Tuesday, 2015-04-14

openstackgerritDean Troyer proposed openstack/python-openstackclient: Defer client imports  https://review.openstack.org/17309800:35
*** stevemar has joined #openstack-sdks02:40
*** stevemar has quit IRC03:06
*** LinuxHorn has quit IRC03:06
*** jacekn has quit IRC03:06
*** stevemar has joined #openstack-sdks03:13
*** LinuxHorn has joined #openstack-sdks03:13
*** jacekn has joined #openstack-sdks03:13
*** britthouser has joined #openstack-sdks03:19
*** britthou_ has joined #openstack-sdks03:21
*** britthouser has quit IRC03:24
openstackgerritDean Troyer proposed openstack/python-openstackclient: Defer client imports  https://review.openstack.org/17309803:50
*** britthou_ has quit IRC04:22
*** britthouser has joined #openstack-sdks04:23
*** terrylhowe has quit IRC04:43
*** maufart has joined #openstack-sdks04:58
*** maufart has quit IRC05:06
*** maufart has joined #openstack-sdks06:47
*** maufart has quit IRC06:53
*** jamielennox is now known as jamielennox|away07:02
*** stevemar has quit IRC07:09
*** chlong has quit IRC07:27
*** terrylhowe has joined #openstack-sdks09:08
*** trown|outttypeww is now known as trown11:53
*** sdague has joined #openstack-sdks13:11
sdagueok, openstack client questions, does anyone have a working version of openstack security group rule add ?13:12
sdague+ openstack security group rule create nova_grenade --proto icmp13:13
sdagueERROR: openstack 'NoneType' object is not iterable13:13
sdaguebut the help tells me that's a completely valid gramar for that command13:14
dtroyer_zzsdague: that's fixed in master13:30
dtroyer_zzchange-id: I017ffd424ca25b6c62193b91068eb1ba7de6c91913:30
dtroyer_zzwait…wrong one13:30
dtroyer_zztoo early13:31
*** britthouser has quit IRC13:32
sdaguedtroyer_zz: so, not usable in grenade because we're at 1.0.2 (juno pinned version)13:33
dtroyer_zzhey, this should actually sound totally familiar…it's the null description bug13:33
dtroyer_zzadd a description13:33
openstackLaunchpad bug 1434172 in python-openstackclient "security group create errors without description" [Low,In progress] - Assigned to Dean Troyer (dtroyer)13:33
dtroyer_zzI haven't worked around it inOSC yet13:33
sdagueyeh I hit 1434172 as well13:34
sdaguebut it's not the issue I'm looking at right now, which is the *rule* create13:34
sdagueI was trying to make these resource scripts in grenade entirely use openstack client, however, that looks like it's not really an option with the amount of issues I'm running into13:36
dtroyer_zzicmp, yeah, that should work, add a —dst-port 0 to work around it13:37
*** sigmavirus24_awa is now known as sigmavirus2413:37
sdagueoh, also, openstack server create doesn't seem to work13:38
sdagueopenstack server create --image cirros-0.3.2-x86_64-uec --flavor m1.tiny nova_server113:38
sdagueERROR: openstack Invalid volume client version '2'. must be one of: 113:38
sdaguemost infuriating because I'm not using any volume commands13:39
dtroyer_zzexport OS_VOLUME_API_VERION=113:39
dtroyer_zzI forget why that started13:40
sdague+ export OS_VOLUME_API_VERION=113:40
sdague+ OS_VOLUME_API_VERION=113:40
sdague+ openstack server create --image cirros-0.3.2-x86_64-uec --flavor m1.tiny nova_server113:40
sdagueERROR: openstack Invalid volume client version '2'. must be one of: 113:40
sdagueoh, there was a typo13:40
sdagueok, yes, that works, but that seems ... suboptimal to make people do that for server create to work13:41
sdagueis there a bug out there for that?13:41
dtroyer_zzI thought there was but can't find it so no13:42
terrylhoweI would think your delayed import changes should solve that volume issue dtroyer_zz13:43
terrylhoweat least if you aren’t using cinder13:43
sdagueyeh, would help, but still doesn't address the fact that if you were trying to use volumes, it doesn't work out of the box13:47
dtroyer_zzterrylhowe: I'm not sure about that, we only import cinderclient.v1.* stuff and cinderclient.exceptions13:47
dtroyer_zzI don't think I ever chansed down exactly what changed to make that start happening13:47
terrylhoweI haven’t seen the issue I’m having a tough time thinking how that might even happen13:49
dtroyer_zzI've seen it, but IIRC only in DevStack installs.13:49
dtroyer_zzsomeone needs to kick those guys13:49
sdaguedtroyer_zz: ah...13:51
sdagueos1:~/devstack(master)> git grep CINDER_VERSION13:51
sdagueopenrc:export CINDER_VERSION=${CINDER_VERSION:-2}13:51
sdagueso it's been embedded in the openrc13:51
sdaguewe should probably get rid of that13:52
sdaguebecause I'm definitely sourcing openrc to bootstrap this script13:52
dtroyer_zzno ownder I never found it in osc… ;)13:53
terrylhoweit’d be nice if you could do security group rules on a nonadmin url13:53
dtroyer_zznove doesn't have admin-specific endpoints13:54
dtroyer_zzand I still can't type before 9:00am13:54
terrylhowehmm, I was getting a an error that it couldn’t find an admin url for that operation13:54
sdagueit's honestly kind of weird that keystone does honestly14:00
sdaguewhat dictates the order of sub commands in lists?14:02
sdaguehttp://paste.openstack.org/show/203873/ - seems really unhelpful in order14:02
dtroyer_zzwe're alpha-sorted everywhere in the source, that's probably an unordered dict dump.14:03
openstackgerritMerged stackforge/python-openstacksdk: Determine limit based on page size  https://review.openstack.org/15595414:03
*** britthouser has joined #openstack-sdks14:03
sdagueis there a budding man page anywhere for osc?14:05
dtroyer_zzthat we do have!14:05
sdaguenot installed by default?14:05
dtroyer_zzteach me how and it will be14:06
sdaguealso, is there an equiv of --poll on 'server create' ?14:06
dtroyer_zz—wait should be our common option for that14:06
* sdague tries14:06
dtroyer_zzhere is the published man page: http://docs.openstack.org/developer/python-openstackclient/man/openstack.html14:07
dtroyer_zzit only covers the global options, commands are separate14:08
dtroyer_zzfwiw, that help output is produced by cliff in cliff.help:HelpCommand.take_action() and should be sorted14:08
dtroyer_zzwhere 'shoud be' == 'needs to be change to be'14:09
sdagueit would be nice if the commands were included14:10
sdagueI wonder if the man page question is one for mordred14:10
dtroyer_zzya know how huge that would be?14:10
sdaguedtroyer_zz: I have a pager that supports search14:10
sdaguersync man page is 2500 lines14:11
sdaguebash man page is 3500 lines14:11
dtroyer_zzwe could probably cat together the detail from all of these… http://docs.openstack.org/developer/python-openstackclient/command-list.html14:11
dtroyer_zzI actually hate how big those are14:11
dtroyer_zzsearch for the set command in bash man page and get back to me14:12
sdaguebecause, honestly, if it's a manpage I'm now outside of hyperlink realm (also, hyperlinked sub documents only make sense if you *know* where it is)14:12
dtroyer_zzyou sound like an emacs-using info guy14:12
sdague /^\s+ set14:13
* dtroyer_zz ducks14:13
sdaguehit n... 8 times?14:13
sdagueno, this is the *anti* info approach14:13
sdagueinfo pages are basically what you have, and I kind of hate14:13
dtroyer_zzso at least we agree on that14:14
sdaguesure :)14:15
sdagueit's terrible for discovery14:15
sdaguebecause you actually need to know enough about the thing to find the thing14:15
sdagueone giant man page means you know where to look, and can search until you find something related14:16
sdaguesmb.conf - 6500 lines :)14:17
sdaguetrying to find bigger ones14:17
sdaguethough I knew that one was a monster14:17
terrylhoweI was all in the cliff help this morning, so sdague https://review.openstack.org/#/c/173341/14:18
dtroyer_zzI'll lean on git as the counter-example… although it can be maddening finding things there if you don't know where to look, proving your point14:18
sdagueterrylhowe: thanks14:19
sdaguedtroyer_zz: right, git man pages are completely useless to learn git14:19
dtroyer_zzbut htat's a discovery problem, after that I think the size does get in the way14:19
sdagueyou have to know git, and they are reasonable to reference advanced options14:20
sdagueI'm not convinced size is an issue14:20
sdaguevery few people read man pages cover to cover14:20
sdaguethe moment you are no longer reading the whole document, but are just searching to some point in it, size is kind of irrelevant (as long as it doesn't crash less)14:21
*** mattfarina has joined #openstack-sdks14:21
sdaguedtroyer_zz: maybe it's just me, but I can't remember the last time I typed "man xxxxx" and the first thing I did next was anything other that '/'14:22
sdagueunrelated, the fact that openstack goes to some shell like thing with no options, does that mean you can #!/usr/local/bin/openstack ?14:23
dtroyer_zzI haven't tried that, but it might work.  It does mean you can redirect into it as in https://review.openstack.org/#/c/164302/14:25
dtroyer_zzbut that has other problems in devstack's use case14:25
dtroyer_zzit saved about 20 seconds when I timed that patch14:25
sdagueis there a way to do var capture and basic conditionals?14:25
dtroyer_zznope, that's into DSL territory and while I'd love to do that we don't14:26
dtroyer_zzmaybe I'll embed Lua14:26
openstackgerritMonty Taylor proposed openstack/os-client-config: Reset cache default to 0  https://review.openstack.org/17334514:26
sdagueok, mordred must now be awake, I really want to figure out the pbr man page thing14:27
sdagueI wonder how many times we have to say mordred to make him appear :)14:27
dtroyer_zzusually not quite three14:27
sdaguethis makes me a little sad as the way -f shell works for server create14:28
mordredsdague: aroo?14:28
mordred"pbr man page thing?14:28
sdaguemordred: how does one install man pages from a python package correctly14:29
sdagueassuming that pbr might be implicated14:29
dtroyer_zzsdague: use -f value instead14:29
dtroyer_zzand -c id14:29
dtroyer_zzreturn exactly one value14:29
mordredsdague: well, you can make them from sphinx14:29
dtroyer_zz-f shell isn't as useful as I had hoped it would be14:29
*** stevemar has joined #openstack-sdks14:30
mordredunless you want to do troff directly14:30
sdaguemordred: right, making the content isn't the issue14:30
sdagueit's pip install python-openstackclient14:30
sdagueactually *installing* man pages correctly14:30
mordredyeah - pbr does that14:30
sdaguemordred: example?14:31
mordredsdague: git-review's setup.cfg installs a manpage14:32
mordredmanpages =14:32
mordred    git-review.114:32
sdaguemordred: nice14:32
mordredmanpages =14:32
mordred    git-review.114:32
mordredhowever- I think there is more magic available14:32
mordrednope. no magic - just that14:33
sdaguemordred: ok, cool, I'll figure out how to add to python-openstackclient14:33
sdagueso, in the git-review case this is already troff14:34
sdagueif it's getting built by sphinx, how do you specify the reference14:35
mordredyeah - that's a bit more difficult, because we don't normally do a sphinx run as part of an install step14:36
mordredI think it would require install becoming "python setup.py build_sphinx install" - except taht's ugly14:37
mordredwe may want to add a feature to pbr14:37
sdagueok, yeh, installing man pages from source would definitely be handy14:37
*** pm90_ has joined #openstack-sdks14:42
*** bknudson has joined #openstack-sdks14:43
sdagueit makes me slightly sad pants that -c addresses,id doesn't work14:47
*** Vultr-Dave has left #openstack-sdks14:48
*** stevemar has quit IRC14:49
dtroyer_zzsdague: try: -c Addresses -c ID14:49
dtroyer_zzmatching case is important, although it shouldn't be14:49
*** stevemar has joined #openstack-sdks14:49
dtroyer_zzand yeah, that should probably handle a list too14:49
dtroyer_zzbut in general, in OSC specifying multiple of anything uses multiple options14:50
sdaguedtroyer_zz: yeh, multiple -c works, which I figured out14:50
sdagueyeh, it's unfortunate, because while lots of things are a huge step forward in UX, things like security group rules feel like a regression because the possitional parameters make a little more sense14:51
*** etoews has joined #openstack-sdks14:51
openstackgerritMonty Taylor proposed openstack/os-client-config: Reset cache default to 0  https://review.openstack.org/17334514:53
*** pm90_ has quit IRC15:01
*** pm90_ has joined #openstack-sdks15:05
openstackgerritMerged openstack/os-client-config: Reset cache default to 0  https://review.openstack.org/17334515:10
*** pm90__ has joined #openstack-sdks15:21
openstackgerritMonty Taylor proposed openstack/os-client-config: Document vendor support information  https://review.openstack.org/17338315:24
*** pm90_ has quit IRC15:24
openstackgerritMerged stackforge/python-openstacksdk: Rework update_attrs so dirty list always updated  https://review.openstack.org/16945115:28
openstackgerritDean Troyer proposed openstack/python-openstackclient: Defer client imports  https://review.openstack.org/17309815:40
*** pm90__ has quit IRC15:47
mordredterrylhowe: thansk! nice cach15:48
openstackgerritMonty Taylor proposed openstack/os-client-config: Document vendor support information  https://review.openstack.org/17338315:48
*** stevemar has quit IRC15:49
*** stevemar has joined #openstack-sdks15:49
openstackgerritTerry Howe proposed stackforge/python-openstacksdk: Create stack  https://review.openstack.org/16406115:51
*** pm90_ has joined #openstack-sdks15:59
openstackgerritDean Troyer proposed openstack/python-openstackclient: Defer client imports  https://review.openstack.org/17309815:59
stevemardtroyer_zz, you sent me down the rabbit hole16:05
stevemarthe biggest of all rabbit holes16:05
* sigmavirus24 gets popcorn16:06
* dtroyer_zz reached for funny top hat16:06
stevemardtroyer_zz, https://launchpad.net/bugs/1443089 came in, and i started to look at our option parsing and glanceclients lib16:06
openstackLaunchpad bug 1443089 in python-openstackclient "image list displays a maximum of 25 images" [High,Confirmed]16:06
dtroyer_zzsomehow I knew the word 'image' would be part of this explanation16:07
stevemarthen after bugging everyone in -glance, and insisting that their client libs are broken...16:07
stevemarsigmavirus24 realizes that we are using our own API16:07
stevemarso it's our own fault, yay16:08
* dtroyer_zz bows16:08
* dtroyer_zz then crawls out of room16:08
sigmavirus24stevemar: you and I did it at the same time16:09
stevemarso i guess if we want to actually list all the images available, then we need to pass in limit=None ? or something to that... sigmavirus24 ?16:09
sigmavirus24dtroyer_zz: want some popcorn?16:09
sigmavirus24stevemar: I don't think you need to pass anything special16:09
terrylhowe25 is such a small page size too16:09
sigmavirus24I think it will "Just Work"™16:09
sigmavirus24If you send page_size=50 or whatever that'll work16:09
sigmavirus2450,100, whatever16:10
* sigmavirus24 wishes he'd have the time this week to work on fixing this16:10
stevemarsigmavirus24, but like you said, arbitrary amounts won't work16:10
sigmavirus24stevemar: that's page_size which only affects how many you get back per-request16:11
sigmavirus24that way you're not making 4 requests for those 103 images16:11
sigmavirus24100 isn't a bad number to pick16:11
dtroyer_zzsince we own the api on this we can do whatever we need to make it work, is it a server-side 'feature' we are working around?16:11
dtroyer_zzcan we not get 'all'?16:11
sigmavirus24If you want 200 that's cool too but you have to consider the stress it' may put on the server16:11
sigmavirus24dtroyer_zz: the server will list "all" if you specify a page size large enough I think16:12
stevemardtroyer_zz ++, getting all would be nice16:12
sigmavirus24I think the server has an arbitrary of max of 1000 though (as someone just mentioned)16:12
dtroyer_zzI did expect sanity limits like that16:12
dtroyer_zzso no matter what we'll need to step through getting the whole list.  it'll look a _LOT_ like the object_store list implementation16:13
sigmavirus24It may or may not be because glance hasn't focused on stability/performance in a few cycles16:13
stevemarsigmavirus24, take the keystone route and just list everything :P16:13
sigmavirus24feature feature feature bug16:13
sigmavirus24in which case I really wish I had one of the RAX tees from PyCon16:14
stevemarthat was a cool tee16:14
dtroyer_zzI gotta check out this 'PyCon' thing sometime…16:14
sigmavirus24best part of never thinking to make my own business cards - briancurtin gets all the emails16:14
stevemardtroyer_zz, would it make sense to allow the user to specify a limit ? like: https://review.openstack.org/#/c/172763/16:15
stevemarand default to 100 or so if it's not set16:15
stevemarinstead of 2516:15
stevemarthis all seems hacked together, i like the keystone approach of using filters instead of pagination  it definitely simplifies the client side16:17
stevemaror maybe i'm just hangry16:18
dtroyer_zzI think it is perfectly fine to have an option for the user to set the limit.  I'd really prefer though to not have one by default.16:20
dtroyer_zzin the keystone LDAP case, I totally understand why 'all' is bad.  but not here16:20
*** pm90_ has quit IRC16:39
openstackgerritEverett Toews proposed stackforge/python-openstacksdk: The first functional test  https://review.openstack.org/17034317:02
*** pm90_ has joined #openstack-sdks17:03
openstackgerritTerry Howe proposed openstack/python-openstackclient: Handle the pagination for image list  https://review.openstack.org/17342017:04
*** pm90_ has quit IRC17:04
*** pm90_ has joined #openstack-sdks17:05
*** trown is now known as trown|lunch17:26
*** sigmavirus24 is now known as sigmavirus24_awa17:30
openstackgerritTerry Howe proposed stackforge/python-openstacksdk: Add the IPv6 subnet attributes  https://review.openstack.org/16875417:31
openstackgerritTerry Howe proposed stackforge/python-openstacksdk: Add the IPv6 subnet attributes  https://review.openstack.org/16875417:32
*** boris-42 has quit IRC17:38
*** chellygel has joined #openstack-sdks18:02
chellygelHey everyone. I asked a question on launchpad today in regards to cliff and how to format the responses. If you get a chance, please take a look and help me out, i'd greatly appreciate it!18:03
chellygelim still a newbie at this whole thing and may be missing something entirely.18:04
*** jamielennox|away is now known as jamielennox18:05
dhellmannchellygel: I think you want the "value" formatter18:07
chellygeldhellmann, thank you! i didnt see that in the documentation!! you have saved me so much work18:15
chellygelill be happy to submit a pr to add some info! :D18:16
*** sigmavirus24_awa is now known as sigmavirus2418:16
openstackgerritMerged openstack/os-client-config: Document vendor support information  https://review.openstack.org/17338318:21
*** tellesnobrega has quit IRC18:25
*** tellesnobrega has joined #openstack-sdks18:25
*** tellesnobrega_ has joined #openstack-sdks18:27
*** tellesnobrega_ has quit IRC18:27
*** tellesnobrega has quit IRC18:29
*** trown|lunch is now known as trown18:32
briancurtinSDK meeting in 15 min in #openstack-meeting-3 -- with terry having been out on vacation and then i had been at pycon (still am), things have been slow. no agenda right now, probably just a conversation on getting back into it with a few of the outstanding reviews18:47
jamielennoxdtroyer_zz: did you get a change to follow up https://review.openstack.org/#/c/168770/18:56
jamielennoxdtroyer_zz: also people start ignoring the _zz when you are known to still respond18:56
dtroyer_zzjamielennox: the zz is a znc problem, I'll go kick it…18:57
*** dtroyer_zz has quit IRC18:58
*** dtroyer has joined #openstack-sdks18:58
openstackgerritMerged stackforge/python-openstacksdk: The first functional test  https://review.openstack.org/17034319:50
sigmavirus24etoews: is trying to lower my life-expectancy =P20:02
etoewsdoing my best20:02
etoewseveryone wants more and more users for their oss projects right?20:02
etoewsjust to confirm. this is the canonical place for our docs right? http://python-openstacksdk.readthedocs.org/en/latest/20:04
etoewsor is there something on openstack.org?20:04
etoewswith a url i'm not aware of20:04
terrylhoweyes etoews20:05
terrylhowereadthedocs that is20:05
etoewsgreat. i prefer that over somewhere on openstack.org. it's more where our users would expect to find it.20:06
*** sigmavirus24 is now known as sigmavirus24_awa21:03
*** trown is now known as trown|outttypeww21:05
*** pm90_ has quit IRC21:07
*** sigmavirus24_awa is now known as sigmavirus2421:17
openstackgerritDean Troyer proposed openstack/python-openstackclient: Add docs for service provider CRUD  https://review.openstack.org/16856221:20
*** mattfarina has quit IRC21:27
*** boris-42 has joined #openstack-sdks21:39
*** pm90_ has joined #openstack-sdks21:40
*** pm90_ has quit IRC21:41
*** pm90_ has joined #openstack-sdks21:41
*** pm90_ has quit IRC22:01
openstackgerritMerged openstack/python-openstackclient: Add warning message if unknown version supplied  https://review.openstack.org/17310222:05
openstackgerritMerged openstack/python-openstackclient: Use cliff deferred help instead of homemade one  https://review.openstack.org/15877922:14
*** britthouser has quit IRC22:44
*** sigmavirus24 is now known as sigmavirus24_awa22:59
*** bknudson has quit IRC23:07
*** chlong has joined #openstack-sdks23:10
*** chlong has quit IRC23:43
*** chlong has joined #openstack-sdks23:45
*** britthouser has joined #openstack-sdks23:50
*** britthouser has quit IRC23:59

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