Wednesday, 2015-01-14

*** chlong has joined #openstack-sdks00:05
*** tellesnobrega has quit IRC00:09
*** tellesnobrega has joined #openstack-sdks00:11
*** stevemar has joined #openstack-sdks00:17
*** stevemar has quit IRC00:27
*** sigmavirus24_awa is now known as sigmavirus2401:05
*** sigmavirus24 is now known as sigmavirus24_awa01:11
*** rmcall has quit IRC01:20
*** dtroyer has quit IRC01:37
*** britthouser has quit IRC01:49
*** britthouser has joined #openstack-sdks01:49
*** sigmavirus24_awa is now known as sigmavirus2401:53
*** dtroyer has joined #openstack-sdks01:55
*** ayoung has joined #openstack-sdks02:01
*** dtroyer has quit IRC02:08
*** tellesnobrega has quit IRC02:09
*** dtroyer has joined #openstack-sdks02:09
*** edleafe has quit IRC02:13
*** edleafe has joined #openstack-sdks02:17
*** tellesnobrega has joined #openstack-sdks02:20
*** dtroyer has quit IRC02:23
*** dtroyer has joined #openstack-sdks02:25
*** stevemar has joined #openstack-sdks02:46
*** lbragstad has quit IRC02:47
*** lbragstad has joined #openstack-sdks02:55
*** briancurtin has joined #openstack-sdks03:00
*** briancurtin has joined #openstack-sdks03:00
*** lbragstad has quit IRC03:08
*** lbragstad has joined #openstack-sdks03:10
*** ayoung has quit IRC03:31
*** dtroyer has quit IRC03:36
*** lbragstad has quit IRC03:38
*** dtroyer has joined #openstack-sdks03:39
*** ayoung has joined #openstack-sdks03:43
*** britthouser has quit IRC03:44
*** britthouser has joined #openstack-sdks03:44
*** lbragstad has joined #openstack-sdks03:47
*** dtroyer has quit IRC03:52
*** dtroyer has joined #openstack-sdks03:58
*** lbragstad has quit IRC04:00
*** lbragstad has joined #openstack-sdks04:02
*** sigmavirus24 is now known as sigmavirus24_awa04:11
*** dtroyer has quit IRC04:13
openstackgerritDean Troyer proposed openstack/python-openstackclient: Copy wiki content to index page
*** dtroyer has joined #openstack-sdks04:21
*** britthouser has quit IRC04:41
*** britthouser has joined #openstack-sdks04:58
*** britthouser has quit IRC04:59
*** britthouser has joined #openstack-sdks04:59
*** briancurtin has quit IRC05:15
*** lbragstad has quit IRC05:52
*** subscope_ has joined #openstack-sdks05:52
*** lbragstad has joined #openstack-sdks05:55
openstackgerritOpenStack Proposal Bot proposed openstack/python-openstackclient: Imported Translations from Transifex
*** briancurtin has joined #openstack-sdks06:16
*** briancurtin has joined #openstack-sdks06:16
*** briancurtin has quit IRC06:20
*** subscope_ has quit IRC06:45
*** stevemar has quit IRC06:47
*** terrylhowe has quit IRC07:10
*** dtroyer has quit IRC07:18
*** dtroyer has joined #openstack-sdks07:26
*** Miouge has joined #openstack-sdks07:38
*** britthouser has quit IRC07:43
*** britthouser has joined #openstack-sdks07:43
*** chlong has quit IRC07:47
*** briancurtin has joined #openstack-sdks08:04
*** briancurtin has quit IRC08:09
*** dtroyer has quit IRC08:10
*** Miouge has quit IRC08:11
*** dtroyer has joined #openstack-sdks08:12
*** lbragstad has quit IRC08:24
*** lbragstad has joined #openstack-sdks08:34
*** dtroyer has quit IRC08:37
*** dtroyer has joined #openstack-sdks08:39
*** Miouge has joined #openstack-sdks08:40
*** dtroyer has quit IRC08:59
*** dtroyer has joined #openstack-sdks09:00
*** Miouge has quit IRC09:10
*** Miouge has joined #openstack-sdks09:15
*** jamielennox is now known as jamielennox|away09:31
*** Miouge has quit IRC09:32
*** Miouge has joined #openstack-sdks09:33
*** dtroyer has quit IRC09:44
*** briancurtin has joined #openstack-sdks09:53
*** dtroyer has joined #openstack-sdks09:55
*** briancurtin has quit IRC09:58
*** terrylhowe has joined #openstack-sdks10:27
*** briancurtin has joined #openstack-sdks11:42
*** briancurtin has quit IRC11:47
*** chlong has joined #openstack-sdks12:05
*** chlong has quit IRC13:02
*** briancurtin has joined #openstack-sdks13:31
*** briancurtin has quit IRC13:36
*** sigmavirus24_awa is now known as sigmavirus2413:50
*** briancurtin has joined #openstack-sdks14:09
*** briancurtin has joined #openstack-sdks14:09
*** mattfarina has joined #openstack-sdks14:11
*** briancurtin has quit IRC14:32
*** pm90_ has joined #openstack-sdks14:58
*** briancurtin has joined #openstack-sdks14:58
*** pm90_ has quit IRC15:01
*** pm90_ has joined #openstack-sdks15:01
*** Miouge has quit IRC15:22
*** britthouser has quit IRC15:23
*** stevemar has joined #openstack-sdks15:24
*** Miouge has joined #openstack-sdks15:36
*** pm90_ has quit IRC15:49
*** britthouser has joined #openstack-sdks16:14
*** pm90_ has joined #openstack-sdks16:30
*** pm90_ has quit IRC16:33
*** pm90_ has joined #openstack-sdks16:34
*** Guest8210 is now known as redrobot16:54
*** bonsai has joined #openstack-sdks17:00
britthouserTrying to use conn.networks.list_networks().  My tenant has a single network.  But what I get back seems to be an infinite iterable.  Am I doing something wrong?17:03
britthouserHere's my code:17:03
bonsaiHi! I'm wondering whether it's possible to monitor an OpenStack Server and be called back when it reaches a certain state, like with the abiquo VirtualMachineMonitor17:06
dtroyerbonsai: AFAIK none of the Python libs have this built-in, the Compute REST API v2.x does not, so you would have to roll your own polling loop.17:08
bonsaidtroyer: Thanks! I suspected as much so it's good to have this confirmed. May I ask what Python's got to do with it?17:10
dtroyerYou didn't specify if you were asking about the REST or a Python interface…and it's 80% python devs in this room17:11
bonsaiOh, OK, sorry, I didn't get that. I'm actually using Java and I came here because this page points to this channel17:12
bonsaiI'll try the jclouds channel instead. Thanks again :)17:15
*** bonsai has left #openstack-sdks17:16
*** terrylhowe has quit IRC17:20
britthouserAdded a second network, and not it just loops over those two network over and over again.  Is that expected? Shouldn't it stop after the first time through the list?17:23
sigmavirus24britthouser: I think I know what's causing it17:24
sigmavirus24briancurtin: are you around?17:25
briancurtinsigmavirus24: what's up17:25
briancurtinbritthouser: uh yeah that shouldn't happen17:25
sigmavirus24 doesn't halt if teh URL doesn't change (like in britthouser's case) because there's always a body for the same URL17:26
sigmavirus24I'm pretty sure we need an extra check17:26
sigmavirus24i.e., if last_url == current_url, halt17:26
briancurtinhmm, will check it out17:26
sigmavirus24Seems to tie in with the problems with how APIs do pagination17:26
sigmavirus24I'm actually kind of surprised that there isn't anything there to change/update page=#17:27
sigmavirus24(or that we don't check the response's links attr17:27
briancurtinsigmavirus24: a bunch of them dont use links, so this is the most general as could be since it's in the base resource17:28
briancurtini dont know what "change/update page=#" is though17:28
briancurtin(pagination is a disaster)17:29
*** terrylhowe has joined #openstack-sdks17:29
sigmavirus24I know it's a disaster, but I the APIs that come to mind for me all use page=1, page=2, page=3, etc. to paginate17:29
briancurtinthat's one i've never seen17:30
sigmavirus24the URL doesn't really change much each time through the loop so I probably need to look at all the APIs we support more closely17:30
* sigmavirus24 's memory could be totally wrong today17:30
britthouserOk.  Glad I'm not doing something silly.  =)17:32
*** bknudson has quit IRC17:33
sigmavirus24yep I'm totally thinking of something else17:34
briancurtinbritthouser: ok i can confirm the same is happening over here. looking into fixing now17:40
sigmavirus24Right, we update
britthouserbriancurtin: Thanks for quick response.  Need me to do anything?17:40
* sigmavirus24 thinks this is related to requests17:41
sigmavirus24nevermind. we do the right thing17:42
briancurtinbritthouser: maybe another test once we get it figured out, but nothing at the moment i can think of17:42
britthouserbriancurtin:  I'm eager to help, just lemme know. =)17:43
briancurtinbritthouser: are you pointing at a devstack or some vendor cloud? it seems like we're trying to page through results when pagination is disabled, which is why we just keep getting the full results back...forever17:50
*** glenc has quit IRC17:51
britthouserbriancurtin: Its just a RHEL OSP install in my lab.17:51
*** glenc has joined #openstack-sdks17:53
briancurtinugh. this might require the Networks Resource gaining it's own implementation of list to cope with the fact that pagination can be disabled (seems to be by default) and we can't discover if it's disabled - however, this is one that will give us a links entry that should have no "next" to tell us we got everything17:59
sigmavirus24briancurtin: I wonder if instead of Network needing it's own list method, we allow a resource to define a different way of doing pagination that list calls/18:01
terrylhoweI kind of think we need some sort of service properties or something and have pagination=false for network18:02
sigmavirus24This way the amount of duplicated code stays super minimal and we let the class define (from a response) what the next page should be18:02
terrylhowemaybe in user preferences18:03
briancurtinterrylhowe: the bad thing is that setting pagination=false when the server actually paginates misses out on that, and we can't discover through the API18:03
terrylhowethe other way would be see if the response is less than the limit18:04
terrylhowesame kind of thing though, you don’t know the limit of the service18:04
briancurtinwe already do that one when a limit is given, but yeah, can't find that out generally18:05
terrylhoweeither way, I think the user should be able to set some of that stuff as service preferences briancurtin18:07
terrylhowesince different versions/configurations may support different features18:07
terrylhowethat way the user can set the service prefrences to whatever works for their  provider18:08
terrylhowepagination, limits, etc.18:08
*** chlong has joined #openstack-sdks18:10
stevellepagination_from_links = true for network?18:13
briancurtinstevelle: if we just do more generic, supports_pagination, and then either try or dont try, nothing else fancy should be needed for network18:15
briancurtinif i just enter "more_data = False" after the get call within list, it works just fine. devstack has two network entries, and setting limit=1 works fine as-is when i let it page18:17
briancurtini need to go have a fun time at the DMV for a while, so i'll get back to this later unless someone jumps on it18:20
*** bknudson has joined #openstack-sdks18:36
*** briancurtin has quit IRC18:54
*** briancurtin has joined #openstack-sdks19:37
stevemardtroyer, someone is eager for an OSC release on the bug page eh20:30
dtroyerstevemar: who? me?  I am but I'm not sure I follow the reference...20:34
stevemardtroyer, someone just commented on 2 bugs with some strongly worded comments20:35
dtroyerah, I wasn't reading comments…20:36
dtroyerso did you still intend to change the Identity default to v3 now?20:37
sigmavirus24stevemar: pro-tip, follow
stevemarsigmavirus24, most of the time i live by that rule, but i should read bug comments :P20:38
dtroyersigmavirus24: nice, I need to put that on replay and stream from the beginning, one or so a day...20:39
sigmavirus24stevemar: once the bug is fixed/fix committed, just put that on loop20:39
sigmavirus24I mean granted, I don't follow my own advice20:39
stevemardtroyer, i dunno, i'm on the fence, it's a KSC variable20:39
sigmavirus24But I'm already well aware that I make terrible life decisions20:39
dtroyerstevemar: oh, I read that as the OSC default...20:40
dtroyer"The Two Classic Online Blunders: Never get into a LAN war in Asia, and never go in to the comments section when your sanity is on the line."20:40
britthouser"No more rhymes now I mean it"..."Anybody want a peanut?"20:41
sigmavirus24britthouser: A+ Princess bride reference20:42
terrylhoweIt would be nice if instead of v3 default, osc would attempt to guess the release based on the url and if not, v3 default20:42
terrylhoweI would of put that in a comment, but …20:43
britthousersigmavirus24: Thanks! =)20:43
sigmavirus24britthouser: If I weren't listening to progressive gaelic death metal right now, I'd totally try to think of some of the other lines from that scene but alas20:44
britthousersigmavirus24: Another time then. =)  Are you in Ireland? Or just fancy gaelic death metal?20:45
stevemarterrylhowe, please add comments, i don't abide by the 'dont read the comments' rule just yet20:45
dtroyerterrylhowe: not sure I understand?  if a version is in the URL we're just supposed to use it, although I have a hack to strip that and do discovery anyway…20:46
sigmavirus24britthouser: it's a welcome change occasionally. One of my TAs in college introduced me to it. She had a really good taste in music20:46
dtroyeralso, not all comments are created equal, some are more equal than others20:46
terrylhoweif authurl.endswith(“v2.0”) identity_version = 2 something like that dtroyer stevemar20:47
terrylhoweI don’t know the current state of what is going on there, maybe you guys have something like that already.  There have been a ton of changes in that area20:48
dtroyerterrylhowe: actually we're looking to do the opposite, strip the v2.0 and do discovery because SO many clouds still hand out auth URLs with v2.0 in them…20:49
britthouserI hate to monopolize the IRC with issues, but I've hit another.  Same RHEL OSP deployment as before.  When I do a conn.image.list_images() I get error, and I'm not sure what it means.20:49
britthouserglance image-list on command line does show me the images I uploaded.20:51
sigmavirus24britthouser: can you add --debug to the glance call and show me the output?20:52
* sigmavirus24 knows glance better than networking =P20:52
terrylhowesounds a bit confusing to me dtroyer I’d expect v2 auth if I supply a v2.0 auth url20:52
sigmavirus24britthouser: ok, mind seeing if glance --os-image-api-version=2 --debug image-list works?20:55
dtroyerterrylhowe: it is confusing.  but so is expecting to use v3 and still handing out v2.0 URLs.  it's only when discovery is expected to be used anyway20:56
terrylhowehanding out v2 auth urls and expecting v3 auth makes absolutely no sense to me21:00
britthousersigmavirus24: It works, would pastebin of the output be useful?21:01
sigmavirus24britthouser: if you'll humor me21:01
dtroyerterrylhowe: handing out v2 urls and expecting discovery to work is the use case.  I don't remember if it is on or off by default in ksc, but it is already there…21:05
sigmavirus24britthouser: I expect you didn't pick a specific image api version in your User Preferences, right?21:15
britthouserShould I have?21:16
briancurtinbritthouser: i entered a bug to track the network list issue - going to try something a little whacky first, then try manual config21:23
britthouserbriancurtin: Thanks!  I hope the DMV was nice to you. =)21:24
briancurtinbritthouser: they were nice enough to give me directions to the proper place to go after waiting in line forever, so sort of haha. (title never showed up, phone people said go to to place X, place X says to go to place Y)21:25
britthouserbriancurtis: I had my license thrown back at me once, so that sounds like a mild visit. =)21:27
britthouserbriancurtin: I also see the same thing with compute.list_keypairs().  I assume its same root cause?21:27
briancurtinbritthouser: i'll take a look, but probably. one that should work, or has been working for me, is any pagination in object_store21:28
britthouserbriancurtis: Ok I'll add that as another datapoint to the same bug.  So to workaround I edited my neutron.conf and set allow_pagination = False, restarted neturon-server, but that didn't fix it.  Should it have?21:30
britthouserbriancurtis: Sorry I set it to True (and then back to false when it didn't work.)21:30
terrylhowebritthouser: was the image list issue solved?   You still getting that key error?21:30
briancurtinbritthouser: hmm, that is concerning. let me find that config file on devstack and play with it - would have thought setting to True would make it work since we're expecting that it does work21:31
britthouserterrylhowe: Still getting it.21:31
briancurtiner, we're expecting that it woudl be true21:31
britthouserbriancurtin: So my setup is RHEL OSP Icehouse, so if it does work for you, I bet there is a bug in neutron itself.21:32
terrylhowebritthouser: was there ever a pastebin of the sdk with debug?  I didn’t see it.  I was wondering if that was page related as well.21:32
terrylhoweand I’m still talking about the glance issue21:33
britthouserterrylhowe: not from sdk, but I'm happy todo that.  Where do I set debug level?  UserPref?21:33
briancurtinterrylhowe: on the SDK/pagination issue, there was no debug one, but i've run it through and can grab if you want. on my devstack we get back a body with two network entries, set the marker to the last one, make the same request, get the same two body entries, set the marker to the last again, and on and on21:34
briancurtinbritthouser: the keypair one is apparently the same issue - i added one kp and just get the same one forever21:36
terrylhowebritthouser: can you recreate the problem with:21:36
terrylhowepython examples/ openstack/image/v1/image.py21:36
britthouserbriancurtin: So I guess that means nova has pagination turned off too?21:37
terrylhoweI assume you have the sdk cloned21:37
briancurtinbritthouser: i'm going to try a few more things and then look at that, but probably yes21:37
britthouserterrylhowe: Yeah I did a git clone. I'll try that.21:37
britthouserterrylhowe: I get: ERROR: examples.common Exception raised: 'images'21:39
terrylhowepython examples/ openstack/image/v1/ —debug21:39
terrylhowebritthouser: ^^^21:39
sigmavirus24britthouser: should openstack/image/v2/ have a
terrylhowesounds like it is getting the same key error, very strange21:40
terrylhoweit should sigmavirus2421:40
terrylhowemaybe no one ever added one21:40
* sigmavirus24 looks at stevelle =P21:40
sigmavirus24Yeah I think it was forgotten21:40
sigmavirus24i'll send a patchset21:40
sigmavirus24so that looks like glance is requiring the request be made to /v{1,2}/images21:45
sigmavirus24but sdk is requesting just /images21:45
terrylhoweyeh sigmavirus24 it is kind of like the sdk isn’t properly setting the version21:46
*** chlong has quit IRC21:47
briancurtini think i entered an issue in LP for that...that's going to be a real hassle21:47
terrylhowebritthouser: does it work if you add on the —os-ap-version image=v1.0   ?21:47
briancurtinsigmavirus24, terrylhowe: i haven't been following but is this what you're dealing with?
terrylhoweI would guess there is something different than expected in service catalog21:48
sigmavirus24briancurtin: I guess it's something like that21:48
terrylhowelooks good, thanks briancurtin21:48
*** stevemar has quit IRC21:48
terrylhowebritthouser: that should of been —os-api-version image=v1.021:49
britthouserterrylhowe: yeah I did s/ap/api/ but still can't seem to get it to run.  I think image=v1.0 is the script name.21:50
terrylhowebritthouser: what I meant was python examples/ openstack/image/v1/ —debug —os-api-version image=v1.021:51
britthouserterrylhowe: I get: error: unrecognized arguments: —-debug -—os-api-version image=v1.021:52
britthousernot trying to be dense here...sorry if there's an obvious way.21:52
britthouserthat I'm missing21:52
briancurtinbritthouser: "python -m examples.list openstack/image/v1/ --debug --os-api-version image=v1.0" is working for me (well, the KeyError, but it's sending out the request)21:56
terrylhoweso it doesn’t put the v1.0 in the url briancurtin ?21:57
terrylhowebritthouser: I think it was just pasting special characters from the IRC client and that was messing you up21:57
briancurtinterrylhowe: nope, "GET"21:57
britthouserbriancurtain: Thanks for that.  I was trying to edit a copy of and add  #!/usr/bin/python  but it wasn't getting anywhere21:57
briancurtinand that just returns that body with teh supported versioned endpoints21:58
terrylhoweyeh, the url is malformed21:58
* sigmavirus24 wonders if glance should return a 40421:58
* sigmavirus24 doesn't know if that route should be defined21:58
briancurtinit's all documented to do this, it's just really, really weird21:59
sigmavirus24briancurtin: just because it's documented doesn't make it right ;)21:59
britthouserterrylhowe: Here's a pastebin if you still need:
briancurtini believe our PHP people request to this unversioned endpoint before any other glance requests to find out where the real endpoint should go, then cache that, then move on to make requests21:59
sigmavirus24glenc is one of our PHP people, yes?22:02
briancurtinyep, and he sits right behind me. once he gets back i'm going to ask what's up there22:04
sigmavirus24briancurtin: I have yet to meet glenc inspite of being in proximity of them several times22:04
terrylhowewell, it is normal to hit the / and get versions, but not so normal to hit /images and get versions22:04
* sigmavirus24 thinks they were at MadisonPHP last year22:04
sigmavirus24terrylhowe: right22:04
sigmavirus24terrylhowe: I'm relatively new to OS, but I can only imagine (no pun intended) that Glance's API wasn't versioned previously and so /images is backwards-compat shim22:05
briancurtinterrylhowe: i think there's something that redirects unversioned requests to / to give you the actual endpoints22:06
openstackgerritIan Cordasco proposed stackforge/python-openstacksdk: Add image v2 proxy
briancurtinor something weird22:06
terrylhoweah, makes sense.  Either way, the sdk should handle that bettere22:06
sigmavirus24That patchset adds a _proxy to openstack/image/v222:07
*** mattfarina has quit IRC22:18
*** terrylhowe has quit IRC22:19
britthousersigmavirus24: is that patch orthogonal to the images issues, or the fix?22:20
sigmavirus24just something I noticed22:21
britthouserok that's what I thought22:21
*** briancurtin has quit IRC22:22
*** briancurtin has joined #openstack-sdks22:25
*** briancurtin has joined #openstack-sdks22:25
*** bknudson has quit IRC22:37
britthouserOk so if I create a UserPref, and then set the version explcitly, I still see the issue.  Which I think is expected since that is equivalent of what I was doing with "--os-api-version image=v1.0", right?22:37
sigmavirus24britthouser: correct22:39
britthouserOk.  I think I'm caught up. =)  And addresses this, right?22:39
sigmavirus24britthouser: to a degree. yes22:42
britthouserOk.  Really appreciate all the help today guys.  This has been a great help.22:44
*** jamielennox|away is now known as jamielennox22:56
*** briancurtin has quit IRC22:57
*** bknudson has joined #openstack-sdks23:03
*** briancurtin has joined #openstack-sdks23:26
*** briancurtin has quit IRC23:27
*** briancurtin has joined #openstack-sdks23:27
*** briancurtin has joined #openstack-sdks23:27
*** pm90_ has quit IRC23:35
*** sigmavirus24 is now known as sigmavirus24_awa23:52
*** britthouser has quit IRC23:56
*** pm90_ has joined #openstack-sdks23:58

Generated by 2.14.0 by Marius Gedminas - find it at!