Thursday, 2010-12-23

*** joearnold has quit IRC00:03
*** sophiap has quit IRC00:13
openstackhudsonProject nova-tarmac build #45,862: FAILURE in 5 min 47 sec:
openstackhudsonProject nova build #325: SUCCESS in 1 min 15 sec:
openstackhudsonProject nova-tarmac build #45,863: STILL FAILING in 3 min 11 sec:
*** dubsquared has joined #openstack00:17
openstackhudsonProject nova-tarmac build #45,864: STILL FAILING in 3 min 1 sec:
*** daleolds has quit IRC00:21
openstackhudsonProject nova-tarmac build #45,865: STILL FAILING in 2 min 57 sec:
vishyeday: any idea why tarmac would just suddenly start complaining about a commit message error?00:28
openstackhudsonProject nova-tarmac build #45,866: STILL FAILING in 2 min 56 sec:
*** irahgel1 has left #openstack00:32
*** dirakx has joined #openstack00:33
*** whaley has joined #openstack00:34
openstackhudsonProject nova-tarmac build #45,867: STILL FAILING in 4 min 49 sec:
openstackhudsonYippie, build fixed!00:37
openstackhudsonProject nova-tarmac build #45,868: FIXED in 3.9 sec:
openstackhudsonProject nova build #326: SUCCESS in 1 min 16 sec:
*** maplebed has quit IRC00:40
*** dubsquared has quit IRC01:00
*** pothos has joined #openstack01:04
vishylooks like one of my trunk merges accidentally overwrote authors of some commits.01:11
*** Cybodog has quit IRC01:14
vishyI think I know what it was.  In one of my branches I merged trunk and fixed conflicts, then i realized i needed to mess with my last commit so i did a shelve/mess around/unshelve/commit, which probably lost the commit history of the merge. Any bzr geniuses know how we can recover the original authors?01:18
*** eday has quit IRC01:21
vishyalso three of my commits were as <>, don't suppose that is fixable either01:26
vishyer root <root@ubuntu> that is01:26
*** aliguori has quit IRC01:31
*** larstobi has quit IRC01:39
*** eday has joined #openstack01:44
*** ChanServ sets mode: +v eday01:45
*** abecc has quit IRC01:46
*** daleolds has joined #openstack01:51
*** _skrusty has quit IRC02:04
*** kashyapc has joined #openstack02:04
openstackhudsonProject nova build #327: SUCCESS in 1 min 15 sec:
vishywell: mailmap is fixed now, but I guess no karma :)02:16
*** _skrusty has joined #openstack02:17
jesse_isn't karma only periodically recalculated02:17
edaynightly I think02:19
*** blpiatt has joined #openstack02:26
*** daleolds has quit IRC02:28
*** mtaylor|afk is now known as mtaylor02:28
blpiattAnyone have a minute to talk nova & guest instance DNS integration?02:28
mtaylordevcamcar: whatcha need packaging-wise?02:31
*** jdurgin has quit IRC02:35
*** jbaker has joined #openstack02:37
*** kevnfx has joined #openstack02:40
edaymtaylor: if you could make nova tarmac use nstead of, that would be swell :)02:41
edaymtaylor: probably needs the --no-virtual-env flag02:41
mtayloreday: why?02:42
mtayloreday: all does is possibly set up a virtual env and then run runtest.py02:43
mtayloreday: or is there something I'm missing02:44
*** littleidea has quit IRC02:47
*** dubsquared has joined #openstack03:06
*** miclorb_ has quit IRC03:08
winston-dHi Gurus.  Is it normal that my uploaded-registered images is being described as 'decrypting' ?03:10
winston-dAnd how long does it take to finish 'decrypting'?03:11
*** Cybodog has joined #openstack03:32
*** hadrian has quit IRC03:49
*** ArdRigh has joined #openstack04:07
*** littleidea has joined #openstack04:07
*** krish has joined #openstack04:09
*** krish has left #openstack04:10
*** jbaker has quit IRC04:24
*** gaveen has joined #openstack04:41
*** AimanA is now known as HouseAway04:50
*** jbaker has joined #openstack04:51
*** MarkAtwood has quit IRC04:51
*** jbaker has quit IRC04:51
*** jbaker has joined #openstack04:51
*** blpiatt has quit IRC04:54
mtayloreday: ?05:11
*** MarkAtwood has joined #openstack05:13
*** ksteward has joined #openstack05:14
*** jbaker has quit IRC05:14
*** dubsquared has quit IRC05:18
edaymtayler : i have a branch that removes, but uses runtest.sh05:23
*** kashyapc has quit IRC05:41
*** kashyapc has joined #openstack05:42
*** ccustine has quit IRC05:48
*** littleidea has quit IRC05:55
*** littleidea has joined #openstack05:58
*** miclorb_ has joined #openstack06:00
*** kevnfx has quit IRC06:00
*** ramkrsna has joined #openstack06:01
*** Abd4llA has joined #openstack06:13
*** sirp1 has quit IRC06:29
mtayloreday: cool. how about if I just have tarmac run nosetests...06:30
mtayloroh - I suppose i need to change the test script so that your branch can go through when approved... k06:31
mtayloreday: done. let me know when your branch has landed - I would love to have tarmac just call nosetests and be done with it06:32
*** joearnold has joined #openstack07:04
*** guigui1 has joined #openstack07:10
vishywinston-d: no it should not take long07:11
vishywinston-d: it may be a bug was introduced recently07:12
vishystackers: a whole bunch of branches have landed recently.  It would be great if people could try run throughs of various installation methods and make sure nothing has broken.  Also I believe we have no logging in the packages right now which is particularly annoying07:14
vishymtaylor: enabling this branch was causing hudson weird build failures:
vishymtaylor: check out the bottom of the console output here:
vishymtaylor: any ideas what could cause that?07:16
winston-dvish, it's been 5 hours, images is still 'decrypting'.  And i am using nova-2010.1 release version07:20
*** cclien has joined #openstack07:28
*** jesse_ has quit IRC07:35
*** jesse_ has joined #openstack07:36
*** ibarrera has joined #openstack07:38
ttxdevcamcar: yes, you should file a bug. Means needs to be split in packaging, unless soren sees a smarter way07:43
ttxdevcamcar: that shouldn't prevent you from landing the feature though... we can fix the package split afterwards.07:45
ttx(as long as the bug is filed :)07:45
winston-dhow can i delete one compute node from database? not using SQL07:45
*** MarkAtwood has quit IRC07:50
*** Abd4llA has quit IRC07:53
*** brd_from_italy has joined #openstack07:53
*** miclorb_ has quit IRC07:54
*** dragondm has quit IRC07:57
*** dragondm has joined #openstack07:57
*** rcc has joined #openstack08:02
*** rcc has quit IRC08:11
*** littleidea has quit IRC08:11
*** joearnold has quit IRC08:28
*** Abd4llA has joined #openstack08:29
olivier_Hi all08:31
*** allsystemsarego has joined #openstack08:33
*** allsystemsarego has joined #openstack08:33
*** ramkrsna has quit IRC08:34
*** irahgel has quit IRC08:42
*** ramkrsna has joined #openstack08:47
olivier_Is the normal behavior that lot's of swift-ring-builder commands return 1 (false) even if the result is correct ?08:50
*** miclorb_ has joined #openstack08:51
olivier_Because I'm writing a shell script that do the full configuration of a swift cluster, and this behavior is a (little) problem08:52
*** irahgel has joined #openstack08:56
*** MarkAtwood has joined #openstack08:59
*** littleidea has joined #openstack08:59
uvirtbotNew bug: #693714 in swift "Incorrect return code with swift-ring-builder" [Undecided,New]
*** irahgel has quit IRC09:14
*** littleidea has quit IRC09:15
*** MarkAtwood has quit IRC09:20
*** MarkAtwood has joined #openstack09:23
*** irahgel has joined #openstack09:28
*** dkocher has joined #openstack09:31
*** kashyapc has quit IRC10:08
*** ibarrera has quit IRC10:10
*** ibarrera has joined #openstack10:10
wedjathello all10:18
*** Abd4llA has quit IRC10:18
wedjatis there a need to restart storage nodes after uploading .gz files from the proxy, after a rebalancing ?10:18
wedjatolivier_: i'm having this behaviour too10:21
olivier_wedjat: Hi, nobody online this morning :-(10:22
*** littleidea has joined #openstack10:23
*** kashyapc has joined #openstack10:25
*** miclorb_ has quit IRC10:35
*** dirakx has quit IRC11:16
*** ctennis has quit IRC11:29
*** ctennis has joined #openstack11:42
*** ctennis has quit IRC11:44
*** ArdRigh has quit IRC11:55
*** Abd4llA has joined #openstack11:59
*** reldan has joined #openstack11:59
*** reldan has quit IRC12:06
*** reldan has joined #openstack12:19
*** dirakx has joined #openstack12:22
*** brd_from_italy has quit IRC12:36
*** gaveen has quit IRC12:43
*** brd_from_italy has joined #openstack12:50
uvirtbotNew bug: #693771 in nova "No more logging with trunk Ubuntu packages" [Undecided,New]
*** rogue780 has joined #openstack12:55
*** rogue780 has quit IRC13:11
*** zul has quit IRC13:16
*** zul has joined #openstack13:17
*** _skrusty has quit IRC13:19
*** reldan has quit IRC13:21
ttxAwesome -- Thanks to recent fixes in bug 661159 and bug 691598, using no-ramdisk official Ubuntu cloud images now works perfectly in Nova.13:23
uvirtbotLaunchpad bug 661159 in nova "Does not allow booting without a ramdisk" [High,Fix committed]
uvirtbotLaunchpad bug 691598 in nova "Metadata service queries fail" [High,Fix committed]
*** brd_from_italy has quit IRC13:28
*** brd_from_italy has joined #openstack13:31
*** _skrusty has joined #openstack13:31
*** blpiatt has joined #openstack13:41
*** hadrian has joined #openstack13:48
*** blpiatt has quit IRC13:56
*** ramkrsna has quit IRC13:59
*** holoway_ is now known as holoway14:00
*** reldan has joined #openstack14:06
notmynamewedjat: no, the rings are automatically reloaded based on the file's mtime14:06
*** littleidea has quit IRC14:17
*** littleidea has joined #openstack14:19
*** zul has quit IRC14:19
*** zul has joined #openstack14:20
*** dillon-w has joined #openstack14:21
dillon-whi all14:22
dillon-whow does nova delete services?14:23
dillon-wwhen I stop one compute node, it is still in services table.14:25
*** zul has quit IRC14:35
*** zul has joined #openstack14:35
*** gondoi has joined #openstack14:37
*** doude has quit IRC14:37
*** doude has joined #openstack14:38
wedjatnotmyname: ok, thanks ! it's really really useful then ! :D14:41
*** jdarcy_ has joined #openstack14:41
*** omidhdl has joined #openstack14:49
*** Podilarius has joined #openstack14:50
*** sophiap has joined #openstack14:50
*** omidhdl has joined #openstack14:50
jeremybolivier_: will you release that script?14:55
*** omidhdl has quit IRC14:55
*** omidhdl has joined #openstack14:56
*** jdarcy has quit IRC14:58
*** jdarcy_ is now known as jdarcy14:59
*** guigui1 has quit IRC15:01
*** omidhdl has quit IRC15:01
*** omidhdl has joined #openstack15:02
*** reldan has quit IRC15:11
*** omidhdl has quit IRC15:11
*** omidhdl has joined #openstack15:13
*** omidhdl has quit IRC15:15
*** omidhdl has joined #openstack15:17
*** omidhdl has quit IRC15:19
*** omidhdl has joined #openstack15:21
*** reldan has joined #openstack15:25
*** omidhdl has quit IRC15:25
*** omidhdl has joined #openstack15:27
*** hggdh has quit IRC15:30
*** omidhdl has quit IRC15:31
*** hggdh has joined #openstack15:32
*** omidhdl has joined #openstack15:32
*** omidhdl has joined #openstack15:36
*** hggdh has quit IRC15:37
*** dkocher has quit IRC15:38
*** omidhdl has joined #openstack15:38
*** sirp1 has joined #openstack15:39
*** kevnfx has joined #openstack15:40
*** omidhdl has quit IRC15:40
*** hggdh has joined #openstack15:41
*** omidhdl has joined #openstack15:41
*** hggdh has quit IRC15:43
*** omidhdl has quit IRC15:45
*** omidhdl has joined #openstack15:45
*** omidhdl has quit IRC15:47
*** kevnfx has quit IRC15:47
*** omidhdl has joined #openstack15:48
*** omidhdl has quit IRC15:49
*** omidhdl has joined #openstack15:51
*** johnpur has joined #openstack15:52
*** ChanServ sets mode: +v johnpur15:52
*** omidhdl has quit IRC15:53
*** omidhdl has joined #openstack15:55
*** fitzdsl has quit IRC15:56
*** fitzdsl has joined #openstack15:56
*** Ryan_Lane has joined #openstack15:58
*** dubsquared has joined #openstack15:59
*** Podilarius has quit IRC16:01
*** daleolds has joined #openstack16:01
*** zul has quit IRC16:05
*** WonTu has joined #openstack16:06
*** hggdh has joined #openstack16:06
*** WonTu has left #openstack16:06
*** jbaker has joined #openstack16:08
*** zul has joined #openstack16:08
ttxRyan_Lane: trying to set importance for bug 681164, I'm not sure I understand what you mean by "creating DNs for searches"16:10
uvirtbotLaunchpad bug 681164 in nova "Use a search to find DNs instead of creating them directly from attributes" [Undecided,New]
*** kevnfx has joined #openstack16:13
*** zul has quit IRC16:14
*** reldan has quit IRC16:14
*** zul has joined #openstack16:16
Ryan_Lanettx: right now the DNs for searches are generated by taking the root, the attribute, and the name of the object. they should be pulled from ldap via a search16:17
Ryan_Lanesince it is possible, and likely, that in a pre-existing ldap infrastructure there will be heirarchy16:18
Ryan_Laneand pre-existing users16:18
Ryan_Laneand naming attributes that won't match the dns we create16:19
Ryan_Lanenaming attributes are the major reason16:19
ttxRyan_Lane: so the impact is performance + risk of namespace collision ?16:19
Ryan_Laneactually, it'll be a little less performant16:20
ttxRyan_Lane: heh16:20
Ryan_Lanethis is for interoperability, and working with existing directories16:21
* ttx settles for "Medium"16:21
Ryan_LaneI'd agree16:21
*** reldan has joined #openstack16:32
annegentleheh. oops16:32
annegentletoo many windows :)16:32
*** kashyapc has quit IRC16:34
mtaylorvishy: yeah- the merge prop didn't have either a description or a commit message, so there was no text to use for a commit message16:35
mtaylorvishy: we're supposed to be blocking those from merge, or responding sensibly - so there's a bug there i guess16:35
jeremybtoo many windows!16:35
annegentlewhy oh why do I hit enter? :)16:36
*** reldan has quit IRC16:37
ttxannegentle: you assigned yourself to bug 681265... does it mean you're working on it ?16:38
uvirtbotLaunchpad bug 681265 in nova ""nova-manage zipfile" step is missing from single.node.install docs" [Low,Triaged]
annegentlettx: Hm, I fixed it. How should I indicate that it's fixed?16:38
ttxannegentle: set it to FixCommitted16:39
ttxannegentle: I'll do it for you, I'm on it16:39
annegentlettx:  done16:40
ttxjaypipes: about bug 686438... is it still in progress ? Should I set it to "In Progress" ?16:42
uvirtbotLaunchpad bug 686438 in glance "Cannot create image: Json Circular reference detected" [Critical,Confirmed]
* ttx can't help to notice that jaypipes works long hours.16:43
mtaylorttx: jaypipes is always working16:44
ttxmtaylor: see
uvirtbotLaunchpad bug 686438 in glance "Cannot create image: Json Circular reference detected" [Critical,Confirmed]16:45
*** zul has quit IRC16:46
*** zul has joined #openstack16:46
*** dubsquared has quit IRC16:49
*** daleolds has quit IRC16:49
jk0vishy: looks like this line in nova/ breaks compute: flags.DEFINE_string('ca_file', 'cacert.pem', _('Filename of root CA'))16:49
jk0doesn't like the _()16:49
jk0in trunk16:49
*** kashyapc has joined #openstack16:50
jk0well, there's a few of them in there actually16:51
*** ibarrera has quit IRC16:52
*** dillon-w has quit IRC16:55
dendrobatesthe _() doesn't like jk016:55
jk0it sure doesn't16:56
*** brd_from_italy has quit IRC16:56
vishyjk0: we can't use gettext ranslation for flags?16:57
jk0apparently not, kept getting NameError: name '_' is not defined16:57
jk0also getting an error about self.driver.init_host() not existing from this commit:
jk0in manage.py16:58
jk0prevents nova-compute from starting16:59
vishyjk0: how does that run? it seems to work for me16:59
vishyjk0: i just ran from trunk16:59
jk0I opened up a python shell and imported XenAPI, then nova.virt.xenapi_conn16:59
jk0that's where the _ exception came up16:59
vishyjk0: ah16:59
vishy_ is loaded by gettext in the bin files17:00
vishytry nova-manage shell bpython17:00
vishynova-manage sh bpython/ipython/python is the best way to do a console, because it loads dependencies and flagfiles properly17:01
jk0damn, that's pretty nice17:02
jk0how did I not know about this before :)17:02
vishyjk0: bpython and ipython are both really sexy if you haven't used them by the way17:02
jk0so does nova-compute in trunk run for you?17:02
* jk0 googles those17:02
vishyjk0: yes i just ran it17:02
*** zul has quit IRC17:03
* ttx leaves for his Christmas break17:03
jk0 I'm getting this17:03
* vishy waves at ttx17:03
jk0but it makes sense17:03
ttxvishy: happy christmas !17:03
jk0I'll just add it17:03
vishyah, yes I added that recently and probably forgot to add it to xenapi17:03
jk0it's no problem, I can add it now17:04
*** seshu has joined #openstack17:04
vishyjust copy the one from fake connection17:04
* vishy forgot that we aren't using any sensible inheritance for the compute drivers17:04
jaypipesttx: hehe, I was at breakfast..17:07
jaypipesttx: looking at bug sec17:07
dabojk0: I add this to my ad hoc scripts:
*** zul has joined #openstack17:08
jk0dabo: what's that do?17:08
jk0well I mean, where do you use it?17:08
dabojk0: it injects the _() function into the builtin namespace17:08
jaypipessirp1: what do you think about
uvirtbotLaunchpad bug 686438 in glance "Cannot create image: Json Circular reference detected" [Critical,Confirmed]17:08
daboif I have a script I'm using to test something17:08
vishyjk0, dabo: nova-manage sh script17:09
vishyit used to work but i haven't tried it for a while.  Supposed to run an external script with all dependencies / flags imported17:09
daboit keeps you from getting the error by emulating what the bin files do17:10
jk0ah I think I remember looking over that code once, doesn't it do something similar to what django does?17:10
sirp1jaypipes: the last comment on that bug seems to indicate you were thinking of a solution, anything ever come to pass from that?17:10
*** reldan has joined #openstack17:11
dabovishy: I don't see that script17:11
jaypipessirp1: that was that model_to_dict() method... remember?  still think we should go that route?17:11
jaypipesvishy: so, trying to challenge my mountain of karma, eh? muhahahahaha17:12
vishyjaypipes: I pulled into the lead :)17:12
sirp1jaypipes: if i'm remembering correctly, the issue with that would be lazy-loading of relationships, since the dictionary would have to contain all of the relations… unless we were to write a lazy-dict class17:12
vishydabo: it is part of a nova-manage command17:13
jaypipessirp1: right... so what do you think?  should we go that route or go the route of managing the properties collection manually?17:13
jaypipesvishy: gah! /me redoubles his effort..17:13
vishydabo: for example if you have a script like the following called test.py17:14
sirp1jaypipes: hmm, for now, i'd lean towards using the model_to_dict method, and all realtions will have to be handled manually (e.g. we will not embed a relation within the model-dict, we'd have to fetch it explicitly if we were interested)17:14
vishydabo: root@i-uswrlzxk:~# trunk/bin/nova-manage shell script test.py17:15
vishydabo: sqlite:////home/vishvananda/trunk/nova.sqlite17:15
jaypipessirp1: gotcha. and I agree with that approach for now, at least, while the registry schema is so simple.17:15
dabovishy: so the idea is that nova-manage builds the whole environment?17:16
vishydabo: it just does the proper imports and loading that the other bin files do, so you don't have to manually load flagfiles, set python path, import gettest, etc.17:17
vishydabo: makes it easier to write little testing or conversion scripts.17:18
vishydabo: another possibility is to make a test case and run it individually with run_tests, but this is simpler for quick and dirty stuff.17:19
jaypipessirp1: btw, just read armando... "is a throw's stone away". :)17:19
dabovishy: ok, thanks. Good to know that that's there17:19
*** DubLo7 has quit IRC17:20
vishydabo: np17:20
* vishy smells a blog future blog post.17:21
*** DubLo7 has joined #openstack17:23
jbrycewow...i know one oss team that's been busy. i have a day of calls and meetings and come back to 300 bug and merge request emails.17:26
openstackhudsonProject nova build #328: SUCCESS in 1 min 21 sec:
jaypipesannegentle, sirp1: review changes made. pls review and approve merge proposal if good...
jaypipesjbryce: :)17:31
sirp1jaypipes: will do17:31
sirp1jaypipes: for the datetime issue in glance, i'm actually thinking of changing the fix: instead, i was thinking of creating a PROTECTED_ATTRS set which included the deleted field and the timestamps, and forbidding those from being set explicitly. thoughts?17:33
*** DubLo7 has quit IRC17:33
_cerberus_If someone could take a quick peek at that would be awesome. I think I've shored up the comments.17:35
jaypipessirp1: that's cool with me.17:37
sandywalshTrying to get trunk to run ... getting:
sandywalshanyone seen "_" not defined lately?17:37
jaypipessandywalsh: that's a bug in nova-combined.17:37
jaypipessandywalsh: happened because the eventlet_merge branch went before the i18n branch and I didn't catch it.  file it as a bug and I'll have a fix immediately on it.17:38
sandywalshjaypipes, thanks!17:38
sandywalsh_cerberus_, I'll pull your branch in the meanwhile (hmm, that sounds bad)17:38
jaypipessandywalsh: the _() method is installed by gettext in all endpoints (nova-combined is a new endpoint that came in with the eventlet_merge branch)...17:38
_cerberus_sandywalsh: :-*17:39
sandywalshjaypipes, yup, it used to work for me. i18n is the main change since.17:39
sandywalshand I'm wearing my bizaar t-shirt today. Although I'm thinking it should be bizarre instead.17:41
jaypipessandywalsh: do you need a fix right away?  if so, just import gettext and then add the line: gettext.install('nova') in nova-combined above the nova imports.17:46
sandywalshjaypipes, nah, I'm ok, I'll review _cerberus_ branch until you're ready17:47
sandywalshdo you still need the bug filed?17:47
*** kevnfx has quit IRC17:48
jaypipessandywalsh: no worries, I'll file it.  trying to catch vishy in karma again, anyway. ;)17:48
*** kevnfx has joined #openstack17:49
*** daleolds has joined #openstack17:49
openstackhudsonProject nova build #329: SUCCESS in 1 min 17 sec:
jaypipessandywalsh: hmmm... it seems that nova-combined already has gettext installed...what rev are you on? (do a bzr revno)17:54
sandywalshjaypipes, 48718:00
uvirtbotNew bug: #693864 in nova "nova-combined needs gettext installed" [Low,Confirmed]
sandywalsh_cerberus_, I'm getting that same "'_' not defined" error as trunk ... it'll have to be a visual review18:04
_cerberus_I haven't merged trunk into that branch since yesterday morning, so all my tests currently pass18:05
sandywalsh_cerberus_, yes, but have you run it?18:05
sandywalsh(installed nova?)18:05
*** kevnfx has quit IRC18:05
*** kevnfx_ has joined #openstack18:05
jaypipessandywalsh: :) it's  fixed in 48918:06
jaypipessandywalsh: just do a bzr pull (or bzr merge lp:nova if you prefer)18:06
sandywalshjaypipes, did you change anything?18:06
_cerberus_sandywalsh: Yeah, I've been making calls against the API all morning18:06
jaypipessandywalsh: nope, armando fixed it in a recent commit..18:06
sandywalshjaypipes, cool ... getting. thx!18:06
sandywalsh_cerberus_, you may want to do the same?18:07
_cerberus_Yup, just merged in trunk and repushed that branch18:07
*** dendrobates is now known as dendro-afk18:10
sandywalshjaypipes, _cerberus_ getting closer:
_cerberus_I've also been running each service separately18:11
sandywalshyeah, seems to be a nova-combined issue18:11
jaypipessandywalsh: heh, that's a new one... could you pastebin your nova.conf pls?18:13
jaypipessandywalsh: looks fine to me... hmm.18:15
sandywalshjaypipes, yeah, that's the one I've been using for a dogs age18:16
*** dendro-afk is now known as dendrobates18:17
sandywalsh_cerberus_, which incantation are you using to run individual services?18:18
_cerberus_in terms of conf or simply starting the service?18:19
_cerberus_./bin/nova-<the thing> --flagfile=<the conf>18:19
sandywalshah, ok ... I'll compare against those18:19
sandywalshI'm not using volumes, so that could be bringing down the house now.18:20
_cerberus_Actually, volumes isn't running at all for me18:20
sandywalshyeah, I usually disable it in nova-combined18:20
sandywalshbut that ain't the problem here.18:21
sandywalshThe entire trace:
jaypipessandywalsh: have you merged your local branch up with trunk r489 yet?18:24
sandywalshyes, that was the most recent merge from _cerberus_18:24
sandywalshI'll try this again in the trunk branch to be sure18:25
*** Abd4llA has quit IRC18:25
jaypipessandywalsh: i18:26
jaypipessandywalsh: gah.. I'm stumped as to why get_connection is the class_str there...18:27
sandywalshso, compute cannot load the compute driver18:27
jaypipessandywalsh: sure it is compute?18:27
jaypipessandywalsh: nm, see that now..18:27
termievishy: heading over to La Officina18:27
*** jc_smith has joined #openstack18:27
*** Ryan_Lane is now known as Ryan_Lane|food18:27
sandywalshjaypipes, which defaults to 'nova.virt.connection.get_connection'18:28
sandywalshwhich isn't a module18:28
sandywalshit's function18:28
vishytermie: its open18:28
vishysandywalsh, jaypipes: apt-get install python-cheetah18:29
sandywalshvishy, I think that was it18:30
jaypipesvishy: already have that..18:30
vishysanywalsh: cool18:30
sandywalshyup, had to disable volume and it's running now18:31
*** sophiap has quit IRC18:31
jaypipesvishy: why is FLAGS.compute_driver defaulted to "'nova.virt.connection.get_connection'"18:31
vishybecause that is how the drivers worked before18:32
jaypipesvishy: before what? sorry, don't know about this...18:32
vishyjaypipes: before we switched to manager/driver classes18:32
sandywalshjaypipes, vishy I think the real issue is that something is eating an exception and the underlying failure is being masked18:33
sandywalshwe should be nesting our exceptions18:33
vishysandywalsh: correct18:34
vishysandywalsh: the NotFound raised in utils.get_object should probably return the underlying exception18:36
edayyeah, I just got bit by that too, and had to install cheetah18:36
sandywalshpython's native exceptions don't handle nesting well (Java actually does a good job there). Perhaps we need to tweak our base exception?18:37
vishysure, we could at least grab the traceback or something18:38
sandywalshso, rather than just rethrow, we can embed it with more context18:38
sandywalshyes, even the traceback would be good18:38
*** Ryan_Lane|food is now known as Ryan_Lane18:38
jaypipes_cerberus_: os-api-refactor approved and going to hudson.18:48
jaypipestr3buchet: ping18:49
tr3buchetwhat's up jaypipes18:51
_cerberus_crud, looks like I bombed on some pylint stuff18:51
jaypipestr3buchet: heya. wondered if you had any response to armando's suggestion about unittest for suspend? have you had a chance to take a look at the new xenapi unittests?18:52
jaypipes_cerberus_: pep8?18:52
*** zul has quit IRC18:52
*** sophiap has joined #openstack18:52
tr3buchetjaypipes: i haven't. trying to get another branch fixed up for merging. i'll take a look.18:52
_cerberus_jaypipes: err, yeah18:53
_cerberus_Blank lines in the class defs18:53
*** zul has joined #openstack18:53
jaypipes_cerberus_: yep, pep8 is, well, very peppy.18:53
*** reldan has quit IRC18:53
jaypipestr3buchet: no worries, just was wondering18:53
*** reldan has joined #openstack18:54
jaypipestr3buchet: also, question: I see that pause was changed from suspended to error, and suspended was added...could you explain this?18:54
tr3buchetwe don't currently use pause18:55
tr3buchetand we have only a few states that we can use to match our existing api18:55
tr3buchetso i set pause to erro18:56
tr3buchetand suspend (which we do use) to suspend18:56
tr3buchetit was either that or have both of them mapped to suspended18:56
jaypipestr3buchet: ok, understood.18:56
jaypipestr3buchet: side question: is there really a difference between the two terms?18:57
tr3buchetpause pauses state leaving memory in tact18:57
tr3buchetsuspend dumps memory to disk18:57
jaypipestr3buchet: aha. good to know :)18:58
tr3buchetsame as the difference between sleep and hibernate18:58
jaypipestr3buchet: I'm no expert in the virt layer, as you might have guessed ;)18:58
tr3buchetoh me either!18:59
jaypipestr3buchet: ok, one final question, promise :)19:00
jaypipestr3buchet: for this:19:00
jaypipes300+ task = self._session.call_xenapi('Async.VM.resume', vm, False, True)19:00
jaypipes301+ self._wait_with_callback(task, callback)19:00
jaypipeswhy not simply call the synchronous version of VM.resume?19:01
* vishy is failing to use nokernel19:01
jaypipestr3buchet: It's not your code, of course (you copy/pasted from the others AFAICT), but I'm wondering why the calls to do something async and then immediately waiting for the result of the call :)19:02
_cerberus_Is there a non-async version of that xen call?19:03
_cerberus_I don't think there's complete parity in the API...19:03
vishyhmm looks like null_kernel code code removed in a merge19:03
vishyprobably my fault19:03
jaypipes_cerberus_: ah, well that would answer that question then :)19:03
_cerberus_jaypipes: well, it *may* answer the questoin19:03
_cerberus_question even19:03
_cerberus_I'm just speculating19:04
tr3buchetmy understanding, which could be mistaken, is that the call back executes when the call finishes so bits higher up the chain can move on and do other things in the mean time19:04
_cerberus_jaypipes: it's possible it's unnecessary. We need to confirm19:06
_cerberus_Good eye19:06
tr3buchetthe callback is defined in the compute manager19:06
tr3buchetand passed to the virt driver19:06
jaypipestr3buchet, _cerberus_: I was just curious, that's all... I mean, if the code were, for instance, looping over dozens of instances and calling asynchronously, I'd understand it a bit more... just for one, then waiting for the result seemed odd.19:06
jaypipesit's not a big deal at all, was just curious :)19:08
_cerberus_Sure, but we want to get it right :-)19:08
tr3bucheteach is a thread19:08
tr3buchetso the function returns immediately19:08
tr3buchetbut sets of a thread waiting for the task to complete19:08
sandywalshjaypipes, my understanding is that async is needed to keep eventlet/twisted from blocking on the rpc call19:09
tr3buchetthat way when the task completes it updates the state19:09
jaypipessandywalsh: ah, now that may indeed be the case...good point.19:09
openstackhudsonProject nova build #330: SUCCESS in 1 min 15 sec:
jaypipestr3buchet: ya, I understood that part...was just wondering why it wasn't done like: session.xenapi.Sync.VM.resume; callback(*args, **kwargs)..19:10
sandywalshjaypipes, the _wait_for_task does an eventlet-friendly block until the call returns.19:10
jaypipestr3buchet: but I think sandywalsh may have hit the nail on the head...19:10
tr3bucheti see isee19:10
tr3buchethe knows more than i do, that's for sure19:10
* sandywalsh doubts that19:10
jaypipestr3buchet: either way, review done. :) one tiny thing to fix was all I could find.19:11
tr3buchetawesome thanks jaypipes. i'll take a look19:11
sandywalshwe're getting to the point where a refactor could be done in that call-path to remove much of the copy-paste code.19:11
jaypipessandywalsh: blueprint-away my friend :)19:12
sandywalsha couple more commands should break the camels back I suspect19:12
sandywalshjaypipes, cactus ;)19:12
jaypipessandywalsh: indeed :)19:12
vishythere seems to be some code missing somewhere from the raw disk branch null_kernel is only showing up in one place19:13
tr3buchetyeah those were so similar that bzr completely screwed the merge19:13
tr3buchethad to do it manually19:13
_cerberus_Speaking of bzr, is there a way to get diff/cdiff to highlight extraneous whitespace so stupid merge issues don't bite me again?19:16
_cerberus_Specifically extra spaces at the end of a line or on blank lines19:16
vishyahhah, so null_kernel just isn't used but the flag is still there i guess19:16
jaypipes_cerberus_: hmm, not sure... lemme check.19:16
jaypipes_cerberus_: though pep8 will pick those up ;)19:16
_cerberus_I didn't see anything in the usage for it19:17
_cerberus_Heh, true19:17
jaypipes_cerberus_: yep... bzr --diff-options=-b19:17
_cerberus_w00t, thanks :-)19:18
_cerberus_(I could have done that myself I suppose :-/ )19:18
jaypipes_cerberus_: or rather bzr diff --diff-options=-b19:18
*** kevnfx_ has quit IRC19:18
jaypipes_cerberus_: --diff-options are passed to the external diff program (which, IIRC, for bzr is /bin/diff -u19:19
_cerberus_Ahh, good to know19:20
jaypipes_cerberus_: so any option in diff --help should be able to be passed in that manner...19:20
*** daleolds has quit IRC19:24
jaypipes_cerberus_: denied again! ;P19:25
_cerberus_That's annoying. I got no reported errors against that file when I did it myself19:27
sandywalsh_cerberus_, pylint or pep819:27
jaypipes_cerberus_: pep8 is a bastard, isn't he? :)19:27
_cerberus_And he is!19:28
sandywalshwhat version? I had problems with an old version19:28
* jk0 <3's pep819:28
_cerberus_Ahh, it appears the real answer is I simply *missed* it the first time19:28
* _cerberus_ is world-class19:28
sandywalshI had to move to 0.6.119:28
jk0I've been running 0.5.0 mostly19:29
jk00.6.1 on my mac19:29
* jk0 updates19:29
jaypipes_cerberus_: yeah, pep8 only shows the first error it finds on a source page, iirc...19:30
jaypipes_cerberus_: so you always want to re-run pep8 bin/* nova after making any changes..19:31
_cerberus_Silly me19:31
jk0I always use --repeat19:31
jaypipesjk0: hmm, good to know...19:31
sirp1jaypipes: updated the glance_datetime_obj proposal with new fix + pep8 fixes19:31
* jaypipes checks off his "learn one thing a day" task../19:31
jaypipessirp1: rock on.19:31
_cerberus_jaypipes: by that metric, I'm set for the next month or so19:31
jaypipes_cerberus_: :)19:31
*** kashyapc has quit IRC19:31
jk0jaypipes: :)19:32
*** jfluhmann_ has joined #openstack19:32
vishyis anyone looking at the lack of logging in packages?  It is very difficult to help people debug installation issues without logging19:33
sandywalshI gauge my learning curve by how many new browser tabs I have open at the end of the day of "stuff to read"19:33
vishy_cerberus_, jaypipes: pep8 --repeat19:33
*** sirp1 has quit IRC19:34
jaypipessirp1: off it goes to hudson.19:34
jaypipesvishy: you're too slow! :)19:35
*** sirp1 has joined #openstack19:35
vishydammit i was behind in scrollback19:35
jaypipesvishy: ++ on logging...19:35
jaypipessandywalsh: that's a good metric.  problem for me is, then my wife comes in after dinner to do her "facebookin" and closes my browser windows down.19:36
vishyttx suggested that it requires redirecting stderr in upstart19:36
* vishy is not familiar with upstart. Anyone?19:36
* sandywalsh opens another browser tab for upstart19:39
*** jkakar has quit IRC19:41
*** jkakar has joined #openstack19:42
*** rbergeron has quit IRC19:42
*** Podilarius has joined #openstack19:44
sandywalshphew, ok, have a running nova again19:46
jaypipes_cerberus_: \o/ yeah! you're in trunk.19:48
sandywalshit's a christmas miracle!19:48
jk0merry f'n christmas, _cerberus_19:48
openstackhudsonProject nova build #331: SUCCESS in 1 min 16 sec:
*** daleolds has joined #openstack19:50
dabothat's my dream: _cerberus_ in a trunk!19:57
_cerberus_I'm in ur trunk, makin ur cookiez19:58
sandywalshI like my _cerberus_ like I like my coffee. Ground up and in the freezer.19:58
* _cerberus_ claps19:58
comstudhey, that's where I store my ex-wife.19:59
comstudwhat a coincidence.19:59
*** reldan has quit IRC20:01
*** iRTermite has quit IRC20:05
dabodo we have code for generating the initial root password on a new instance? Or are we just hardcoding something like 'qwerty' for now?20:08
Ryan_Lanedabo: I'd imagine they don't have root passwords20:09
Ryan_Lanekey injection + passwordless sudo20:09
daboRyan_Lane: we need them for Rackspace's use case. I can create my own routine, but wonder if someone had already done something similar.20:10
Ryan_Laneyou can set it when you create the image, I believe20:10
*** ctennis has joined #openstack20:10
jeremybdabo: pwgen?20:11
dabojeremyb: where is that used?20:12
jeremybdabo: it's a debian package and i guess it's available in other distros too20:13
dabojeremyb: oh, I misunderstood. I thought you were saying that it was already being used in OpenStack.20:13
Ryan_Laneyeah, would be nice if openstack would inject the root password20:14
Ryan_Laneshouldn't be too hard20:14
Ryan_Laneto add, that is20:14
*** kernelpanic_ has joined #openstack20:21
*** dendrobates is now known as dendro-afk20:25
*** hggdh has quit IRC20:28
*** HouseAway is now known as AimanA20:28
*** ppetraki has quit IRC20:36
*** henrichrubin has quit IRC20:38
*** ppetraki has joined #openstack20:44
*** MarkAtwood has quit IRC20:51
*** gondoi has quit IRC20:51
*** johnpur has quit IRC20:56
Ryan_Lanewhen I do a describe instances using the EC2 api, it doesn't seem to return IP address information about an instance. I see dnsname and privatednsname... how do I go about getting the IP?21:03
jeremybgethostbyname? :-P21:09
*** jesse_ has quit IRC21:10
Ryan_LaneI need the IP to make a DNS entry with ;)21:10
jeremybis there any requirement that the nova-controller and the mediawiki be the same node?21:11
*** hggdh has joined #openstack21:12
Ryan_LaneI'm doing everything via apis or ldap calls21:12
*** jesse_ has joined #openstack21:18
Ryan_LaneI guess I'll use dnsname for now. not really what I want to do21:18
vishyroot passwords :(21:20
Ryan_Lanewhy frown on root passwords? :)21:21
Ryan_Laneroot passwords aren't evil21:21
XenithI never understood that view.21:21
Ryan_Lanethey are kind of necessary, really21:22
XenithThen again, I've also seen the opposite, where devs decry sudo as the most evil thing ever. :)21:22
vishyit is injecting them that is evil21:23
vishybecause you have to write code that understands images and hypervisors21:23
Ryan_Laneahhh. ok21:23
Ryan_Lanedon't we already do that with ips and keys?21:24
jeremybcan't you just inject with puppet?21:24
vishyonly because we have to21:24
Ryan_Lanejeremyb: that's what I plan on doing21:24
Ryan_Lanesame with keys21:24
vishyif you are using a properly configured guest you don't have to inject21:24
*** jdarcy has quit IRC21:24
Ryan_LaneI'm using the IP injection though21:24
vishyyeah and that is an ugly hack21:24
vishythere was discussion about attaching a cdrom device with information21:25
Ryan_Laneah. that would be nice21:25
vishyso a script could retrieve info, do setup, then detach the cd somehow21:25
vishythat seems like the best strategy to avoid hacking the image21:26
Ryan_Laneyeah. that sounds like a good idea21:26
Ryan_Lanewould be nice if that CD was customizable. there's other stuff I'd like to stick in there21:26
Ryan_Lanevishy: so what's the method of getting an instances IP address using the API?21:28
vishywhich api?21:28
Ryan_LaneI really need to switch to the openstack one at some point21:28
Ryan_Lanethe attributes I get are dnsname and privatednsname21:29
vishyprivatednsname is the ip21:29
Ryan_Lanegreat. thanks21:29
vishydnsname will show the public ip if there is one associated21:29
Ryan_Laneah. good to know21:29
Ryan_Lanenova doesn't currently handle DNS, does it?21:30
jeremybis that the same with amazon?21:30
Ryan_LaneI'd hate to know I'm reimplementing it in my web frontend21:30
*** shane_ has joined #openstack21:31
XenithWhat was that django web frontend someone was working on?21:33
alekibangoXenith: will be released this year21:34
alekibangoby ansolabs21:34
Ryan_Lanenot sure. I started working on a php one before I knew that one was being written21:34
XenithI was going to implement one in rails, but I'd like to take a look at the django one first.21:34
alekibangoRyan_Lane: php? dont we have enough python web frameworks?21:34
jeremybalekibango: it's a mediawiki extension21:35
alekibangolol even rails21:35
Ryan_Laneit's a mediawiki extension21:35
jeremybi still don't even really understand why21:35
XenithWell, I don't know django. :)21:35
jeremybbut so be it21:35
Ryan_Laneour volunteer devs are php people21:35
alekibangojeremyb: freedom  is good21:35
Ryan_Laneand I already have really good ldap code written in mediawiki.21:35
jeremybs/in/for/ ? :)21:36
Ryan_Laneyes :)21:36
Ryan_LaneI also get to use SMW and external data if I use mediawiki21:37
Xenithalekibango: Is this the one?
Ryan_Laneplus documentation pages21:37
Ryan_LaneXenith: that's different21:37
Ryan_Lanethat will give access to VM consoles via the web21:37
alekibangoXenith: waht is coming is django based nova web interface21:38
alekibangofor managing hosts21:38
*** shane_ has quit IRC21:38
jesse_alekibango: actually anso is doing the work but it is released by nasa21:43
alekibangoaha, ty for explaining21:43
XenithAny information about it about yet? Or are we waiting for a general release first?21:45
alekibangoXenith: be patient, it will come out this year21:46
XenithJust curious is all, since my current project is winding down and my next project is openstack.21:48
jesse_Xenith: devcamcar is the person leading the project21:49
jesse_I just pinged him to login if he has time21:49
*** miclorb has joined #openstack21:51
*** sophiap has quit IRC21:51
Ryan_Lanealekibango: this year as in 2010, or as in within a year?21:54
*** ppetraki has quit IRC22:03
*** seshu has left #openstack22:03
jeremybhah, important distinction :)22:07
Ryan_Lanejeremyb: another plus as doing it as a mediawiki extension, it gets localized by translatewiki :)22:09
jeremybRyan_Lane: do we have non english users?22:10
Ryan_LaneI'm going to make the extension non-wmf specific22:11
Ryan_Lanethough we do have non-english users. but we expect developers to write documentation in english22:12
Ryan_Lanethe cluster may expand past using it for test/dev though. so it's a possibility we will22:12
vishyXenith: scalr is also porting to interface with openstack
*** shane_ has joined #openstack22:29
*** piken_ has joined #openstack22:30
*** piken has quit IRC22:31
*** jesse_ has quit IRC22:31
*** kevnfx has joined #openstack22:34
notmynameannegentle: question about when API docs would be ready
*** shane_ has quit IRC22:47
*** Podilarius has left #openstack22:50
uvirtbotNew bug: #693963 in nova " wants sudo privs" [Undecided,New]
*** soosfarm_ has joined #openstack22:58
*** shane_ has joined #openstack23:00
*** jesse_ has joined #openstack23:01
*** allsystemsarego has quit IRC23:01
*** londo_ has quit IRC23:07
*** londo_ has joined #openstack23:07
*** londo_ has quit IRC23:08
*** londo_ has joined #openstack23:09
*** londo_ has joined #openstack23:13
*** piken_ has quit IRC23:18
*** piken_ has joined #openstack23:18
*** shane_ has quit IRC23:21
*** miclorb has quit IRC23:22
tr3buchetcan i get some christmas love for my suspend branch:
tr3buchetit needs merging23:36
*** lvaughn has quit IRC23:42
*** sophiap has joined #openstack23:44

Generated by 2.14.0 by Marius Gedminas - find it at!