Saturday, 2015-08-01

*** dogweather has joined #openstack-swift00:00
*** dogweather has left #openstack-swift00:02
openstackgerritCarlos Cavanna proposed openstack/swift: Improving statistics sent to Graphite.  https://review.openstack.org/20265700:10
*** theannegentle has quit IRC00:12
*** aix has quit IRC00:23
claygtorgomatic: what do you think about patch 18567900:24
patchbotclayg: https://review.openstack.org/#/c/185679/00:24
torgomaticclayg: it's the best patch in the whole world00:24
claygyou tried to help and then acoles_away and me were like w/e - then I tried to help - and now it's in conflict00:24
claygyeah but I think the diskfile refactor made it in - so that one should be good I think00:25
torgomaticyeah, I should rebase that one00:25
claygoh there was just that crappy comment - i guess I can pick it back up00:25
claygoh - yeah that's better :)00:25
*** DericHorn-HP has quit IRC00:26
claygup to you - it was all your good idea - I'm happy to shepered it in for you too00:26
torgomaticI'll try to get to it pretty soon00:27
claygkthnxbai00:27
*** DericHorn-HP has joined #openstack-swift00:33
*** DericHorn-HP has quit IRC00:33
*** NM1 has joined #openstack-swift00:37
openstackgerritClay Gerrard proposed openstack/swift: cleanup from reaper fixes  https://review.openstack.org/20810600:41
*** hrou has quit IRC00:43
*** hrou has joined #openstack-swift00:43
*** haigang has joined #openstack-swift00:52
*** NM1 has quit IRC01:40
claygcharz: you may be weekend already -> community cluster functests failed with weird error -> [swiftqa03] out: Could not open requirements file: [Errno 2] No such file or directory: '/home/swiftqa/swift-2.3.0-244-g8c2d6f6/requirements.txt'02:00
openstackgerritOpenStack Proposal Bot proposed openstack/python-swiftclient: Updated from global requirements  https://review.openstack.org/8925002:20
openstackgerritOpenStack Proposal Bot proposed openstack/swift: Updated from global requirements  https://review.openstack.org/8873602:21
*** haigang has quit IRC02:39
*** proteusguy has joined #openstack-swift02:39
ccavanna_clayg: Hi Clay. Quick question about automated checks. I checked in some fixes and I noticed that "SwiftStack Cluster CI check" ran. I didn't see those runs before.02:45
*** haigang has joined #openstack-swift02:56
*** hrou has quit IRC02:57
*** haigang has quit IRC03:14
*** haigang has joined #openstack-swift03:16
*** zul has quit IRC03:22
claygccavanna_: I think those were disabled while charz was using those nodes for other stuff - so now they're coming back on and - well i think they're broke03:23
claygcharz: will fix them tho - he's a ninja03:24
claygya'll have a good weekend!03:24
ccavanna_clayg: they look broken :-)03:25
ccavanna_clayg: or my code is broken in a fantastic way :-p03:26
*** jrichli_ has quit IRC03:26
ccavanna_clayg: on a serious note :-), it seems a setup issue. It's returning 401 Unauthorized when trying to establish a connection.03:29
*** zul has joined #openstack-swift03:37
*** nadeem has joined #openstack-swift04:01
*** mfalatic has quit IRC04:11
notmynamegood evening04:23
*** DericHorn-HP has joined #openstack-swift04:35
*** ccavanna_ has quit IRC04:37
*** DericHorn-HP has quit IRC05:05
*** proteusguy has quit IRC05:10
*** changbl has joined #openstack-swift05:21
charzccavanna_ clayg: thanks for notice, I'm working on it. :-)05:24
notmynamecharz: thanks05:28
charznotmyname: np, have a nice weekend!05:30
*** nadeem has quit IRC05:30
*** zaitcev has quit IRC05:33
*** DericHorn-HP has joined #openstack-swift05:38
*** DericHorn-HP has quit IRC05:39
tamizh_geekHey clayg: About patch 204196, I'm not sure if I understand your comments fully.05:47
patchbottamizh_geek: https://review.openstack.org/#/c/204196/05:47
tamizh_geekclayg: available for a quick chat ?05:47
notmynametamizh_geek: I'm looking, but don't expect clayg to be here. it's 11pm on friday night for him05:56
*** DericHorn-HP has joined #openstack-swift05:56
tamizh_geeknotmyname: oh sorry! thanks for your help.05:56
*** DericHorn-HP has quit IRC05:58
notmynametamizh_geek: so i think his concern is that we don't always pass a request when dealing with swob objects. and that will result in inconsistent results (ie people noticing that their error message they are seeing /doesn't/ do json)05:59
notmynametamizh_geek: and the way to fix it is audit the entire codebase for that. and that would be...rather larger...than your current patch05:59
notmynametamizh_geek: at least that's how I read it ;-)06:00
tamizh_geeknotmyname: Firstly, there are no test failures for this. And also pardon my ignorance - If the swob response is not passed any request object, and if the swob itself is cooking up a request object - in this case it will return a non-JSON response.06:00
tamizh_geekwhich is not changing the default behaviour existing before this patch right?06:01
notmynameI haven't looked at the test coverage yet, so I can't comment there ;-)06:02
notmynamebut yeah, defaulting to the current behavior is good06:02
notmynamebut I think his thought is that we'll have app developers sending an Accept header and sometimes seeing it respected and sometimes not. that's the inconsistency06:03
notmynameuntil the request object is plumbed through in every call06:04
tamizh_geekOk. I'll be glad to send another patch to bring in a consistent behaviour so that all over the code base, the swob.Response is always instantiated with a swob.Request object.06:05
tamizh_geekBut, one thing I want to confirm is, is the fact that "swob.Response cooks up a request object if we don't send one" is correct?06:05
tamizh_geekAFAIK, that doesn't happen. We haven't noticed this before because, the error response don't really care about what came up in a request. So they are *mostly* instantiated without a request object and everything was fine.06:08
tamizh_geekSo there are two ways to handle this. 1) Bring in consistency by making the Request object a mandatory argument to the Response class. 2) Inside the Response#_get_error_response_string check if the request is None: then default to sending text/html response.06:09
tamizh_geekany comments ?06:10
notmynamelooking :-)06:11
tamizh_geeknotmyname: Thanks :) Hope I'm not pestering you in a weird time!06:12
notmynameno, not pestering. I was working on refactoring some other code. I've been working on some community tracking via commits and reviews and the organically-grown codebase needs an overhaul to be able to give me the new data I want06:13
notmynametamizh_geek: ah. swob does "cook up" a request06:15
notmynamein master's version of swob, look at line 138706:16
notmynamebasically, when a response is called, the request is generated if it wasn't passed in to the __init__06:16
*** haigang has quit IRC06:16
tamizh_geekhmmm., checking.06:17
tamizh_geekoh yeah!06:18
tamizh_geekso in that case, in real life scenarios, the request object will be recreated from the env. And the 'Accept' headers will still be respected.06:19
tamizh_geekAm i right? @notmyname06:19
notmynameyes, that sounds right, but I don't currently remember everywhere that env comes from when calling the swob object06:23
notmynametamizh_geek: however, I think i've figured out the last part of what clayg said06:23
notmyname(it normally takes me a bit ;-)06:23
tamizh_geekhehe.06:23
tamizh_geekand what was it?06:24
notmynamethe reason you had to add the blank request is because of the way the test is getting the content on line 1412 in your patch06:24
notmynamethat test was "bypassing the swob machinery" by accessing some hidden attributes of the response object06:25
notmynameso the gist he linked shows how to do it without adding the blank request06:25
notmynameie https://gist.github.com/clayg/1e8cbd0b90eb66eb3ae0#file-request-is-should-patch-L2606:25
notmynamedirectly call the response with an empty env and a callable that does nothing. this makes swob create the request and carry on06:26
tamizh_geekooh yeah. since the response is a callable, I just have to call it with a empty env.06:26
tamizh_geekright!06:26
notmynameat least, I think that's all that's happening. that's based on the assumption that not creating the blank request cause the test to fail in your patch (based on your comment. I haven't run it)06:27
tamizh_geeklet me try it right away.06:28
notmynametamizh_geek: the other thing to try is to send a DELETE to an object that doesn't exist and an accept header. try with and without your patch. that will tell you if "the env is set so it should just work" , as you supposed above, is actualy what happens06:32
notmynametamizh_geek: ie the other part of the gist from clayg06:32
tamizh_geeknotmyname: clayg is right. The old tests pass without a explicit request object when change the way how I elicit the response.06:32
*** DericHorn-HP has joined #openstack-swift06:32
*** DericHorn-HP has quit IRC06:32
notmynameok, that's good06:32
tamizh_geekhmm yeah, will check a DELETE request now.06:33
notmynameand so the question (for us devs) is if it works without modifying the object server to pass in the request or if the object server must be modified before the json will be returned06:34
notmynameand if it's the latter, that means that users will see inconsistent error message formats, even if when they send the accept header, until we audit the entire codebase for that06:35
*** csmart has quit IRC06:36
*** csmart has joined #openstack-swift06:36
notmynameclayg: it only took me 30 minutes to understand your 7-line comment! and you had code and everything! ;-)06:37
tamizh_geeknotmyname: Got it now. Thanks for awesome explanation. Appreciate very much.06:37
notmynametamizh_geek: what did you see with the 404 on delete?06:38
tamizh_geekFiring up my vm.06:38
notmynameheh, I should too ;-)06:38
notmynamehmm..seemed to work on my machine. I think06:41
notmynameah, wait06:43
notmynamejust a delete is the wrong test06:43
tamizh_geeknotmyname: yes -  https://gist.github.com/tamizhgeek/d509dbff10c9dbb5284806:43
notmynameyeah, that's what I saw too, but it's the wrong code path06:44
tamizh_geekits working.06:44
notmynamehang on06:44
tamizh_geekshould I just write a functional test and verify it ?06:44
notmyname...that test still worked. one more thing to try06:46
notmynameok that passed too06:47
tamizh_geekcool!06:47
notmynametamizh_geek: are you looking at the object server code?06:47
notmynamethe delete method?06:47
tamizh_geekyeah at it.06:48
notmynameline 835 is what clayg modified in his gist. because it doesn't pass the request in06:48
notmynameso to get that code path, we need a delete on a missing object with an x-delete-at header06:48
notmynameand that works on my machine06:48
notmynameso in order to really break it, I removed line 85406:48
notmynamethe request=request, on the call to whatever ends up as the response class06:49
notmynameand that still works, from what I can tell06:49
tamizh_geekhmmm ok ok.06:49
notmynameand I switch back to master, and I get html always06:50
tamizh_geekso is this a good to go? I can add this confusion scenario as a func test and resubmit the patch ?06:50
notmynameso, point being, I think you're right about the env being passed to the auto-created request and it works06:50
tamizh_geeknotmyname: ?06:52
notmynameI would like to see the test_507() method cleaned up as clayg suggested in his gist. it will make the test better overall. let me look at the rest of the patch and I'll leave a review06:53
tamizh_geeknotmyname: thanks. will do that.06:53
notmynameas to a functional test, it would be nice, of course, but I wouldn't hold this up for it.06:54
notmynamewell, don't hold me to that yet. let me think on it while I look at what you already have ;-)06:55
tamizh_geekok then. will clean up test_507 and submit a patch right away. The func test will add in the very next patch, quoting this conversation.06:56
tamizh_geeknotmyname: sure :)06:56
charznotmyname: around?06:58
notmynamecharz: yeah ;-)06:58
charznotmyname: Did you know why setup.py is required pbr and it not includes in requirement.txt?06:59
charznotmyname: just curious07:00
notmynamecharz: because it's setup.py that is parsing requirements.txt with pbr. so it needs to be available before requirements.txt is even looked at, which is why it wasn't added there. by the time requirements.txt is looked at, by definition you already have pbr installed07:00
notmynamehowever, if it makes things easier to include pbr in requirements.txt, I'd certainly support adding it07:01
openstackgerritAzhagu Selvan SP proposed openstack/swift: Respect 'Accept' header in error responses  https://review.openstack.org/20419607:03
charznotmyname: oh, that's make sense to use it to parse requirements.txt.07:03
charzIn general, if I use `pip install swift` from pip repo, pbr is not required, is it correct?07:04
notmynamecharz: in general I don't think that will work because swift isn't on pypi07:04
notmynameas far as I know07:04
notmynamecharz: but in general, pbr should not be required if you install swift from packages07:05
charznotmyname: oh! that's what I meant. Thanks07:06
tamizh_geeknotmyname: uploaded a new patch 20419607:06
patchbottamizh_geek: https://review.openstack.org/#/c/204196/07:06
tamizh_geekthanks for your help!07:06
notmynameyup, just pulled it07:06
notmynametamizh_geek: out of curiosity, what time zone are you in? and what brings you to swift?07:07
notmynametamizh_geek: for some code archeology, here's why your patch isn't causing problems: https://review.openstack.org/#/c/19422/07:13
notmynameclayg: ^07:13
notmynametamizh_geek: I'm running the last set of tests now07:17
*** DericHorn-HP has joined #openstack-swift07:20
*** DericHorn-HP has quit IRC07:26
tamizh_geeknotmyname: I'm in IST. I am based out of Chennai, India.07:35
notmynameah, great!07:35
tamizh_geeknotmyname: Few months back, I came across the Openstack and swift while discussing something between friends. Checked this out, looked great. Always wanted to contribute to a open source project.07:36
tamizh_geekSwift looked very interesting.07:36
tamizh_geekSo I'm here :)07:36
notmynameyay!07:37
notmynameI'm very happy you're here07:37
tamizh_geeknotmyname: Thanks!07:37
claygnotmyname: oh of course - when it makes up a response it's doing it from the environ *actual* environ - duh07:38
notmynametamizh_geek: so you got me to look into a functional test for this, and I've got one. but it's not passing yet07:38
tamizh_geekwhich is that? May be I can help ?07:39
notmynameI delete a missing object and send an accept header and still get html back07:40
tamizh_geekby missing object you mean a non-existent object right ?07:41
notmynameright07:41
notmynamehttps://gist.github.com/notmyname/6d4f936215edc2ad59b007:42
tamizh_geeklooks like the response is json and the content-type is not.07:45
notmynameright07:46
notmynameah, happens with curl too07:47
notmynameso theres an issue there07:47
notmynamethe unit tests aren't checking that either07:48
tamizh_geekyeah right. My bad! will fix it upload a new patch. thanks for the excellent catch.07:50
notmynameclayg: are you staying up for much linger?07:51
notmynamelonger07:51
claygnotmyname: idk, i'm watching rome on prime07:51
claygummm 411 is still text/plain too :\07:51
notmynameclayg: heh, ok :-)07:51
notmynameIf y'all do get it straightened out, i'll take another look tomorrow night. but I've got to be up in about 5.5 hours to go to rainier so I should probably get to bed ;-)07:52
notmynameclayg: or just watch rome, 'cause it's late ;-)07:53
claygso check_object_creation returns text/plain07:56
tamizh_geeknotmyname: actually unit tests pass. the test_507_json_response explicitly tests the resp.content_type and it passes. The curl to delete a non existent object returns the correct content-type header as well : https://gist.github.com/tamizhgeek/8932641245299cab9b1c08:11
tamizh_geekclayg: which is the 411 test you were mentioning?08:11
*** DericHorn-HP has joined #openstack-swift08:35
*** DericHorn-HP has quit IRC08:39
*** marzif has quit IRC08:40
*** ig0r_ has joined #openstack-swift09:40
*** DericHorn-HP has joined #openstack-swift09:45
tamizh_geekclayg: Got it from the gerrit09:49
*** ig0r_ has quit IRC09:53
tamizh_geekweird enough I get 404 as the response for the same curl : https://gist.github.com/tamizhgeek/bb876c7267d668310faa09:54
*** ig0r_ has joined #openstack-swift09:59
*** ig0r_ has quit IRC10:07
*** ig0r_ has joined #openstack-swift10:10
*** silor has joined #openstack-swift10:14
*** ig0r_ has quit IRC10:26
*** silor has quit IRC10:39
openstackgerritNakagawa Masaaki proposed openstack/swift: SAIO on Ubuntu 14.04 can not pass unittest.  https://review.openstack.org/20779011:02
*** sanchitmalhotra has joined #openstack-swift11:06
*** sanchitmalhotra has quit IRC11:18
janonymousDoes current eventlet will work on py3 , Is Eventlet required to be upgraded for py3 :  eventlet 0.17.3 now fully supports Python 3 , or is there a plan to move to asyncio11:47
*** SkyRocknRoll has joined #openstack-swift11:54
*** haigang has joined #openstack-swift11:57
*** haigang has quit IRC11:58
*** ig0r_ has joined #openstack-swift11:58
*** foobar__ has joined #openstack-swift12:51
foobar__Hi everybody. Is there anybody who can help me with tempurls in swift?12:52
foobar__I have setup everything and it works fine. Now I need an additional query parameter at the end of my proxy pipeline. TempURL seems to remove these parameters. Is there any way to do this? Headers are not working, since I wan't to embedd the documents into a website and, depending on the query parameter render a specific pregenerated thumbnail...12:53
*** haigang has joined #openstack-swift13:14
*** haomaiwang has quit IRC13:18
*** haomaiwang has joined #openstack-swift13:18
*** jrichli_ has joined #openstack-swift13:28
*** ig0r_ has quit IRC13:28
*** jrichli_ has quit IRC13:42
*** haomaiwang has quit IRC14:01
*** haomaiwang has joined #openstack-swift14:01
*** changbl has quit IRC14:07
*** ig0r_ has joined #openstack-swift14:12
*** ig0r_ has quit IRC14:16
*** kcaj has quit IRC14:21
*** kcaj has joined #openstack-swift14:26
*** hrou has joined #openstack-swift14:26
*** DericHorn-HP has quit IRC14:29
*** sanchitmalhotra has joined #openstack-swift14:36
*** haigang has quit IRC14:52
*** ig0r_ has joined #openstack-swift14:53
*** sanchitmalhotra has quit IRC14:53
*** hugokuo has quit IRC14:55
*** hugokuo has joined #openstack-swift14:58
*** ig0r__ has joined #openstack-swift15:15
*** ig0r_ has quit IRC15:18
*** xnox has quit IRC15:25
*** xnox has joined #openstack-swift15:27
*** hrou has quit IRC15:53
*** pcaruana has joined #openstack-swift15:56
*** foobar__ has quit IRC15:58
*** alejandrito_ has joined #openstack-swift16:12
*** DericHorn-HP has joined #openstack-swift16:50
*** jrichli has joined #openstack-swift17:19
*** DericHorn-HP has quit IRC17:29
*** DericHorn-HP has joined #openstack-swift17:48
*** mowly has joined #openstack-swift17:59
mowlyHi all. I have got a question. I am trying since 6 hours without success. Maybe anybody can support me? I have a formpost setup but cannot read out the element within the multipart request. The object ist successfully stored in swift but I need to generate a thumbnail with this object.18:01
mowlyWhen I call request.body_file I get a _CappedFileLikeObject18:01
mowlyhow do i create an image object from this? I saw there is a .read() function, but this do not work... Any tips maybe??18:02
mowlywhen i upload the file via swift cli client everything works perfect. In this case is simply use the request.body to create the thumbnail18:07
*** DericHorn-HP has quit IRC18:09
*** superflyy has joined #openstack-swift18:12
*** SkyRocknRoll has quit IRC18:17
*** DericHorn-HP has joined #openstack-swift18:29
*** ig0r_ has joined #openstack-swift18:32
*** jrichli_ has joined #openstack-swift18:34
*** superflyy has quit IRC18:34
*** jrichli has quit IRC18:37
*** hrou has joined #openstack-swift18:41
*** ig0r_ has quit IRC18:42
*** DericHorn-HP has quit IRC18:50
openstackgerritJanie Richling proposed openstack/swift: Trivial Key Master for encryption  https://review.openstack.org/19374919:01
*** mordred has quit IRC19:20
*** ig0r__ has quit IRC19:21
*** mordred has joined #openstack-swift20:12
openstackgerritJanie Richling proposed openstack/swift: Cryptography module to be used by middleware  https://review.openstack.org/19382620:17
*** mowly has quit IRC20:23
*** olletsocmit has joined #openstack-swift20:55
*** olletsocmit has left #openstack-swift20:57
*** kcaj has quit IRC21:12
*** kcaj has joined #openstack-swift21:15
*** alejandrito_ has quit IRC21:21
openstackgerritMerged openstack/python-swiftclient: Block comment PEP8 fix.  https://review.openstack.org/20584921:33
openstackgerritJanie Richling proposed openstack/swift: Encrypter and Decrypter middleware - object body only  https://review.openstack.org/19419121:59
*** jrichli_ has quit IRC22:00
*** jrichli has joined #openstack-swift22:05
*** NM has joined #openstack-swift22:08
*** jrichli has quit IRC22:09
openstackgerritOpenStack Proposal Bot proposed openstack/python-swiftclient: Updated from global requirements  https://review.openstack.org/8925023:44
*** kcaj has quit IRC23:46
*** kcaj has joined #openstack-swift23:46

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!