | |
| *** SkyRocknRoll has joined #openstack-swift | 12:08 | |
| *** jasondotstar has joined #openstack-swift | 12:10 | |
| *** kei_yama has quit IRC | 12:13 | |
| *** km has quit IRC | 12:13 | |
| *** jordan__ has joined #openstack-swift | 12:18 | |
| *** silor has joined #openstack-swift | 12:19 | |
| *** thurloat_isgone is now known as thurloat | 12:21 | |
| *** jordanP has quit IRC | 12:22 | |
| *** SkyRocknRoll has quit IRC | 12:26 | |
| *** jordan__ has quit IRC | 12:31 | |
| *** jordanP has joined #openstack-swift | 12:31 | |
| *** jasondotstar has quit IRC | 12:42 | |
| *** SkyRocknRoll has joined #openstack-swift | 12:42 | |
| *** jasondotstar has joined #openstack-swift | 12:55 | |
| *** petertr7_away is now known as petertr7 | 12:58 | |
| *** mordred has joined #openstack-swift | 13:03 | |
| *** bkopilov has joined #openstack-swift | 13:07 | |
| *** ppai has quit IRC | 13:07 | |
| *** CR7 has quit IRC | 13:09 | |
| *** amoturi has joined #openstack-swift | 13:09 | |
| *** cdelatte has joined #openstack-swift | 13:16 | |
| *** jasondotstar has quit IRC | 13:18 | |
| *** delattec has joined #openstack-swift | 13:26 | |
| *** cdelatte has quit IRC | 13:29 | |
| *** fifieldt_ is now known as fifieldt | 13:36 | |
| *** SkyRocknRoll has quit IRC | 13:36 | |
| *** jkugel has joined #openstack-swift | 13:37 | |
| *** NM has joined #openstack-swift | 13:44 | |
| *** eranrom has joined #openstack-swift | 13:57 | |
| *** jlhinson has joined #openstack-swift | 13:57 | |
| *** pgbridge has quit IRC | 13:58 | |
| eranrom | Got a question on user Vs. system metadata: Take for example x-container-read, is that a sysmeta in nature or user metadata in nature. On one hand the documnetation states that ACLs are examples of core swift features that should have been sysmeta and on the other they are set by users | 13:59 |
| eranrom | and are not filtered out in responses | 13:59 |
| *** blmartin has joined #openstack-swift | 14:02 | |
| *** blmartin has quit IRC | 14:09 | |
| *** acampbel11 has joined #openstack-swift | 14:11 | |
| *** barker has joined #openstack-swift | 14:12 | |
| *** jrichli has joined #openstack-swift | 14:16 | |
| *** wbhuber has joined #openstack-swift | 14:18 | |
| *** SkyRocknRoll has joined #openstack-swift | 14:19 | |
| *** SkyRocknRoll has joined #openstack-swift | 14:19 | |
| *** pgbridge has joined #openstack-swift | 14:20 | |
| kragniz | joeljwright: yeah, I spent a while looking at that gate-tempest-dsvm-neutron-src-python-swiftclient failure a couple of days ago | 14:26 |
| kragniz | joeljwright: it's pretty odd | 14:26 |
| joeljwright | there's surely no reason for it | 14:28 |
| joeljwright | all the patch does is rename some non-kwargs | 14:28 |
| joeljwright | so it must be an infra problem | 14:29 |
| joeljwright | but all the other project dsvm neutron tests are ok | 14:29 |
| *** jasondot_ has joined #openstack-swift | 14:32 | |
| *** blmartin has joined #openstack-swift | 14:34 | |
| blmartin | Good morning all | 14:34 |
| *** MooingLemur has joined #openstack-swift | 14:34 | |
| kragniz | joeljwright: yeah, it would be worrying if that actually broke things | 14:36 |
| *** minwoob has joined #openstack-swift | 14:46 | |
| *** NM has quit IRC | 14:48 | |
| *** NM has joined #openstack-swift | 14:49 | |
| jrichli | blmartin: good morning | 14:50 |
| *** NM has quit IRC | 14:50 | |
| *** jasondot_ has quit IRC | 14:51 | |
| *** jasondotstar has joined #openstack-swift | 14:51 | |
| *** NM has joined #openstack-swift | 14:52 | |
| *** wbhuber has quit IRC | 14:52 | |
| *** zhill has joined #openstack-swift | 14:56 | |
| *** mmcardle1 has joined #openstack-swift | 14:56 | |
| *** mmcardle has quit IRC | 14:57 | |
| *** barker has quit IRC | 15:01 | |
| *** blmartin has quit IRC | 15:04 | |
| *** blmartin has joined #openstack-swift | 15:05 | |
| portante | so the swift guys are just working on a patch, if it has not been submitted already, https://review.openstack.org/184189, which up the number of workers per-disk to get higher levels of concurrenty | 15:07 |
| portante | oh, hey swift guys | 15:07 |
| portante | was err ... ah ... um ... just talking about you guys | 15:08 |
| * portante goes and finds the right channel now ... | 15:08 | |
| *** haomaiw__ has joined #openstack-swift | 15:08 | |
| swifterdarrell | portante: lol | 15:08 |
| *** haomaiwa_ has quit IRC | 15:09 | |
| *** acampbel11 has quit IRC | 15:14 | |
| *** barker has joined #openstack-swift | 15:16 | |
| *** wbhuber has joined #openstack-swift | 15:23 | |
| *** janonymous_ has joined #openstack-swift | 15:27 | |
| janonymous_ | Hi , this might seem lame but could anybody please give me an idea about : pile = GreenAsyncPile(len(conns)); for response in pile: ... | 15:29 |
| janonymous_ | This is a snippet of obj.py | 15:29 |
| janonymous_ | how does it work when no response is in the pile ? | 15:30 |
| *** NM has quit IRC | 15:31 | |
| *** gyee has joined #openstack-swift | 15:32 | |
| janonymous_ | say out of 5 response only 2 came , how does the loop goes ? could somebody help on this | 15:33 |
| *** Fin1te has joined #openstack-swift | 15:46 | |
| notmyname | good morning, world | 15:46 |
| notmyname | eranrom: as a first pass, sysmeta is x-object-sysmeta-*. and user meta is x-object-meta-* :-) | 15:47 |
| *** haomaiwang has joined #openstack-swift | 15:48 | |
| notmyname | joeljwright: any progress on the swiftclient gate issue from yesterday? | 15:48 |
| joeljwright | not really | 15:49 |
| joeljwright | on #openstack-qa discussing it now | 15:49 |
| *** haomaiw__ has quit IRC | 15:49 | |
| pgbridge | good morning | 15:50 |
| joeljwright | notmyname: potentially related to bug 1463200 but "that one's still lacking in any significant analysis. as i said there could be more than one cause" | 15:50 |
| openstack | bug 1463200 in OpenStack Compute (nova) "check-tempest-dsvm-multinode-full fails due to "Failed to compute_task_migrate_server: No valid host was found"" [Undecided,New] https://launchpad.net/bugs/1463200 | 15:50 |
| *** SpamapS has joined #openstack-swift | 15:52 | |
| SpamapS | HI! I'm looking at a possible bug in swiftclient.service where some methods take options['header']and some take optoins['headers'] and it seems the actual implementation is a) untested and b) inconsistent. Before I get too deep, is there anybody here paying attention that might be familiar with this and able to shed light on why the distinction may be important? Thanks. | 15:53 |
| SpamapS | s/optoins/options/ | 15:54 |
| notmyname | SpamapS: if joeljwright isn't available, I can help. (I just got online this morning and am going through emails, etc) | 15:55 |
| joeljwright | SpamapS: I'm here | 15:55 |
| joeljwright | SpamapS: If you point me at the locations I can have a look and tell you if there's anything important there (or a bug) | 15:56 |
| SpamapS | joeljwright: https://git.openstack.org/cgit/openstack/python-swiftclient/tree/swiftclient/service.py#n1680 | 15:57 |
| SpamapS | joeljwright: I believe that should be 'headers' | 15:57 |
| SpamapS | joeljwright: according to the docs for upload | 15:57 |
| SpamapS | joeljwright: but download takes 'header' | 15:57 |
| SpamapS | joeljwright: and we did just now find that 'headers' does not work on upload. | 15:58 |
| SpamapS | I'm working on a test that will break before I get too deep | 15:58 |
| joeljwright | I can't find anywhere that it calls options['headers'] | 15:59 |
| joeljwright | but I think I know why it's option['header'] | 15:59 |
| *** acampbell has joined #openstack-swift | 15:59 | |
| eranrom | notmyname: yeah I got that. thanks. Just wondering if we are not missing CDMI's equivalent of 'data system metadata' which is a system metadata that can be manipulated by the user instructing the system how it should work (e.g. ACLs) the current definition of sysmeta is more of a metadta that cannot be manipulated by the user, but rather for internal usage. Anyway, its not very important | 15:59 |
| joeljwright | the argparse option in the CLI is '—header' which produces a list of key:value pairs | 16:00 |
| SpamapS | joeljwright: https://git.openstack.org/cgit/openstack/python-swiftclient/tree/swiftclient/service.py#n1205 | 16:00 |
| SpamapS | joeljwright: docs say 'headers' | 16:00 |
| *** amoturi has quit IRC | 16:00 | |
| joeljwright | ah, in that case it's a doc bug | 16:00 |
| SpamapS | joeljwright: let me try 'header' | 16:01 |
| joeljwright | probably because I wrote that and subliminally agree with you :) | 16:01 |
| *** jasondotstar is now known as jasondotstar|afk | 16:01 | |
| joeljwright | all the code usages are options['header'], so it's certainly a bug in the docstring | 16:01 |
| notmyname | eranrom: yeah, I'd say there's almost 3 classes of metadata: sysmeta that's hidden and set by the system (eg encryption or EC stuff); user meta where the user can do whatever (x-object-meta-*); and then "other" like content-type, etag, ACLs, etc that are more like first-class HTTP headers (or otherwise don't follow another naming pattern) | 16:02 |
| notmyname | eranrom: but the "other" category can sometimes be set by the user directly and sometimes indirectly. eg content-type is directly set. timestamp and etag are indirectly set | 16:02 |
| eranrom | noymyname: Agree, I guess this makes it 4 classes :-) | 16:03 |
| SpamapS | joeljwright: cool, I'm testing using that, but I wanted to make sure I wasn't crazy. Thanks for confirming. | 16:03 |
| SpamapS | joeljwright: should I push up a patch? | 16:03 |
| joeljwright | SpamapS: yes please, a patch would be great | 16:04 |
| *** jordanP has quit IRC | 16:04 | |
| notmyname | eranrom: your 4th class is stuff liek the acls? | 16:04 |
| openstackgerrit | Clint 'SpamapS' Byrum proposed openstack/python-swiftclient: Fix docstring typo for SwiftService.upload https://review.openstack.org/193187 | 16:05 |
| eranrom | notmyname: exactly. 3rd is like etag, 4th ACLs | 16:05 |
| joeljwright | SpamapS: thanks | 16:05 |
| notmyname | eranrom: with your container sync stuff, you're building a whitelist of headers to sync, right? | 16:06 |
| *** haomaiw__ has joined #openstack-swift | 16:06 | |
| notmyname | eranrom: is that set per container or in a config somewhere? | 16:06 |
| SpamapS | joeljwright: Néni zač. :) | 16:07 |
| *** haomaiwang has quit IRC | 16:07 | |
| notmyname | SpamapS: joeljwright: approved | 16:07 |
| eranrom | notmyname: I started off thinking it should be in a config, but now leaning towards a per container definition, which brings me to that 4th class... | 16:07 |
| notmyname | eranrom: in order to filter the configured headers? | 16:08 |
| eranrom | notmyname: yep | 16:08 |
| notmyname | eranrom: I'd definitely want to see starting small for now and leaving out ACLs as something to sync | 16:09 |
| eranrom | notmyname: that is: the metadata that holds "whitlist" per container | 16:09 |
| notmyname | eranrom: right. if you let users set what to sync, then you have to filter that based on what is allowed to sync | 16:09 |
| joeljwright | notmyname: fastest swiftclient patch ever! | 16:09 |
| notmyname | eranrom: ie set(allowed headers) | set(requested headers) => set(headers that are synced) | 16:10 |
| notmyname | eranrom: or is that &? point is, the intersection of those sets :-) | 16:10 |
| *** petertr7 is now known as petertr7_away | 16:10 | |
| notmyname | intersection is | and union is & right? | 16:11 |
| joeljwright | notmyname: also a good chance to see if this gate-tempest-dsvm-neutron-src-python-swiftclient bug strikes again | 16:11 |
| *** haomaiwa_ has joined #openstack-swift | 16:11 | |
| *** haomaiw__ has quit IRC | 16:11 | |
| notmyname | joeljwright: was there a patch that landed, or is it random? | 16:11 |
| joeljwright | still trying to work that out | 16:11 |
| joeljwright | notmyname: there are a no other swiftclient patches in a -verified state | 16:12 |
| joeljwright | notmyname: but I need to dig through them to see if people have been rechecking endlessly | 16:12 |
| eranrom | notmyname: yeah, |,& can be interpreted that way. Anyway, I was thinking more of a flag saying user metadata yes/no and a white list of additional headers, but am open to suggestions. | 16:12 |
| notmyname | joeljwright: I didn't see many (any?) rechecks in my gerrit emails this morning | 16:12 |
| notmyname | eranrom: yeah, that might work, but I'd be worried about the categorization/naming. ie naming is hard and making that clear (what is "user metadata" and is that clear to a user/operator)? | 16:14 |
| eranrom | notmyname: good point. an alternative would be just a white list. | 16:14 |
| notmyname | eranrom: my first guess (and thus can likely be improved) is to have a whitelist with x-cotnaienr-meta-* pattern and a list of explicit headers | 16:14 |
| notmyname | eranrom: yeah. a whitelist with a * pattern for stuff we know is always syncable (x-container-meta-*) | 16:15 |
| eranrom | notmyname: agree. | 16:15 |
| notmyname | I've seen this pattern somewhere else in swift. with objects | 16:15 |
| eranrom | will look for this | 16:15 |
| notmyname | eranrom: I think it's the object server allowed_headers | 16:15 |
| eranrom | ok, thanks | 16:15 |
| notmyname | eranrom: eg you can send any header you want for an object PUT. but swift will only remember the x-object-meta-* headers and the headers listed in allowed_headers on the object server config | 16:16 |
| eranrom | is there a container equivalent? | 16:16 |
| eranrom | I can look it up... | 16:16 |
| openstackgerrit | Romain LE DISEZ proposed openstack/swift: Close connections from SegmentedIterable on client disconnection https://review.openstack.org/193192 | 16:17 |
| notmyname | eranrom: IIRC, no | 16:17 |
| eranrom | notmyname: ok, perhaps its high time we have one :-) | 16:17 |
| notmyname | eranrom: seems like a thing independent of container sync syncing metadata :-) | 16:18 |
| eranrom | notmyname: agree. | 16:18 |
| notmyname | eranrom: personally, I wouldn't prioritize arbitrary container header setting as a thing. there's a lot of other stuff that I've heard a lot of end-user and operator requests for :-) | 16:19 |
| *** aix has quit IRC | 16:20 | |
| eranrom | notmyname: so you take back: "whitelist with x-cotnaienr-meta-* pattern and a list of explicit headers" | 16:20 |
| notmyname | eranrom: as stuff to sync or as stuff to set? | 16:21 |
| acoles | notmyname: eranrom iirc there is some filtering of headers passed to container/account servers in the proxy controllers | 16:21 |
| acoles | notmyname: eranrom take a look around swift/proxy/controllers/base.py transfer_headers() | 16:21 |
| eranrom | notmyname: I thought you were referring to a stuff to set, but I guess you meant stuff to sync | 16:22 |
| eranrom | acoles: sure will do | 16:22 |
| *** panbalag has joined #openstack-swift | 16:22 | |
| acoles | its been a while since i looked at it, so i hope i'm not confusing matters | 16:22 |
| *** openstackgerrit has quit IRC | 16:22 | |
| *** openstackgerrit has joined #openstack-swift | 16:23 | |
| acoles | but IIRC for object requests most headers get passed through to object server backend, where allowed_headers is applied, but for container and account the proxy controller filters more | 16:23 |
| *** panbalag has quit IRC | 16:23 | |
| notmyname | torgomatic: can you take a look at https://review.openstack.org/#/c/193192/ please? | 16:25 |
| eranrom | acoles: Any chance you refer to pass_through_headers in controllers/container.py? | 16:26 |
| acoles | eranrom: yes, so there you see the container/account specific 'allowed' list | 16:27 |
| eranrom | acoles: ok, getting back to the question of what container metadata to replicate as part of container sync. How about having a boolean flag at the container level saying replicate metadata yes/no if set to yes then the replcated metadata would be x-container-meta-* plus operators whilelist in container-sync.conf which can be a subset of the container 'allowed' list? | 16:29 |
| *** gsilvis has quit IRC | 16:30 | |
| eranrom | s/replcated/replicated | 16:30 |
| *** nadeem has joined #openstack-swift | 16:32 | |
| *** nadeem has quit IRC | 16:33 | |
| *** jasondotstar|afk is now known as jasondotstar | 16:33 | |
| *** nadeem has joined #openstack-swift | 16:33 | |
| acoles | eranrom: maybe. you would need to have the client send that new container level flag but somewhere in swift you'll need to translate the client generated header to a x-container-sysmeta-* key so it gets persisted. Then container sync can query it when required. | 16:37 |
| *** gsilvis has joined #openstack-swift | 16:37 | |
| acoles | eranrom: i guess it could be a yes/no or a client generated whitelist. depends on use cases. | 16:37 |
| *** NM has joined #openstack-swift | 16:40 | |
| eranrom | acoles: right. If I got notmyname, then he would like to start with a limited client configuration, e.g. no client generated whitlelist. | 16:40 |
| *** ni638629 has joined #openstack-swift | 16:40 | |
| acoles | eranrom: k, i will defer to whatever notmyname is saying, i was just being nosey and dropping a hint about the transfer_headers filter really :) | 16:41 |
| eranrom | acoles: ok, thanks. Question: If I translate the client's header to x-container-sysmeta-* key, how can the client tell the state? My understanding is that sysmeta is not visible but perhaps I am wrong | 16:42 |
| eranrom | aclose: state = whether this is a yes or a no to replicate container metadata | 16:42 |
| acoles | eranrom: you are correct. you would also need to translate the sysmeta value back to client header key in responses | 16:42 |
| eranrom | acoles: ok gotcha, np | 16:43 |
| *** ni638629 has left #openstack-swift | 16:43 | |
| acoles | sysmeta cannot be sent to/from client | 16:43 |
| acoles | but it does get persisted for you! | 16:43 |
| *** jasondotstar is now known as jasondotstar|afk | 16:44 | |
| eranrom | acoles: ok, is there a special naming convention for user defined metadata that is persisted as sysmeta (such as ACLs) | 16:45 |
| *** jasondotstar|afk is now known as jasondotstar | 16:46 | |
| *** jasondotstar is now known as jasondotstar|afk | 16:46 | |
| eranrom | acoles: or as ACLs should have been... | 16:46 |
| acoles | eranrom: heh, yeah there is stuff that predates sysmeta that gets plumbed all the way through, sysmeta just saves you some of that plumbing. | 16:48 |
| acoles | eranrom: i would suggest you use a header with key stemmed X-Container-Sync-* since it relates to X-Container-Sync-To etc | 16:49 |
| eranrom | acoles: good idea! will do. | 16:49 |
| eranrom | aclose: notmyname: I gotta go, this was very helpful, thanks. I will update the bug and code accordingly. | 16:50 |
| *** jasondotstar|afk is now known as jasondotstar | 16:53 | |
| *** jasondotstar is now known as jasondotstar|afk | 16:53 | |
| *** eranrom has left #openstack-swift | 16:55 | |
| *** NM has quit IRC | 17:01 | |
| *** geaaru has quit IRC | 17:02 | |
| *** krykowski has quit IRC | 17:03 | |
| *** jasondotstar|afk has quit IRC | 17:03 | |
| *** barker has quit IRC | 17:05 | |
| *** barker has joined #openstack-swift | 17:07 | |
| *** blmartin has quit IRC | 17:09 | |
| *** wbhuber has quit IRC | 17:11 | |
| *** NM has joined #openstack-swift | 17:14 | |
| *** petertr7_away is now known as petertr7 | 17:14 | |
| *** wbhuber has joined #openstack-swift | 17:16 | |
| *** jasondotstar has joined #openstack-swift | 17:20 | |
| *** jasondotstar is now known as jasondotstar|afk | 17:20 | |
| *** jasondotstar|afk is now known as jasondotstar | 17:20 | |
| *** rledisez has joined #openstack-swift | 17:21 | |
| *** acampbell has quit IRC | 17:21 | |
| *** barker has quit IRC | 17:22 | |
| openstackgerrit | paul luse proposed openstack/swift: EC Ssync: Update parms to include node and frag indices https://review.openstack.org/191521 | 17:25 |
| rledisez | swifterdarrell: hello darrell | 17:26 |
| *** jasondotstar is now known as jasondotstar|afk | 17:26 | |
| rledisez | swifterdarrell: i think i might need some help about writing unit test for https://review.openstack.org/#/c/193192/ | 17:26 |
| rledisez | swifterdarrell: i don’t see how to start, because it would involve launching a wsgi server, and inside an app establishing a connection from SegmentedIterable | 17:27 |
| *** barker has joined #openstack-swift | 17:27 | |
| rledisez | swifterdarrell: is it doable inside a unit test? seems a bit « too much ». what do you think? | 17:27 |
| *** mmcardle1 has quit IRC | 17:28 | |
| *** joeljwright has quit IRC | 17:28 | |
| *** jasondotstar|afk is now known as jasondotstar | 17:29 | |
| *** jlhinson has quit IRC | 17:29 | |
| *** barker has quit IRC | 17:31 | |
| *** barker has joined #openstack-swift | 17:34 | |
| notmyname | rledisez: if you look at the existing unit tests for FakeApp (there are several), you'll see the way we've tackled that sort of thing in the past | 17:38 |
| *** barker has quit IRC | 17:38 | |
| rledisez | notmyname: ok, i’ll check that. thx | 17:39 |
| notmyname | hmm... it is unfortunate that the string SegmentedIterable doesn't appear anywhere in the test/unit directory | 17:40 |
| *** breitz has quit IRC | 17:42 | |
| *** breitz has joined #openstack-swift | 17:42 | |
| rledisez | yeah, i grepped that and it’s not tested at all for now | 17:42 |
| notmyname | rledisez: it's indirectly tested by the dlo and slo range read tests | 17:44 |
| *** jlhinson has joined #openstack-swift | 17:50 | |
| *** openstackgerrit has quit IRC | 17:50 | |
| *** openstackgerrit has joined #openstack-swift | 17:51 | |
| *** marzif has joined #openstack-swift | 17:52 | |
| *** acampbell has joined #openstack-swift | 17:53 | |
| *** pgbridge has quit IRC | 17:54 | |
| *** fthiagogv has joined #openstack-swift | 17:56 | |
| *** pgbridge has joined #openstack-swift | 17:56 | |
| torgomatic | hm, looks like there's no unit tests at all for the closing of the underlying app iters | 17:57 |
| torgomatic | I did this http://paste.openstack.org/show/AG0em7UM9ZHSmp3T5KjU/ and nothing failed | 17:57 |
| notmyname | however, it reports 99% coverage :-/ | 18:02 |
| torgomatic | those lines are the 1% | 18:06 |
| *** jasondotstar is now known as jasondotstar|afk | 18:07 | |
| openstackgerrit | Merged openstack/swift: Use just IP, not port, when determining partition placement https://review.openstack.org/191970 | 18:10 |
| *** acoles is now known as acoles_away | 18:11 | |
| *** jasondotstar|afk is now known as jasondotstar | 18:15 | |
| *** jasondotstar is now known as jasondotstar|afk | 18:15 | |
| *** ronenkat has quit IRC | 18:16 | |
| openstackgerrit | Merged openstack/python-swiftclient: Fix docstring typo for SwiftService.upload https://review.openstack.org/193187 | 18:19 |
| *** petertr7 is now known as petertr7_away | 18:26 | |
| *** NM has quit IRC | 18:27 | |
| *** jasondotstar|afk has quit IRC | 18:27 | |
| *** NM has joined #openstack-swift | 18:31 | |
| openstackgerrit | Michael Barton proposed openstack/swift: go: document replication https://review.openstack.org/193244 | 18:33 |
| notmyname | redbo: thanks! | 18:36 |
| *** marzif has quit IRC | 18:37 | |
| *** mariusv has joined #openstack-swift | 18:38 | |
| *** mariusv has quit IRC | 18:38 | |
| *** mariusv has joined #openstack-swift | 18:38 | |
| *** marzif has joined #openstack-swift | 18:40 | |
| *** lcurtis has joined #openstack-swift | 18:41 | |
| *** mariusv has quit IRC | 18:42 | |
| *** petertr7_away is now known as petertr7 | 18:42 | |
| *** mariusv has joined #openstack-swift | 18:43 | |
| *** mariusv has quit IRC | 18:43 | |
| *** mariusv has joined #openstack-swift | 18:43 | |
| *** mariusv has quit IRC | 18:43 | |
| *** NM has quit IRC | 18:47 | |
| *** marzif has quit IRC | 18:47 | |
| *** jasondot_ has joined #openstack-swift | 19:01 | |
| *** jasondot_ is now known as jasondotstar|afk | 19:07 | |
| *** wer_ has quit IRC | 19:07 | |
| *** acampbell has quit IRC | 19:08 | |
| *** blmartin has joined #openstack-swift | 19:10 | |
| *** uschreiber_ has joined #openstack-swift | 19:23 | |
| *** uschreiber_ has quit IRC | 19:25 | |
| *** jasondotstar|afk is now known as jasondot_ | 19:27 | |
| *** jasondot_ is now known as jasondotstar|afk | 19:27 | |
| *** jasondotstar|afk has quit IRC | 19:27 | |
| *** acampbell has joined #openstack-swift | 19:31 | |
| *** zhill has quit IRC | 19:32 | |
| *** NM has joined #openstack-swift | 19:33 | |
| torgomatic | oh man, this closing stuff is a total mess | 19:36 |
| *** jasondotstar has joined #openstack-swift | 19:36 | |
| torgomatic | there's just tons of places where we don't close WSGI iterables | 19:36 |
| rledisez | actually, i’m looking at the server where i applyed my patch | 19:36 |
| rledisez | the patch does not fix everything there is still leakage | 19:36 |
| torgomatic | heh, yeah, no kidding :| | 19:37 |
| notmyname | remember that one time we though "yeah, we handle sockets well now. no more socket leaks. yay" | 19:39 |
| torgomatic | software is a cross between gardening and nailing jello to a tree | 19:43 |
| openstackgerrit | Darrell Bishop proposed openstack/swift: Allow 1+ object-servers-per-disk deployment https://review.openstack.org/184189 | 19:44 |
| *** jasondotstar is now known as jasondotstar|afk | 19:47 | |
| swifterdarrell | notmyname: clayg: torgomatic: tdasilva: mattoliverau: portante: kota: cschwede: rebased and ready to go ^^^^^^^^^^^^ | 19:49 |
| *** jasondotstar|afk is now known as jasondotstar | 19:52 | |
| openstackgerrit | Minwoo Bae proposed openstack/swift: EC Reconstructor: Do not reconstruct existing fragments. https://review.openstack.org/193279 | 19:53 |
| *** delatte has joined #openstack-swift | 19:59 | |
| *** jlhinson has quit IRC | 20:01 | |
| *** zaitcev has joined #openstack-swift | 20:02 | |
| *** ChanServ sets mode: +v zaitcev | 20:02 | |
| clayg | i love this go community! | 20:02 |
| clayg | I'm trying to figure out why I can't get hummingbird to show up when I 'godoc -http=:6060' | 20:02 |
| clayg | find this https://code.google.com/p/go/issues/detail?id=1540 | 20:02 |
| *** delattec has quit IRC | 20:02 | |
| clayg | and the comment by russ "working-as-intended" | 20:03 |
| clayg | "if you top letting yourself get pushed around by FHS you will be happier" | 20:03 |
| clayg | it's like "all those guys trying to make distros are idiots - just put the files where my tool wants them and it will be fine" | 20:03 |
| *** jasondotstar is now known as jasondotstar|afk | 20:06 | |
| *** jasondotstar|afk is now known as jasondotstar | 20:06 | |
| clayg | i'm not sure i follow the subsequent discussion - but it *looks* like the debian package just carries the "make godoc work with symlinks" as a patch!? | 20:06 |
| *** jasondotstar is now known as jasondotstar|afk | 20:06 | |
| clayg | that's awesome | 20:06 |
| openstackgerrit | Michael Barton proposed openstack/swift: go: restructure cmd/hummingbird.go https://review.openstack.org/188939 | 20:07 |
| *** jasondotstar|afk is now known as jasondotstar | 20:08 | |
| swifterdarrell | clayg: It's not their fault the rest of the world is doing it wrong... | 20:10 |
| *** wbhuber_ has joined #openstack-swift | 20:10 | |
| *** wbhuber has quit IRC | 20:10 | |
| *** SkyRocknRoll has quit IRC | 20:12 | |
| zaitcev | clayg: It is sometimes necessary to carry a fork. Usually it's avoided by distros because it's labor intensive. | 20:14 |
| *** jlhinson has joined #openstack-swift | 20:22 | |
| *** delattec has joined #openstack-swift | 20:25 | |
| *** delatte has quit IRC | 20:28 | |
| *** aix has joined #openstack-swift | 20:33 | |
| openstackgerrit | Samuel Merritt proposed openstack/swift: Get better at closing WSGI iterables. https://review.openstack.org/193303 | 20:34 |
| torgomatic | rledisez: I've just proposed a patch that should plug more of the holes. If you have time, could you try it out and see if it works (and comment on the patch with your findings)? | 20:35 |
| *** Fin1te has quit IRC | 20:35 | |
| *** silor has quit IRC | 20:37 | |
| rledisez | torgomatic: sure, i’ll give it a try | 20:37 |
| *** marzif has joined #openstack-swift | 20:41 | |
| *** fthiagogv has quit IRC | 20:46 | |
| clayg | minwoob: hey did you run tests on patch 193279 - i think we're not doing a great job mocking response headers | 20:52 |
| patchbot | clayg: https://review.openstack.org/#/c/193279/ | 20:52 |
| minwoob | clayg: Yeah, I was trying to get them to work. | 20:55 |
| minwoob | clayg: I'll keep digging at it. | 20:55 |
| clayg | yeah with your change the mocked responses need to set the fragment index in the response - or otherwise they'll all be "None" and thye'll get skipped | 20:56 |
| clayg | you could make your test verify the header is not-None before it throws it away (letting the tests get off w/o having to set the header) - but that seems like cheating | 20:56 |
| minwoob | Right. | 20:58 |
| *** Fin1te has joined #openstack-swift | 21:00 | |
| *** nadeem has quit IRC | 21:00 | |
| *** jamielennox is now known as jamielennox|away | 21:01 | |
| *** nadeem has joined #openstack-swift | 21:02 | |
| *** marzif has quit IRC | 21:03 | |
| *** acampbell has quit IRC | 21:07 | |
| clayg | minwoob: so this test might get you started -> https://gist.github.com/clayg/4aebb0a21589d485785e | 21:09 |
| *** jamielennox|away is now known as jamielennox | 21:09 | |
| * notmyname is currently running benchmark tests for the async container updates patch | 21:09 | |
| clayg | it doesn't cover the case of duplicates - you need a seperate test for that - and it doesn't blow up on master because the jerasure backend doesn't have the same problem as isa_l | 21:09 |
| clayg | notmyname: doesn't the go server already do that? | 21:10 |
| notmyname | perhaps. I think so | 21:10 |
| notmyname | https://review.openstack.org/#/c/189080/ | 21:11 |
| *** petertr7 is now known as petertr7_away | 21:11 | |
| *** Fin1te has quit IRC | 21:18 | |
| minwoob | clayg: Got it. Thank you! | 21:20 |
| openstackgerrit | Merged openstack/swift: go: document replication https://review.openstack.org/193244 | 21:28 |
| *** ianbrown has joined #openstack-swift | 21:29 | |
| *** jlhinson has quit IRC | 21:32 | |
| *** jlhinson has joined #openstack-swift | 21:33 | |
| *** rledisez has left #openstack-swift | 21:34 | |
| *** jrichli has quit IRC | 21:34 | |
| *** jlhinson has quit IRC | 21:42 | |
| *** blmartin has quit IRC | 21:45 | |
| notmyname | baseline run finished | 22:04 |
| *** peluse has quit IRC | 22:15 | |
| *** minwoob has quit IRC | 22:18 | |
| *** jkugel has quit IRC | 22:20 | |
| mattoliverau | Morning | 22:22 |
| notmyname | clayg: greenthread.GreenThread().wait() doesn't take any arguments :-( | 22:22 |
| notmyname | hi mattoliverau | 22:22 |
| notmyname | clayg: but if wait() blocks the reactor (at least for the current thread), then I can't wrap it in a Timeout either or I'll just get hub starvation, right? | 22:25 |
| clayg | wat? | 22:27 |
| notmyname | I could do sawn_n and then eventlet.sleep(100), but that sounds pretty horrible | 22:27 |
| clayg | with Timeout(): gt.wait() | 22:27 |
| notmyname | the timeout won't have a chance to fire until after wait() finishes | 22:27 |
| clayg | if the gt blocks spawn or spawn_n won't make a difference | 22:27 |
| notmyname | (there's a little bit of a question there) | 22:27 |
| clayg | notmyname: heh - one of us is confused :D | 22:28 |
| notmyname | I vote for me :-) | 22:28 |
| * clayg goes to hack | 22:28 | |
| notmyname | my first test shows I'm the confused one | 22:31 |
| notmyname | https://gist.github.com/notmyname/0004945577d4b3cd2855 finishes in about 2 seconds | 22:31 |
| notmyname | ie showing that .wait() doesn't block the reactor until the method finishes | 22:32 |
| clayg | this works sort of like I expect - https://gist.github.com/clayg/1313818bbf96688d0427 | 22:33 |
| clayg | notmyname: lol | 22:33 |
| clayg | notmyname: only so many ways to spell it I guess | 22:33 |
| notmyname | :-) | 22:33 |
| clayg | you beat me tho :\ | 22:33 |
| notmyname | I didn't worry with pesky usage strings ;-) | 22:33 |
| clayg | anyway - I think it would acctually make the diff a little smaller because less test churn | 22:34 |
| clayg | they fact that we won't have to revert it after we release it is a bonus | 22:34 |
| notmyname | yeah, the thing that convinces me that you are right is the user expectation that normally the listings are right | 22:35 |
| clayg | *less test churn | 22:35 |
| clayg | fucking users | 22:35 |
| clayg | storage would be easy if a) we didn't have to give the data back or b) no one used it | 22:35 |
| notmyname | I liked it whenyou said somethign earlier, but I had already started down this road and wanted to see what it looked like (so I wasn't ignoring you earlier) | 22:35 |
| clayg | :P | 22:35 |
| clayg | I SHALL NOT BE IGNORED!!!! | 22:35 |
| notmyname | but, FWIW, I'm currently runnign the patch that's in gerrit in my test. should be done in ~45 minutes | 22:36 |
| clayg | sure np, you can tell me to stow it too | 22:36 |
| notmyname | basically, 3600 seconds of zero-byte PUTs to 1 container (2 proxies, 2 storage servers, all spinning disks) | 22:36 |
| *** wbhuber_ has quit IRC | 22:38 | |
| notmyname | current master was able to do, well... it's taking a while to report what it did | 22:39 |
| notmyname | 95% latency of .441 with max latency of 5.561 | 22:39 |
| *** delattec has quit IRC | 22:45 | |
| zigo | swifterdarrell: Any update on fixing the 1.0.7m release? | 22:45 |
| swifterdarrell | zigo: I heard that the upstream dudes were making a 1.0.8 that will be awesome | 22:46 |
| *** jasondotstar has quit IRC | 22:46 | |
| zigo | swifterdarrell: I don't need awesome, I just need something that works ! :/ | 22:46 |
| swifterdarrell | zigo: best bet would be to get in contact with paul, tushar, or kevin directly | 22:47 |
| zigo | swifterdarrell: I only have the email address of Tushar, can you give me the others? | 22:48 |
| swifterdarrell | zigo: probalby not, but clayg surely can! | 22:48 |
| swifterdarrell | clayg: (nevermind, send emails via PM) | 22:50 |
| *** ianbrown has quit IRC | 22:54 | |
| *** NM has quit IRC | 22:55 | |
| *** openstackgerrit has quit IRC | 22:55 | |
| *** tellesnobrega has quit IRC | 22:55 | |
| *** klrmn has quit IRC | 22:55 | |
| *** EmilienM has quit IRC | 22:55 | |
| *** dmsimard has quit IRC | 22:55 | |
| clayg | ho: holy crap man - thanks for the responses in the review on patch 149930 - I don't think i understood quite a badly the olso.policy's register check pattern was fighting us :\ | 22:56 |
| patchbot | clayg: https://review.openstack.org/#/c/149930/ | 22:56 |
| *** lcurtis has quit IRC | 22:57 | |
| *** km has joined #openstack-swift | 22:59 | |
| *** NM has joined #openstack-swift | 23:06 | |
| *** openstackgerrit has joined #openstack-swift | 23:06 | |
| *** tellesnobrega has joined #openstack-swift | 23:06 | |
| *** klrmn has joined #openstack-swift | 23:06 | |
| *** EmilienM has joined #openstack-swift | 23:06 | |
| *** dmsimard has joined #openstack-swift | 23:06 | |
| *** ianbrown has joined #openstack-swift | 23:06 | |
| *** marzif has joined #openstack-swift | 23:10 | |
| openstackgerrit | Samuel Merritt proposed openstack/swift: Get better at closing WSGI iterables. https://review.openstack.org/193303 | 23:12 |
| *** nadeem has quit IRC | 23:18 | |
| *** kei_yama has joined #openstack-swift | 23:19 | |
| *** NM has quit IRC | 23:42 | |
| *** thurloat is now known as thurloat_isgone | 23:45 | |
| *** chlong has joined #openstack-swift | 23:47 | |
| *** ho has joined #openstack-swift | 23:47 | |
| ho | good morning! | 23:53 |
| torgomatic | hi | 23:55 |
| ho | torgomatic: hello! | 23:57 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!