Monday, 2016-08-22

*** m_kazuhiro has joined #openstack-swift00:54
*** jfizzlezizzle5 has quit IRC01:22
*** zhengyin has joined #openstack-swift01:39
kota_Hell02:21
kota_tyhoon has come to Tokyo so staying at home02:22
*** Jeffrey4l has joined #openstack-swift02:24
notmynamekota_: yikes. https://weather.com/storms/hurricane/news/japan-tropical-storm-mindulle-lionrock-kompasu02:24
notmynamenot one, but three02:24
notmynamekota_: stay safe02:24
kota_notmyname: yeah but i think one is now at Tokyo and the others are heading for far so maybe it'll be fine by this night02:27
*** csmart has quit IRC02:33
*** csmart has joined #openstack-swift02:35
*** tqtran has joined #openstack-swift03:02
*** tqtran has quit IRC03:06
*** chrichip has quit IRC03:23
*** chrichip has joined #openstack-swift03:23
*** m_kazuhiro has quit IRC03:44
*** aagrawal has joined #openstack-swift04:01
*** david-lyle has quit IRC04:37
mahaticgood morning04:52
*** chenhuayi has quit IRC04:53
*** pcaruana has quit IRC04:55
mahaticnotmyname: helpful info on rolling upgrades, thanks! one more question, once we do Swift (standalone) rolling upgrades test, can we be assured of the same results for a swift rolling upgrades test within an OpenStack environment?05:03
mahaticOr would we need to do check a few more things if it's in an OpenStack environment?05:03
mahaticneed to check*05:03
notmynamemahatic: well, yes and no05:03
notmynameultimately, yes. because a prod environment will have a lot of pieces and you gotta make sure they're all good05:04
notmynamebut also "no" in that, if all the pieces are holding to their API and staying self-contained, then it *should* work05:04
notmynameof course "should" is why we test05:05
notmynamebut what I mean is that if you're monitoring with ceilometer and doing auth with keystone, then it's important to make sure they are upgraded too05:05
mahaticyou mean we gotta make sure explicitly it works in an openstack environment?05:05
notmynamebut what we're responsible for as the swift team is making sure our upgrades work and things still work together05:06
notmynameso I guess the answer depends on what you're trying to do05:06
notmynameif it's something around prod deployments, then test everything. if it's about openstack gating, then test just swift05:06
notmyname(i think)05:06
mahaticyes production deployments05:07
mahaticnotmyname: thanks! that is helpful05:12
notmynamethen yeah, checking that eg keystone and swift still work together is important05:12
notmynamebut that's already covered in the gate, I think05:12
mahaticokay05:13
notmynamethe grenade tests, IIRC, test that swift and keystone still work together (and other projects. it's the full "test all of devstack" suite)05:14
*** jamielennox is now known as jamielennox|away05:18
mahaticnotmyname: yeah I thought so too, but we were wondering if that is updated enough (for a few projects it is not)05:20
mahaticthat being grenade05:22
*** aswadr_ has joined #openstack-swift05:22
*** links has joined #openstack-swift05:38
*** links has quit IRC05:38
*** dabukalam has quit IRC05:39
*** dabukalam has joined #openstack-swift05:39
*** nadeem has joined #openstack-swift05:48
*** aagrawal has quit IRC05:50
*** silor has joined #openstack-swift06:07
*** silor1 has joined #openstack-swift06:14
*** silor has quit IRC06:16
*** silor1 is now known as silor06:16
*** geaaru has joined #openstack-swift06:25
*** aagrawal has joined #openstack-swift06:37
*** aagrawal has quit IRC06:43
*** aagrawal has joined #openstack-swift06:45
*** abhinavtechie has joined #openstack-swift06:48
*** aagrawal has quit IRC06:49
*** rcernin has joined #openstack-swift06:51
*** pcaruana has joined #openstack-swift06:52
*** dfflanders has joined #openstack-swift06:54
*** klrmn has quit IRC06:58
*** abhinavtechie has quit IRC06:58
*** abhinavtechie has joined #openstack-swift07:01
*** tesseract- has joined #openstack-swift07:09
*** rcernin has quit IRC07:25
*** cbartz has joined #openstack-swift07:25
*** joeljwright has joined #openstack-swift07:42
*** ChanServ sets mode: +v joeljwright07:42
*** asettle has joined #openstack-swift08:05
*** geaaru has quit IRC08:06
openstackgerritJoel Wright proposed openstack/python-swiftclient: Add copy object method  https://review.openstack.org/28020008:06
joeljwrightmorning08:07
joeljwright:)08:07
*** geaaru has joined #openstack-swift08:18
*** asettle has quit IRC08:27
*** asettle has joined #openstack-swift08:35
*** nadeem has quit IRC08:53
*** takashi has joined #openstack-swift08:57
*** tqtran has joined #openstack-swift09:06
*** tqtran has quit IRC09:11
*** sheel has joined #openstack-swift09:22
*** klrmn has joined #openstack-swift09:24
*** hosanai has quit IRC09:48
*** haypo has joined #openstack-swift09:57
openstackgerritVictor Stinner proposed openstack/swift: py3: tox.ini: use substituation to py35 commands  https://review.openstack.org/35852710:00
openstackgerritVictor Stinner proposed openstack/swift: py3: Port test_splice to Python 3  https://review.openstack.org/35852810:00
openstackgerritVictor Stinner proposed openstack/swift: py3: port memcached client to Python 3  https://review.openstack.org/35852910:00
*** dfflanders has quit IRC10:25
*** openstackgerrit has quit IRC11:03
*** openstackgerrit has joined #openstack-swift11:04
*** tqtran has joined #openstack-swift11:08
*** tqtran has quit IRC11:13
*** maestropandy has joined #openstack-swift11:28
*** openstackstatus has quit IRC11:36
*** openstackstatus has joined #openstack-swift11:39
*** ChanServ sets mode: +v openstackstatus11:39
*** asettle has quit IRC11:46
*** maestropandy has left #openstack-swift11:53
*** takashi has quit IRC11:57
*** asettle has joined #openstack-swift11:57
*** maestropandy1 has joined #openstack-swift12:03
*** maestropandy1 has left #openstack-swift12:04
*** Jeffrey4l_ has joined #openstack-swift12:15
*** Jeffrey4l has quit IRC12:16
*** asettle has quit IRC12:19
*** asettle has joined #openstack-swift12:29
*** klrmn has quit IRC12:47
*** abhinavtechie has quit IRC12:51
*** joeljwright1 has joined #openstack-swift12:55
*** kei_yama has quit IRC12:57
*** joeljwright1 has quit IRC13:07
*** bigdogstl has joined #openstack-swift13:48
*** briancurtin has quit IRC13:51
*** briancurtin has joined #openstack-swift13:52
*** bigdogstl has quit IRC13:53
*** hurricanerix has quit IRC14:10
*** tqtran has joined #openstack-swift14:11
openstackgerritDonagh McCabe proposed openstack/swift: Corrections for the API specifications in api-ref  https://review.openstack.org/35476714:14
*** tqtran has quit IRC14:15
*** brnelson has joined #openstack-swift14:15
*** silor has quit IRC14:17
openstackgerritDonagh McCabe proposed openstack/swift: Corrections for the API specifications in api-ref  https://review.openstack.org/35476714:18
*** sheel has quit IRC14:26
*** hurricanerix has joined #openstack-swift14:29
*** bigdogstl has joined #openstack-swift14:33
*** cdelatte has joined #openstack-swift14:36
*** bigdogstl has quit IRC14:43
*** zaitcev has joined #openstack-swift14:59
*** ChanServ sets mode: +v zaitcev14:59
*** silor has joined #openstack-swift15:00
*** silor has quit IRC15:02
*** silor has joined #openstack-swift15:03
notmynamegood morning15:05
*** silor1 has joined #openstack-swift15:07
*** asettle has quit IRC15:08
*** silor has quit IRC15:08
*** silor1 is now known as silor15:08
notmynamethis week I'll be traveling to NYC. tomorrow and wednesday is OpenStack East, and on Thurs-Fri is the ops midcycle meetup15:09
notmynameI'll be on-and-off in IRC, but ping me if I can do anything for you and I'll get back to you asap15:10
notmynamejoeljwright: in the midst of travel this week, I'm planning on a serious look at swiftclient for a release15:10
joeljwrightnotmyname: timburke and I have been trying to get this copy patch into shape15:11
notmynameyep, I saw that :-)15:11
joeljwrightI'll speak to him later about any other pacthes we need to land15:11
joeljwrightenjoy NYC15:11
*** abhinavtechie has joined #openstack-swift15:12
*** Jeffrey4l_ has quit IRC15:13
*** bigdogstl has joined #openstack-swift15:14
*** cbartz has quit IRC15:18
*** klrmn has joined #openstack-swift15:22
*** Jeffrey4l_ has joined #openstack-swift15:22
*** cdelatte has quit IRC15:30
*** bigdogstl has quit IRC15:32
*** bigdogstl has joined #openstack-swift15:33
*** Administrator__ has quit IRC15:36
*** Administrator__ has joined #openstack-swift15:37
*** asettle has joined #openstack-swift15:39
*** bigdogstl has quit IRC15:48
*** manous has joined #openstack-swift15:48
*** bigdogstl has joined #openstack-swift15:56
*** Jeffrey4l_ has quit IRC16:00
*** Raymii has quit IRC16:02
*** Raymii has joined #openstack-swift16:02
*** airdisa has joined #openstack-swift16:03
*** furlongm has joined #openstack-swift16:04
*** furlongm_ has quit IRC16:06
*** klrmn has quit IRC16:07
*** asettle has quit IRC16:11
*** bigdogstl has quit IRC16:12
timburkeeither it was eerily quiet all weekend, or my bouncer didn't do its job very well :(16:13
*** Jeffrey4l_ has joined #openstack-swift16:14
*** bigdogstl has joined #openstack-swift16:14
*** silor has quit IRC16:14
*** asettle has joined #openstack-swift16:23
*** tesseract- has quit IRC16:28
*** asettle has quit IRC16:28
*** pcaruana has quit IRC16:29
*** Jeffrey4l_ has quit IRC16:33
*** bigdogstl has quit IRC16:36
*** dmorita has joined #openstack-swift16:45
*** abhinavtechie has quit IRC16:48
*** manous has quit IRC16:49
*** nadeem has joined #openstack-swift16:51
*** manous has joined #openstack-swift17:06
*** SkyRocknRoll has joined #openstack-swift17:10
*** bigdogstl has joined #openstack-swift17:10
*** tqtran has joined #openstack-swift17:13
*** bigdogstl has quit IRC17:16
*** tqtran has quit IRC17:17
*** Suyi has joined #openstack-swift17:20
joeljwrighttimburke: cheers for the comments17:24
*** asettle has joined #openstack-swift17:25
joeljwrighttimburke: I did make a version of the copy patch that required container and object to be set, but I felt like I was changing things too much. Your comments have made me reconsider - I'll throw up another updated version later.17:25
joeljwrighttimburke: and I'll keep adding tests as I go :D17:26
timburkejoeljwright: cool, thanks. that just seems like a better api to me17:26
joeljwrightyeah, me too tbh17:26
joeljwrightI had the patch all ready17:26
timburkemakes it easier :-)17:26
joeljwrightI just decided against it when I was checking out teh other methods that follow teh patterm Marek used17:27
joeljwrightit will certainly clean up that horrible URL parsing logic17:27
joeljwrightI was thinking I wanted you to dissuade me from that approach while I was writing it :)17:27
joeljwrightyou should have a new version to look at in 3 or 4 hours time17:28
timburkehappy to be a source of dissuasion :P17:28
joeljwright:D17:29
*** joeljwright has quit IRC17:30
*** asettle has quit IRC17:31
*** tqtran has joined #openstack-swift17:35
*** zul has quit IRC17:40
*** zul has joined #openstack-swift17:48
*** manous has quit IRC18:03
*** vint_bra has joined #openstack-swift18:13
*** geaaru has quit IRC18:14
*** manous has joined #openstack-swift18:16
openstackgerritThai Tran proposed openstack/swift: WIP Investigating Account ACL  https://review.openstack.org/35671518:16
tqtranhello, if i wanted to add tests to the patch above, where is the best place to put them? seems like we have tests for ACL but they only cover general http cases, not whether a certain user actually has access based on their ACL or not.18:19
*** lcurtis has joined #openstack-swift18:19
tqtranis this the best place? https://github.com/openstack/swift/blob/master/test/functional/test_access_control.py18:20
claygtqtran: idk, there's some stuff in functional/test_(account|container).py too?18:43
*** manous has quit IRC18:49
tqtranclayg: thanks for the pointer, i'll take a look.18:51
*** lcurtis has quit IRC18:52
*** manous has joined #openstack-swift19:01
openstackgerritMohit Motiani proposed openstack/swift: Add regions and zones in ring-builder.py  https://review.openstack.org/35062519:08
*** nadeem has quit IRC19:23
*** nadeem has joined #openstack-swift19:24
*** SkyRocknRoll has quit IRC19:24
openstackgerritMohit Motiani proposed openstack/swift: Fix typos and grammer in builder.py  https://review.openstack.org/35882519:30
*** klrmn has joined #openstack-swift19:30
*** Glourflump has joined #openstack-swift19:34
*** Glourflump has left #openstack-swift19:39
*** rcernin has joined #openstack-swift19:41
openstackgerritClay Gerrard proposed openstack/swift: Fix typos and grammer in builder.py  https://review.openstack.org/35882519:46
*** manous has quit IRC19:55
*** asettle has joined #openstack-swift19:59
*** asettle has quit IRC19:59
*** manous has joined #openstack-swift20:08
*** tqtran has quit IRC20:09
openstackgerritVictor Stinner proposed openstack/swift: py3: port memcached client to Python 3  https://review.openstack.org/35852920:11
*** tqtran has joined #openstack-swift20:13
zaitcevDoes anyone else see Hummingbird mysteriously refusing to log (to syslog or anywhere else)?20:18
claygthat sounds vaugly familiar20:18
zaitcevIf you didn't positively identify and fix it, I'll have a look then20:19
claygzaitcev: I had to do this one time -> https://github.com/swiftstack/vagrant-swift-all-in-one/commit/97dfbcb1484792fbf375ba83886e7e233ff530e820:19
zaitcevoooh20:20
timburketorgomatic: i've been thinking a bit about bug 1598093 today; looks like it affects SLOs, too. what's you opinion on whether SegmentedIterable should expect quoted segment paths vs. doing the quoting itself? i figured your work that led to patch 352657 might be a nice third data-point21:07
openstackbug 1598093 in OpenStack Object Storage (swift) "Dynamic Large Object does not work with "?" in prefix" [Undecided,New] https://launchpad.net/bugs/159809321:07
timburkeah, looks like patchbot isn't here. https://review.openstack.org/#/c/352657/ - the multi-range GET for SLO guy21:08
torgomatictimburke: SegmentedIterable is just gonna take whatever it's given and jam it into PATH_INFO in a subrequest, and I think WSGI says that PATH_INFO is unquoted21:10
torgomaticso I'd expect unquoted things; after all, we're talking to the proxy server via function calls, not network requests, so we ideally don't have any escaping going on21:10
claygisn't there a unix-ism about doing decode/encode close to wire/disk?21:12
claygi.e. "Agree with torgomatic - it's almost always the right call"21:13
torgomatictimburke: heh, this is lame... SegmentedIterable sticks the query string onto the path so it can call make_subrequest(), which splits the path on ? to get the query string back out :|21:13
timburketorgomatic: ah! there's the stupid. thanks21:14
*** joeljwright has joined #openstack-swift21:15
*** ChanServ sets mode: +v joeljwright21:15
timburketorgomatic: although...which seems less bad? let make_subrequest grow a query_string arg (which, if present, should maybe prevent us from unquoting path), or just have SegmentedIterable pass in quoted paths as make_subrequest has documented? seems like there's a lot of quoting/unquoting pretty much no matter what :-/21:18
torgomatictimburke: so I guess in this case, it's a failing of the DLO middleware to pass that thing into make_subrequest in the right way for it to end up in just PATH_INFO21:18
torgomatictimburke: I like having make_subrequest() take a query string. I just looked at all the call sites, and they either (a) don't even mention query strings, or (b) stick a hardcoded query string onto the path21:19
mattoliverauMorning21:30
zaitcevGius, who's our encription person? jrichli?21:35
mattoliverauzaitcev: yup21:36
jrichliyo21:36
jrichlii gotta catchup with scrollback21:37
claygzaitcev: or acoles_ timburke or tdasilva - you can use me in a pinch - i'm pretty sure I understand the swift parts anyway21:37
zaitcevjrichli: https://github.com/openstack/swift/blame/master/doc/source/overview_encryption.rst#L72 says "encryption" in the pipeline, but isn't it encrypter and decrypter now?21:37
jrichlithat is the correct thing to add to the pipeline.  that was one of the later changes21:38
clayglook at the docs being all correct and shiz ;)21:39
jrichlithe middleware files are still called decrypter and encrypter, but you use encrtyption in the pipeline21:39
zaitcev"that" being what's documented? etc/proxy-server.conf-sample too?21:39
zaitcevjrichli: thanks21:39
torgomatictimburke: check out https://gist.github.com/smerritt/ca5a6701d46c55b43059e4d0e302c2f9 for a first pass at cleaning up the query string crud21:47
torgomaticI'm suspicious that it might fix bug 1598093 as wel, but I haven't tried it.21:47
openstackbug 1598093 in OpenStack Object Storage (swift) "Dynamic Large Object does not work with "?" in prefix" [Undecided,Confirmed] https://launchpad.net/bugs/1598093 - Assigned to Tim Burke (1-tim-z)21:47
timburketorgomatic: thanks!21:51
claygtorgomatic: timburke: so it's a good idea that make_subrequest and make_env have different ideas about the quotedness of path?21:57
torgomaticclayg: it's like a balloon filled with stupid... I push down over here, but it just pops up more over there21:58
torgomatic(but no, that's confusing and will lead to people getting it wrong)21:59
claygtorgomatic: I think blank would get confused/mad if you give if quoted path's btw21:59
claygfrom docstring on swob.Request.blank ->         :param path: encoded, parsed, and unquoted into PATH_INFO21:59
claygi don't really know what it means by "parsed" - but I'm pretty sure encoded and unquoted both mean things to me21:59
claygand one of them matches with how I sorta think make_env works - but I'm not sure what the WSGI spec says about encoding - probably that it's assumed some sort of encoded byte string (utf8?) because nothing else makes sense in python2?22:01
timburkeclayg: apparently it means https://github.com/openstack/swift/blob/2.9.0/swift/common/swob.py#L832 -- so i guess we can even say something like 'http://saio:8080/v1/...' !?22:02
claygtimburke: that seems like a weird convience because we use blank mostly in tests?22:03
timburkeclayg: i'm fairly certain wsgi assumes all non-body bytes are latin-1 :-/22:04
timburkewhich led us to https://review.openstack.org/#/c/348322/5/swift/common/header_key_dict.py22:04
torgomaticclayg: Request.blank() takes quoted paths, so "/with%spaces" turns into PATH_INFO: "/with spaces"22:05
claygoh dear goodness22:05
torgomaticer, that's %20, but you know what i mean22:06
claygtorgomatic: lol - you're right!  when the doc string says "unquoted" - it means "totally fucking quoted because https://github.com/openstack/swift/blob/master/swift/common/swob.py#L847"22:07
timburkeclayg: 'cause describing what you're going to do to the argument is a sane way to write docstrings...22:08
*** bigdogst_ has joined #openstack-swift22:10
claygtimburke: oh rofl - sorry yeah I read it wrong22:13
clayg"this argument will be ..." - oh man I sure hope I wrote that22:13
timburke*totally* sane. why wouldn't you write it that way?22:14
clayg27c7abfb - YES!22:14
timburkelol22:14
*** bigdogst_ has quit IRC22:14
*** jamielennox|away is now known as jamielennox22:15
claygi guess the weird thing about most strings is that unquote tends to converge eventually - so you can mostly unqote unquoted strings and it pretty rarely does the wrong thing for you22:17
clayg"did you really *want* "%20" - better quote that noise"22:18
timburkeyup. only a problem if you want a literal %xx in there, which tends to be unlikely22:19
claygwild22:19
*** vint_bra has quit IRC22:33
openstackgerritJoel Wright proposed openstack/python-swiftclient: Add copy object method  https://review.openstack.org/28020022:34
*** joeljwright has quit IRC22:40
*** tqtran has quit IRC22:41
claygwow patch 280200 is huge!  next time we try to do this we should remember to start with just the client stuff - then the service stuff - then the shell stuff - the scope of that change is a beast!22:41
*** tqtran has joined #openstack-swift22:44
claygtimburke: where's patchbot?  is this something notmyname normally fixes?22:44
timburkeclayg: yeah; i think it's the same reason that https://not.mn/swift/swift_community_dashboard.html is currently down22:45
claygnotmyname: we're hopelessly lost without you - please hear our woes22:46
* clayg expects notmyname in an airplane now -> http://www.aidthoughts.org/wp-content/uploads/2013/11/obiwan.png22:47
clayga disturbance in the #openstack-swift22:48
klrmntimburke: clayg: he's moved the dashboard to http://d.not.mn/swift_community_dashboard.html22:48
*** nadeem has quit IRC22:52
mattoliveraunotmyname's server has issues over the weekend22:55
mattoliverauhe does have http://d.not.mn/swift_community_dashboard.html for the dashboard22:55
mattoliverauklrmn beat me too it :P22:56
*** asettle has joined #openstack-swift23:02
*** asettle has quit IRC23:10
*** kei_yama has joined #openstack-swift23:23
*** airdisa has quit IRC23:31

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