Thursday, 2011-02-24

*** sebastianstadil has joined #openstack00:02
*** lamar has quit IRC00:04
*** joearnold has joined #openstack00:05
openstackhudsonProject nova build #564: SUCCESS in 1 min 41 sec:
openstackhudsonTarmac: Adds colors to output of tests and cleans up run_tests.py00:07
openstackhudson* sets working directory for nose to nose/tests so it loads faster00:07
openstackhudson* moves db into nova/tests/test.sqlite00:08
openstackhudson* deletes the db in instead of before running tests00:08
*** dragondm has quit IRC00:09
*** dfg has quit IRC00:11
*** dragondm has joined #openstack00:12
vish1I know you all want faster tests:
*** vish1 is now known as vishy00:14
*** acmurthy has joined #openstack00:15
*** mahadev has joined #openstack00:15
*** nelson has quit IRC00:16
*** nelson has joined #openstack00:16
sirp-hey guys: could we get a few more eyes on this merge-prop; trunk is currently broken00:18
*** Cybo has joined #openstack00:18
*** MUILTFN has quit IRC00:21
sirp-this basically reverts r709 for the moment so we can get trunk stable, with the idea of re-applying some form of 709 down the road00:21
*** dprince has joined #openstack00:25
*** zul has quit IRC00:26
*** mahadev has quit IRC00:27
*** MUILTFN has joined #openstack00:31
justinsbpvo: I saw you approved the revert of r709?00:33
pvoyes, I did. By altering the behavior, it breaks backwards compatibility with images that use the norarc as is.00:35
*** glange has joined #openstack00:35
justinsbpvo: I'd ask you to reconsider your approval00:35
justinsbpvo: Here's why: (1) OpenStack API is not yet stable00:36
justinsbpvo: (2) If compatiblity with novarc was your concern, there's a better fix00:36
sirp-justinsb: yeah, lots of us are using novarc's that would break immediately if we adopt this change....i'm all for fixing this, but it's clear that this needs more discussion (there is a large disagreement on the best approach)00:37
justinsbpvo: (3) I think it's very poor form not to address my concerns00:37
justinsbpvo: (4) This is a major step backwards, and we're going to have to go through all the pain again00:37
justinsbpvo: (5) That's it for now :-)00:37
justinsbsirp-: So why not use the patch which fixes novarc ?00:37
openstackhudsonProject nova build #565: SUCCESS in 1 min 42 sec:
openstackhudsonTarmac: Revert commit 709. This fixes issues with the Openstack API causing 'No user for access key admin' errors.00:37
sirp-justinsb: the revert709 is just a stop gap so we can continue development against trunk while we have this discussion; it's *not* meant to silence discussion :)00:38
dragondmThis change(r709) has consequences that caused all existing systems using existing configs to break.   Doing a refactor of the design oddities with auth between the  2 apis is a good thing, but I think it needs to be proposed as a blueprint.00:38
pvoNot saying the concerns can't be addressed, but the behavior needs to be debated and discussed at least on the ML or in irc before submitting a merge that changes the behavior?00:38
justinsbWhy not merge the novarc fix instead?00:38
justinsbThe old behaviour was so broken I can't believe you guys are defending it...00:38
justinsbpvo: Did you look at the patch before reverting it?00:39
pvoat the r709? yes.00:39
dragondmErmm.. the old behavior was odd, perhaps. But not actually nonfunctional.00:39
justinsbGreat, well, I look forward to going through this all over again when we fix it again00:40
justinsbBig step forward for OpenStack00:40
*** Cybo has quit IRC00:40
justinsbI'm going to open the blueprint00:42
pvojustinsb: I think thats a great idea.00:43
*** burris has quit IRC00:44
*** joearnold has quit IRC00:45
justinsbOK - I've cooled down now - sorry guys00:46
justinsbSo what did it break that was so important?00:46
sirp-justinsb: no worries, man :)  the process broke down a bit, we should have had this discussion  on the mergeprop before it go this far00:46
*** adjohn has joined #openstack00:47
sirp-justinsb: certainly a good debate to have00:47
justinsbsirp-: Well, I think it's obvious we need to fix it.  It didn't occur to me that anyone would object00:47
justinsbsirp-: What did it break?00:47
pvojustinsb: agree that we should have common auth. Just need to have a little heads up before it changes.00:47
justinsbpvo: OpenStack API is not locked though, right?  We're not in backwards compatibility mode yet?00:48
*** j05h has quit IRC00:48
pvothe spec was approved00:48
justinsbThere's a spec?  It was approved?00:49
sirp-justinsb: it was just generating a 'user not found' error when i tried to use nova-tools to hit the OSAPI00:49
justinsbsirp-: Right, the credentials had changed.  But it must have broken something important to warrant the panic amongst the Rackspace crew?00:49
*** dprince has quit IRC00:50
dragondmYah, pretty much all of our dev systems broke when we pulled trunk and few knew why.00:50
sirp-justinsb: well it broke our ability to functionally test any new code developed against trunk00:50
justinsbdragondm: But we talked about that in private IRC this morning and I gave you the solution and went off to fix it00:51
justinsbsirp-: Do you mean the smoketests?  They just use the EC2 API.00:51
justinsbsirp-: Or so I thought!00:52
justinsbsirp-: Do you have tests for the OpenStack API?  Can you contribute them to the project?00:53
sirp-justinsb: no no,  i mean 'poor-mans functional testing' i.e. running through the process by using the command line tools; since nova-tools barfed on the authing, we couldn't verify 'builds', etc00:53
pvojustinsb: we're hoping to correct a lot of the issues we're having with more automated testing00:53
justinsbsirp-: Why not just write a test and contribute it?00:53
pvojustinsb: working on that.00:54
justinsbpvo: Yes.  I'm busy coding tests, which is how I found the auth issue.  Of course, you've broken my tests now, so I hoped it was for a very good reason00:54
pvojustinsb: this also isn't the first time we've been clobbered, so maybe we were a bit sensitive to it as well.00:54
justinsbpvo: clobbered?00:54
pvopulled trunk; can't build00:54
pvooh, and I meant "we are also working on that" not "I hope you are working on that" :)00:55
dragondmAlso, we tried the novarc patch, and it didn't seem to work. We were still getting 'user not found'00:55
justinsbpvo: That's how I took it :-)00:55
justinsbdragondm: That would have been a useful contribution to the thread.00:56
dragondmYah, We had to break off for a few hours to d an interview00:56
pvojustinsb: cool.00:56
dragondmSo I was away from IRC00:56
pvoback later. ttfn00:57
*** pvo has quit IRC00:57
*** KenD has left #openstack00:58
*** MUILTFN has quit IRC01:01
*** dysinger has joined #openstack01:01
*** zul has joined #openstack01:04
*** MUILTFN has joined #openstack01:06
*** burris has joined #openstack01:11
*** joearnold has joined #openstack01:13
*** MUILTFN has quit IRC01:17
*** cc_freenode has quit IRC01:18
*** MUILTFN has joined #openstack01:23
*** jsgotangco has joined #openstack01:29
*** vvuksan has joined #openstack01:37
*** MUILTFN has quit IRC01:37
*** MUILTFN has joined #openstack01:37
*** vvuksan has quit IRC01:39
*** Ryan_Lane has quit IRC01:44
*** zenmatt has quit IRC01:45
*** acmurthy has quit IRC01:46
*** ivan has quit IRC01:47
*** nii has joined #openstack01:47
*** nii has left #openstack01:48
*** gregp76 has left #openstack01:48
*** sebastianstadil has quit IRC01:48
*** ivan has joined #openstack01:53
*** littleidea has quit IRC01:56
*** littleidea has joined #openstack01:57
*** vvuksan has joined #openstack01:58
*** littleidea has quit IRC01:58
*** littleidea has joined #openstack01:58
*** te has quit IRC02:03
*** MUILTFN has quit IRC02:05
*** romain_lenglet_ has joined #openstack02:06
*** ovid has joined #openstack02:16
*** ovid is now known as Guest9059702:16
*** romain_lenglet_ has quit IRC02:19
*** MUILTFN has joined #openstack02:20
*** Guest90597 is now known as ovidwu02:22
*** mray has joined #openstack02:22
*** MUILTFN has quit IRC02:25
*** MUILTFN has joined #openstack02:27
*** vvuksan has quit IRC02:29
*** acmurthy has joined #openstack02:30
*** MUILTFN has quit IRC02:32
*** exlt_ is now known as exlt02:32
*** ChanServ sets mode: +v exlt02:33
*** MUILTFN has joined #openstack02:36
*** abecc has joined #openstack02:36
*** magritte has quit IRC02:37
*** jongwooh has joined #openstack02:38
*** MUILTFN has quit IRC02:39
*** dragondm has quit IRC02:40
*** abecc has quit IRC02:41
justinsbvishy: Loving the color-coded test output :-)02:41
*** vvuksan has joined #openstack02:41
vishyjustinsb: you should merge my other patch (speeds tests up by 2x and prevents leaks02:41
*** jongwooh has left #openstack02:42
justinsbvishy: That sounds great... but then I couldn't commit without having that branch as a pre-req ?02:43
jaypipessirp-, justinsb: admin client for Glance looking good...
vishyjustnsb: true i should find another core person to review it so it can get merged02:43
vishyjaypipes: I want to add some data to images to allow for different net/block drivers02:44
justinsbjaypipes: Looks good, but probably not meant for me!02:44
vishyjaypipes: I assume this will have to go into glance as well02:44
vishyjaypipes: my thinking is that it uses metadata from kernel if there is one, otherwise it uses the image02:45
jaypipesvishy: glance just stores and caches the images and metadata about the image. it allows you to put whatever metadata you want to with the image..02:45
vishyjaypipes: coolness02:45
jaypipesjustinsb: true :)02:45
*** mahadev has joined #openstack02:48
kpepplevishy: i can't tell who is github@ (bzr blame limitation) ... but can you explain why we are cat'ing run_test.err.log in the file ? this will break tests for all new bzr branch-es of lp:nova (since it doesn't exist) ...02:50
vishykpepple: probably a merge error with various branches02:51
vishykpepple: iif we merge my branch here it is fixed there02:52
vishyoops my bad02:52
*** jc_smith has quit IRC02:53
vishythat was an error it should be catting run_tests.log02:53
*** jsgotangco has quit IRC02:54
*** vvuksan has quit IRC02:54
*** Ryan_Lane has joined #openstack02:55
kpepplevishy: cool. thx.02:55
vishykpepple: that code path should only run if it gets a short error message02:56
vishyso it shouldn't break future branches i think02:56
*** mahadev has quit IRC02:57
*** joearnold has quit IRC02:58
*** joearnold has joined #openstack02:58
*** mahadev has joined #openstack03:03
*** gregp76 has joined #openstack03:03
*** mahadev has quit IRC03:04
*** maplebed has quit IRC03:06
*** acmurthy has quit IRC03:07
*** maplebed has joined #openstack03:10
*** lamar has joined #openstack03:14
*** pvo has joined #openstack03:15
*** pvo has quit IRC03:15
*** pvo has joined #openstack03:15
*** ChanServ sets mode: +v pvo03:15
*** dprince has joined #openstack03:23
*** mray has quit IRC03:26
*** mray has joined #openstack03:31
*** kashyap has joined #openstack03:32
*** gregp76 has quit IRC03:32
*** mahadev has joined #openstack03:34
*** mahadev has quit IRC03:38
*** dendro-afk is now known as dendrobates03:41
*** dubsquared has joined #openstack03:43
dubsquaredjaypipes: in the house?03:44
*** mahadev has joined #openstack03:45
*** mdomsch has joined #openstack03:45
*** hadrian has quit IRC03:46
*** Dumfries has quit IRC04:06
*** joearnold has quit IRC04:07
*** deshantm_xenorg has quit IRC04:12
*** guynaor has left #openstack04:12
*** deshantm_xenorg has joined #openstack04:13
*** MUILTFN has joined #openstack04:24
*** mdomsch has quit IRC04:32
*** ccustine has joined #openstack04:39
*** maple_bed has joined #openstack04:40
*** maplebed has quit IRC04:41
*** clayg has quit IRC04:41
*** clayg_ has joined #openstack04:42
*** dprince has quit IRC04:45
*** MUILTFN has quit IRC04:45
*** ccustine has quit IRC04:51
*** gregp76 has joined #openstack04:56
*** dprince has joined #openstack04:56
*** zenmatt has joined #openstack04:58
*** sandywalsh has joined #openstack04:58
*** mahadev has quit IRC05:00
*** dprince has quit IRC05:01
*** kashyap has quit IRC05:01
*** troytoman has joined #openstack05:02
*** aimon has joined #openstack05:04
*** mray has quit IRC05:04
*** mray has joined #openstack05:08
*** mray has joined #openstack05:08
*** troytoman is now known as troytoman-away05:10
*** [ack] has quit IRC05:14
*** [ack] has joined #openstack05:16
*** clayg_ is now known as clayg05:16
*** gaveen has quit IRC05:17
*** dendrobates is now known as dendro-afk05:22
*** kashyap has joined #openstack05:24
*** Ryan_Lane has quit IRC05:29
*** gaveen has joined #openstack05:31
*** gaveen has quit IRC05:36
*** miclorb_ has quit IRC05:36
*** f4m8_ is now known as f4m805:39
*** MUILTFN has joined #openstack05:40
*** mahadev has joined #openstack05:41
*** gaveen has joined #openstack05:48
*** mahadev_ has joined #openstack05:49
*** mahadev has quit IRC05:49
*** gaveen has quit IRC05:50
*** gaveen has joined #openstack05:51
*** gaveen has quit IRC05:53
*** pvo has quit IRC05:58
*** lamar has quit IRC05:59
*** clayg has quit IRC06:02
*** clayg_ has joined #openstack06:03
*** mray has quit IRC06:06
openstackhudsonProject nova build #566: SUCCESS in 1 min 41 sec:
openstackhudsonTarmac: Change missed reference to run_tests.err.log06:07
*** clayg_ is now known as clayg06:11
*** berendt has joined #openstack06:12
*** joearnold has joined #openstack06:20
*** pvo has joined #openstack06:26
*** miclorb_ has joined #openstack06:27
*** pvo has quit IRC06:31
uvirtbotNew bug: #724143 in nova "KeyError: 'metadata' error when creating server with OS API" [Undecided,New]
*** joearnold has quit IRC06:37
*** joearnold has joined #openstack06:45
*** clayg has quit IRC06:47
*** MUILTFN has quit IRC06:48
*** clayg_ has joined #openstack06:48
*** jesse_ has joined #openstack06:51
*** jesse_ is now known as anotherjesse06:51
anotherjessevishy: Network size (14) must be a power of two. Try 1606:52
anotherjessevishy: perhaps we can do a check for that?06:52
anotherjessecooleo - i'll write a bugzor06:52
vishyi think there is a bug06:52
berendtbugs for novatools has to be opened in nova project?06:53
anotherjesseah, I'll assign it to myself06:53
vishythere is definitely one for checking that the cidr fits06:53
vishy(with a fix)06:53
berendtor maybe someone knows this one:
vishynot sure if it also checks for power of 206:53
berendti included my novarc06:53
anotherjessevishy: i'll read06:53
justinsbberendt: Are you running trunk?06:55
berendtjustinsb: rev 71306:55
berendtjustinsb:  727. Revert commit 709. This fixes issues with the Openstack API causing 'No user for access key admin' errors. (detail)06:55
justinsbberendt: Yeah - 709 was my commit, so sorry about that06:56
justinsbberendt: If you update past 727 you'll get back to the behaviour you had before06:56
berendtjustinsb: no problem.. i'll update our packages06:58
justinsbbernedt: Just as a warning, I also introduced another bug with metadata being required for server creation when using the OpenStack API07:01
justinsbberendt: I just posted a hotfix, but it's not in trunk yet.  Is that going to affect you?07:02
berendtcan't say it at the moment.. playing the first time with the openstack api today07:02
berendti used ec2 until now07:02
justinsbberendt: OK!  I think the EC2 API is much more stable at the moment07:03
berendtjustinsb: i can't use euca-desribe-images in combination with glance.. :(07:04
justinsbberendt: What's the issue with using euca-describe-images and glance?  Is this a known bug?07:04
justinsbberendt: Could well be the case, I'm just not familiar with it07:05
uvirtbotLaunchpad bug 723966 in glance "euca-describe-images: Unexpected value of isPublic  for image None" [Undecided,New]07:05
justinsbberendt: Thanks!07:06
*** joearnold has quit IRC07:08
*** zenmatt has quit IRC07:11
*** lvaughn_ has quit IRC07:13
*** lvaughn has joined #openstack07:19
*** CloudChris has joined #openstack07:20
*** naehring has joined #openstack07:20
*** joearnold has joined #openstack07:23
*** littleidea has quit IRC07:26
*** joearnold has quit IRC07:30
*** anotherjesse has quit IRC07:31
ttxwow, lots of bugmail.07:35
*** winston-d has joined #openstack07:37
*** jesse_ has joined #openstack07:38
*** gasbakid has joined #openstack07:46
uvirtbotNew bug: #724159 in nova "Fixes for bug 723298 inadverntently broke flatdhcp" [High,In progress]
*** mahadev_ has quit IRC07:52
*** mgoldmann has joined #openstack07:52
*** rcc has joined #openstack08:00
*** drico has joined #openstack08:05
*** allsystemsarego has joined #openstack08:07
*** allsystemsarego has joined #openstack08:07
*** czajkowski has quit IRC08:10
*** czajkowski has joined #openstack08:10
berendti dropped and recreated my nova database and now i want to resync it with "nova-manage db sync".. but nothing happens.. does nova-manage store the latest synced db revision somewhere on the local filesystem?08:10
*** gregp76 has quit IRC08:16
*** eikke has joined #openstack08:17
berendtberendt: solved.. some local issues08:18
*** gaveen has joined #openstack08:21
*** gaveen has joined #openstack08:21
*** gaveen has quit IRC08:27
*** clayg_ is now known as clayg08:28
*** Nacx has joined #openstack08:30
*** gasbakid has quit IRC08:31
*** dmd17 has joined #openstack08:31
*** sateesh has joined #openstack08:32
berendtare the python-novatools usable at the moment?08:33
*** gasbakid has joined #openstack08:33
*** skiold has joined #openstack08:36
*** calavera has joined #openstack08:47
Ge0RgeR^alekibango :are you online?08:59
*** BelaBoincMe has joined #openstack08:59
*** winston-d has quit IRC09:07
*** MarcMorata has joined #openstack09:09
*** czajkowski has quit IRC09:10
*** czajkowski has joined #openstack09:10
*** irahgel has joined #openstack09:10
vishyvagrant single node and multinode smoketests are running (with my patches).  We're going to plug them into a local hudson tomorrow.09:23
alekibangoGe0RgeR^: yes now09:26
*** BelaBoincMe has left #openstack09:30
*** miclorb_ has quit IRC09:32
*** djbobyd has joined #openstack09:38
djbobydHi all09:39
djbobydI was wondering if anyone tried to connect to nova via elasticfox extension09:40
*** ramkrsna has joined #openstack09:41
ttxdjbobyd: I know some people do.09:41
*** adiantum has joined #openstack09:41
ttxnever tried it myself.09:41
djbobyddo they have success , because I always get 400 Bad status09:41
ttxI think so, I've seen some bugs reported but nothing blocking iirc09:42
djbobydmaybe there is some configuration I am doing wrong09:42
djbobyddo you know if there is something special regarding the configuration of the EC2 API09:43
ttxdjbobyd: I don't know, like I said I haven't tried myself.09:43
djbobydok, thanks anyway09:43
*** metoikos_ has quit IRC09:46
berendtttx: do you think we should discuss the topic "remove the image injection" on the mailing list or is it sufficient to discuss it in my merge proposal?09:46
*** Steppy has joined #openstack09:47
ttxberendt: we'll need to discuss if we want to remove it on the ML, or at the design summit09:48
berendtttx: it think removing it is the right way09:48
ttxberendt: we might need to have the discussion before we add new layers of it, though09:48
*** gasbakid has quit IRC09:50
Ge0RgeR^ttx: about the instances getting lost after reboot09:51
Ge0RgeR^since iam not able to restore them and permanantly lost ...09:51
ttxGe0RgeR^: i expect people to comment on that ML thread and discuss if that should be considered a bug.09:52
*** dmd17 has quit IRC10:17
Ge0RgeR^ttx: but this is getting lost permanently10:26
Ge0RgeR^not able to restore them in any way10:26
Ge0RgeR^this is my nova.conf i have after trunk/ppa10:27
*** londo_ has quit IRC10:47
*** londo_ has joined #openstack10:47
*** adiantum has quit IRC11:01
*** NicoGC has joined #openstack11:43
berendtis the definition of the "cloud controller" on correct?11:44
berendt"with the cloud controller component representing the global state"11:44
berendti would call "nova-api" and "nova-scheduler" the cloud controller, is this right?11:44
*** Guest79035 is now known as soeren11:46
*** freeflyi1g has quit IRC11:48
*** freeflying has joined #openstack11:49
*** dmd17 has joined #openstack11:51
*** Steppy has left #openstack11:57
*** sebastianstadil has joined #openstack11:59
*** h0cin has joined #openstack12:00
*** NicoGC has quit IRC12:00
*** fabiand_ has joined #openstack12:11
*** _0x44_flughafen is now known as _0x4412:17
*** ctennis has quit IRC12:20
*** mgoldmann has quit IRC12:22
*** irahgel1 has joined #openstack12:23
*** irahgel has quit IRC12:23
*** adiantum has joined #openstack12:27
*** naehring has quit IRC12:32
ttxberendt: I think the "cloud controller" in that description is actually RabbitMQ.12:34
berendtttx: I think this should be fixed12:34
*** fabiand_ has quit IRC12:35
*** vvuksan has joined #openstack12:35
*** kashyap has quit IRC12:35
*** zul has quit IRC12:36
*** ctennis has joined #openstack12:36
*** ctennis has joined #openstack12:36
*** anticw has quit IRC12:41
*** sandywalsh has joined #openstack12:41
*** anticw has joined #openstack12:43
*** fabiand_ has joined #openstack13:02
*** Ge0RgeR^ has quit IRC13:23
*** littleidea has joined #openstack13:24
*** Ge0RgeR^ has joined #openstack13:24
*** berendt has quit IRC13:25
*** fitzdsl has quit IRC13:30
*** fitzdsl has joined #openstack13:30
*** uksysadmin has joined #openstack13:34
*** littleidea has quit IRC13:36
uksysadminhi - got a multi-node set up question: I've got a controller running nova-{api,network,scheduler,objecstore,compute} rabbitmq and mysql. I've got a second node running nova-compute - so in essence I've two compute nodes.13:37
uksysadminI can ping each other, everything seems ok13:37
uksysadminwhen I look in mysql on the controller I see my compute nodes (and scheduler and network) listed13:37
*** drico has quit IRC13:38
uksysadminwhen I launch an instance, I get "scheduling" as it can't run any more on the controller... it doesn't seem to want to spawn instances on my standalone compute node13:38
*** drico has joined #openstack13:38
uksysadminwhat information does the api/controller need in order to send the request to that node? does it try to contact via the name in the table - say "nova2"... and if it can't do a dns lookup fails?13:39
*** kashyap has joined #openstack13:44
*** mray has joined #openstack13:48
*** hadrian has joined #openstack13:48
ttxjaypipes: ping me when you have 5 min.13:50
*** Ge0RgeR^ has quit IRC13:51
*** kashyap has quit IRC13:51
uvirtbotNew bug: #724317 in nova "openssl error due to openssl.cnf.tmpl file provided by OpenStack" [Undecided,New]
jaypipesttx: pong13:53
*** mgoldmann has joined #openstack13:54
ttxjaypipes: your extra lp:nova/bexar branch wasn't merged in time... but might trigger another round13:54
uvirtbotLaunchpad bug 714678 in nova/bexar "Missing items in release tarball" [Critical,Fix committed]13:54
ttxjaypipes: I think we need to fix those .pyc files. What's your take on it ?13:55
ttxif we fix that, we'll include the migrate fix as well.13:55
ttxMore delays, but meh13:56
jaypipesttx: no worries :)13:56
*** pvo has joined #openstack13:57
ttxjaypipes: do you think we should fix those .pyc files inclusion before releasing ?13:57
*** pvo has quit IRC13:57
*** pvo has joined #openstack13:57
*** ChanServ sets mode: +v pvo13:57
*** mray has quit IRC13:57
jaypipesttx: yeah, the pyc files should definitely not be in there.13:57
ttxjaypipes: (see last comment on
uvirtbotLaunchpad bug 714678 in nova/bexar "Missing items in release tarball" [Critical,Fix committed]13:57
ttxok. I'll file a new bug, push to cactus, push to bexar.13:58
* ttx starts to hate this point release business.13:58
jaypipesttx: cheers, and ahreed.13:58
ttxand include the migrate fix in the list.13:58
*** adiantum has quit IRC13:59
*** mray has joined #openstack14:00
*** ppetraki has joined #openstack14:06
*** zenmatt has joined #openstack14:12
*** gondoi has joined #openstack14:12
uvirtbotNew bug: #724325 in nova/bexar "Generated tarballs include two .pyc files" [High,In progress]
*** nelson has quit IRC14:19
*** djbobyd has quit IRC14:22
*** sandywalsh has quit IRC14:25
*** sandywalsh has joined #openstack14:25
jaypipesttx: FYI, I pushed the glance admin tool branch. You can see what it looks like in the whiteboard on the blueprint:
uksysadmin(sorted with a sudo nova-manage db sync from the compute node)14:27
sandywalshspent hours on the plane catching up on ML threads. Now I'm under water again :/14:27
*** mray has quit IRC14:37
jaypipessandywalsh: heh, no shit.14:41
*** RobertLaptop has quit IRC14:42
*** pvo has quit IRC14:43
ttxjaypipes: seen that, looks cool.14:45
*** troytoman-away is now known as troytoman14:45
*** CloudChris has quit IRC14:48
*** kashyap has joined #openstack14:48
ttxnova-core: simple review on the critical path to 2011.1.1:
*** abecc has joined #openstack14:55
*** troytoman is now known as troytoman-away14:59
*** pvo has joined #openstack15:03
*** Shanix has joined #openstack15:04
Shanixhi i need help with openstack15:04
Shanixi am able to spawn instances but cannot ssh15:05
Shanixanyone??? who could refer me to a link15:09
Shanixor any hint15:09
vvuksanwhat kind of networking are you using ?15:10
Shanixi am trying to connect from host machine15:11
Shanixand i cant15:11
Shanixi dont get any error at all15:11
Shanixeven in the logs15:11
vvuksanare you using flatnetworking of flatdhcp15:11
vvuksanyou can look at console logs for the instance15:12
Shanixi use the standard dcpbridge15:12
Shanixi am new in this maybe i didnt configure it well15:12
*** lamar has joined #openstack15:13
vvuksanwhat does your nova.conf look like ?15:13
Shanixcan i paste it here?15:13
*** mray has joined #openstack15:15
*** f4m8 is now known as f4m8_15:15
Shanixare you there vvuksan?15:16
Shanixi just pasted the nova.conf15:16
vvuksanshanix: that looks reasonable15:16
vvuksanbtw you can post to in the future15:16
vvuksanare you running both nova-compute and nova-network?15:17
Shanixyes both are running15:17
vvuksanare you running them with --nodaemon option ?15:17
vvuksanjust to see that they haven't errored out ?15:17
Shanixthey are running at daemon15:17
Shanixi checked the logs they seem to be ok15:17
Shanixi am also running rabbitmq15:18
Shanixand redis15:18
vvuksani don't know if redis works any more15:18
vvuksanwhich release is this ?15:19
Shanixi am also not sure but it was in the manual i followed so i installed it15:19
*** RobertLaptop has joined #openstack15:19
Shanixhow do i check for release?15:19
vvuksanwhich one did you download ?15:19
Shanixi installed it through apt-get15:20
Shanixfrom the link provided in openstack.org15:20
Shanixthis repo -> add-apt-repository ppa:nova-core/trunk15:21
vvuksanit's the trunk15:21
vvuksanwhen you do euca-describe-instances15:21
Shanixi guess should be the latest15:21
*** bcwaldon has joined #openstack15:21
vvuksandoes it show instance as running?15:21
*** ovidwu_ has joined #openstack15:21
Shanixi was thinking maybe some routing issue?15:22
Shanixi am running it on ubuntu 10.1015:22
vvuksani doubt it15:23
vvuksanon the compute node go to /var/lib/nova/instances15:23
Shanixcoz i can ping the virtual ip on my side but cant the instance one15:23
Shanixconsol.log is empty15:24
Shanixconsole.log i meant15:25
*** dprince has joined #openstack15:25
Shanixis that an issue?15:25
Shanixi mean there?15:27
vvuksanthat's not a good sign15:30
Shanixwhat should i check next?15:30
vvuksantype virsh list15:30
*** kazu has joined #openstack15:30
vvuksanon the compute node15:30
vvuksando you get anything ?15:30
Shanixyes instance running15:31
Shanixis that good?15:32
Shanixvvuksan: i guess its good that it shows running15:33
lamarbcwaldon: I think you might have unsaved changes15:33
vvuksanShanix: yes that's good15:35
*** ramkrsna has quit IRC15:35
vvuksanShanix: question is why there is no console log15:35
Shanixhow would i know?15:35
vvuksanare you sure you are looking in the correct instance directory ?15:35
Shanixshall i look here  /var/log/libvirt/qemu/15:36
Shanixi pasted in private is that correct?15:37
*** kazu has quit IRC15:40
*** kazutaka has joined #openstack15:41
*** kazutaka has quit IRC15:41
*** Shanix has quit IRC15:42
*** Shanix has joined #openstack15:42
*** kazu has joined #openstack15:43
*** mahadev has joined #openstack15:45
*** sateesh has quit IRC15:48
*** mahadev has quit IRC15:50
*** calavera has quit IRC15:51
*** mray has quit IRC15:52
*** mahadev has joined #openstack15:52
*** dragondm has joined #openstack16:01
*** troytoman-away is now known as troytoman16:01
*** mahadev has quit IRC16:10
*** bcwaldon has quit IRC16:10
*** zul has joined #openstack16:11
jaypipesdubsquared: around?16:13
*** Shanix has quit IRC16:14
*** bcwaldon has joined #openstack16:17
*** mahadev has joined #openstack16:17
openstackhudsonProject nova build #567: SUCCESS in 1 min 40 sec:
openstackhudsonTarmac: Globally exclude *.pyc files from generated tarballs.16:17
*** johnpur has joined #openstack16:19
*** ChanServ sets mode: +v johnpur16:19
*** j05h has joined #openstack16:21
jaypipesanybody else having a problem pushing to LP?16:24
*** dubsquared1 has joined #openstack16:26
dubsquared1jaypipes: in the house?16:26
jaypipesdubsquared1: yup.16:26
dubsquared1in response to question, my install is ubuntu + kvm16:27
openstackhudsonProject nova build #568: SUCCESS in 1 min 41 sec:
dubsquared1so it's using the default s3 image service, nova-objectstore16:27
openstackhudsonTarmac: This fix checks whether the boot/guest directory exists on the hypervisor.16:27
openstackhudsonIf that is not the case, it creates it.16:27
openstackhudsonAnd this time I'm pretty sure I'm not breaking anything with this fix! (see bug 723298)16:27
uvirtbotLaunchpad bug 723298 in nova "FlatManager should not call driver.init_host()" [Low,Fix committed]
jaypipesdubsquared: ah, ok. and why don't you want to use nova-objectstore?16:28
ttxok, final merge proposals up for 2011.1.1...16:29
*** enigma has joined #openstack16:30
ttxnova-core, jaypipes, jk0: please consider giving the two reviews on some love, so that I can generate new candidate tarballs / packages for testing early tomorrow16:30
*** kashyap has quit IRC16:30
ttxoh, jk0 is already on it :)16:31
jk0ttx: already done :)16:31
ttxjk0: consider looking into the other one,
jk0looking now16:31
*** skiold has quit IRC16:32
*** Nacx has quit IRC16:33
*** Nacx has joined #openstack16:34
*** dirakx has joined #openstack16:35
*** uksysadmin has quit IRC16:39
*** littleidea has joined #openstack16:42
openstackhudsonProject nova-bexar build #9: SUCCESS in 1 min 39 sec:
openstackhudsonTarmac: Pull fix for LP #717467 (migrate.versioning ImportError) into Bexar 2011.1.1 release.16:42
uvirtbotLaunchpad bug 717467 in nova/bexar "from migrate.versioning import exceptions fails with recent venv" [Medium,In progress]
*** kashyap has joined #openstack16:50
*** burris has quit IRC16:56
*** fabiand_ has quit IRC16:58
*** lamar has quit IRC16:59
*** kazu has quit IRC17:04
*** xtoddx- is now known as xtoddx17:05
openstackhudsonProject nova build #569: SUCCESS in 1 min 41 sec:
openstackhudsonTarmac: Support HP/LeftHand SANs.  We control the SAN by SSHing and issuing CLIQ commands.  Also improved the way iSCSI volumes are mounted: try to store the iSCSI connection info in the volume entity, in preference to doing discovery.  Also CHAP authentication support.17:07
openstackhudsonCHAP support is necessary to avoid the attach-volume command require an export on the SAN device.  If we had to do that, the attach command would have to hit both the volume controller and the compute controller, and that would be complex.17:07
*** dirakx has quit IRC17:12
xtoddxjustinsb: around?17:14
justinsbxtoddx: Hi!17:16
xtoddxjustinsb: just wondering what needed to change with api to make it testable?17:16
justinsbxtoddx: Sorry, not sure what you're referring to?  An email or a branch?17:17
xtoddxjustinsb: there are some helpers sprinkled around to return wsgi apps to run tests againss17:17
xtoddxjustinsb: sorry, i'm looking at the servicify-nova-api branch17:17
xtoddxjustinsb: and working my way though test-openstack-api-servers now17:18
justinsbxtoddx: Ah ok... so servicify is to support the test-openstack-api-servers branch17:18
justinsbxtoddx: So the test-openstack-api-XX branches all bring up services in-process - nova-api, nova-compute, nova-scheduler etc17:19
xtoddxjustinsb: so it looks like the test-api-servers is integration tests?17:19
justinsbxtoddx: No, they're part of the unit test suit.  They're "integrated" tests.  If they were integration tests, I couldn't include them in the unit test suite.17:20
justinsbxtoddx: ;-)17:20
rykerIs the cloudstack in any way related to openstack?  I'm wondering why they announce openstack releases on their blog page?
justinsbxtoddx: So the changes to servicify nova-api just refactored the bin/nova-api code into a service17:20
justinsbxtoddx: Just like the other services (the non-WSGI ones)17:20
xtoddxjustinsb: so the end result is you stand up a full service instead of just using __call__ on the wsgi app?17:21
justinsbxtoddx: Yes.  I construct an HTTP endpoint, and call it using httplib.17:21
*** dirakx has joined #openstack17:22
justinsbxtoddx: Aiming to catch the problems where I write passing unit tests but then things break trivially17:22
justinsbxtoddx: I sort of need some help on how to shut down the WSGI app if you know about this stuff - can't figure that one out yet!17:23
*** maple_bed is now known as maplebed17:23
xtoddxjustinsb: that is a good question17:23
xtoddxjustinsb: i'll see what I can dig up17:23
justinsbxtoddx: Thanks.  But anyway, what are you thinking about testing using the full services?  Love it or hate it?  (I can't imagine there's going to be much opinion in between)17:24
xtoddxjustinsb: I always love new tests!17:24
justinsbxtoddx: :-)17:24
justinsbxtoddx: I imagine it'll be pulled out into smoke_tests longer term, once we start running them regularly17:25
xtoddxjustinsb: I could see that, but at least you have the steps being written now.  I like that they run with unit tests, which is what is holding the smoketests back.17:26
dprincejustinsb: See my comments on how to reproduce the metadata issue. In the ticket.17:26
dprincejustinsb: I'm also working on some integration/functional tests that will help us catch this stuff quicker.17:27
*** burris has joined #openstack17:27
justinsbdprince: Thanks for the confirmation that the hotfix solves the first problem, and the additional bug report17:27
*** sateesh has joined #openstack17:27
justinsbdprince: I'll add metadata test cases to my "integrated" tests17:27
justinsbdprince: Feel free to branch and extend them also if you'd like :-)17:28
dprincejustinsb: Sure. Will do.17:29
*** gregp76 has joined #openstack17:30
*** acmurthy has joined #openstack17:33
*** lamar has joined #openstack17:38
*** eikke has quit IRC17:39
*** tr3buchet has joined #openstack17:46
*** zul has quit IRC17:52
tr3buchetthere we are17:57
*** mahadev has quit IRC18:00
dubsquared1creiht:  in the house?18:00
openstackhudsonProject nova build #570: SUCCESS in 1 min 51 sec:
openstackhudsonTarmac: Some first steps towards resolving some of the issues brought up on the mailing list related to documenting flags.18:03
openstackhudsonThis patch helps expose some of the features of gflags that we were sort of skipping past, such as organizing flags by the file that defined them in help.18:03
openstackhudsonThere are still some things to fix, for example all the flags defined in nova.flags should be moved to other files, and the ones that are there currently show up under nova.log, but previously they were all in one spot anyway so this is still fully a step forward.18:03
openstackhudsonManager import was moved into Service's init so that the dynamic flags get loaded earlier so we can show them in the help.18:03
*** dprince has quit IRC18:04
jaypipesjustinsb: nice ML post. I'll respond when I've got a free moment... bit swamped today. :(18:08
*** bcwaldon has quit IRC18:08
justinsbjaypipes: Thanks.18:09
justinsbjaypipes: On the HP SAN merge request, can you set it back away from Approved back to Merged please18:09
*** j05h has quit IRC18:10
*** lamar has quit IRC18:10
justinsbjaypipes: Background: I think it already merged clean.   But I had accidentally pushed trunk to the pre-req branch (metadata3)18:10
justinsbjaypipes: I'll check that it did merge clean, but I think if you Approve it again it could cause problems for the merge-bot18:10
justinsbjaypipes: Actually, I was able to do it - never mind18:11
justinsbjaypipes: Interesting permissions model launchpad has :-)18:11
*** Nacx has quit IRC18:13
jaypipesjustinsb: k18:13
*** KenD has joined #openstack18:13
*** doude_ has quit IRC18:17
*** berendt has joined #openstack18:17
berendti'm trying to run iscsi, but i got the following error: can anyone assist?18:18
*** mdomsch has joined #openstack18:20
*** RobertLaptop has quit IRC18:21
*** mgoldmann has quit IRC18:21
*** RobertLaptop has joined #openstack18:21
justinsbberendt: Could you paste the output of an ls on your /dev/disk/by-path/ directory - so we can tell whether it's some sort of permissions issue, or whether the file genuinely isn't there18:23
*** doude has joined #openstack18:23
*** mahadev has joined #openstack18:24
berendtjustinsb: i don't know why.. but i tried it again and now it's working.. strange18:25
justinsbberendt: I think it's actually a timing issue18:25
justinsbberendt: So discovery completes, but the volume isn't yet available18:25
justinsbberendt: I think my patch for HP SANs just landed in trunk, which includes a retry loop18:26
berendtjustinsb: yes the patch is there.. i'll update.. ;)18:26
justinsbberendt: Because I came across the same issue using SANs.  I think the logic is common to all iSCSI targets18:26
tr3buchetjaypipes: that's a doozie18:28
*** dendro-afk is now known as dendrobates18:28
* tr3buchet digs in18:28
jaypipestr3buchet: just letting you know since ewan specifically requested a review from ya.. :P18:29
tr3buchetyeah i've gotten behind on emails a bit (agile training)18:30
*** KenD has quit IRC18:30
tr3buchetsorry to both of you18:30
*** RobertLaptop has quit IRC18:31
jaypipestr3buchet: no worries mate, totally understand. :)18:31
*** RobertLaptop has joined #openstack18:31
*** zul has joined #openstack18:32
*** KenD has joined #openstack18:32
*** fabiand_ has joined #openstack18:41
*** MarcMorata has quit IRC18:44
*** sebastianstadil has quit IRC18:45
openstackhudsonProject nova build #571: SUCCESS in 1 min 41 sec:
openstackhudsonTarmac: Easy and effective fix for getting the DNS value from flag file, when working in FlatNetworking mode.18:49
openstackhudsonProject nova-bexar build #10: SUCCESS in 1 min 38 sec:
openstackhudsonTarmac: Globally exclude *.pyc files from generated tarballs. Fix backported from Cactus.18:50
*** berendt has quit IRC18:52
kpeppleryker: is part of the openstack community
kpeppleryker: although last i checked their solution was mostly java-based ...18:54
*** sateesh has quit IRC18:54
uvirtbotNew bug: #724493 in swift "Bad use of webob.Request.attr and webob.Response.attr" [Low,Triaged]
*** dubsquared1 has left #openstack18:57
*** citral has joined #openstack18:57
*** jlmjlm has joined #openstack18:58
*** dendrobates is now known as dendro-afk18:59
*** mdomsch has quit IRC18:59
*** Pentheus has joined #openstack19:01
jk0vishy: serously <3'ing your test fixes19:02
vishykk0: glad my 8 hours of tearing my hair out was well spent19:03
*** berendt has joined #openstack19:04
*** KenD has quit IRC19:05
*** mdomsch has joined #openstack19:06
*** Ryan_Lane has joined #openstack19:06
openstackhudsonProject nova build #572: SUCCESS in 1 min 41 sec:
openstackhudsonTarmac: Don't require metadata (hotfix for bug 724143)19:07
uvirtbotLaunchpad bug 724143 in nova "KeyError: 'metadata' error when creating server with OS API" [High,In progress]
uvirtbotNew bug: #724503 in nova "network restart fails with postgresql 9" [Undecided,New]
jaypipesdevcamcar: ping!19:16
uvirtbotNew bug: #724507 in swift " line_data encoding" [Low,Triaged]
*** irahgel1 has left #openstack19:17
*** KenD has joined #openstack19:18
*** clauden has joined #openstack19:18
rykerkpepple: I didn't realize was part of the openstack community.  That seems odd to me.  I thought openstack is basically a competitor/alternative to's cloudstack.  Seems kind of like Microsoft being part of the Ubuntu community. :)19:18
*** doude has quit IRC19:19
*** dendro-afk is now known as dendrobates19:20
kpeppleryker: funny you should bring up microsoft ... that's part of's involvement AFAIK --
ttxryker: yes, contributed Hyper-V support in openstack19:23
Glaceehyper-v is not even supported in yet19:24
*** berendt has quit IRC19:24
ttxryker: see lightning talk at
*** alekibango_ has joined #openstack19:25
*** maplebed has quit IRC19:26
*** bcwaldon has joined #openstack19:27
rykerlol.  yeah, i guess that makes sense as far as microsoft providing hyper-v support.19:27
*** maplebed has joined #openstack19:27
rykerbut is another cloud platform for existing virtualization technologies, like kvm, xen, etc. same as openstack, right?19:27
Glaceeryker: yes19:28
rykerany idea what their involvement in the openstack community means?19:28
ryker*thier = cloud.com19:28
*** KenD has quit IRC19:28
Glaceesame as onapp,enomaly and others19:28
*** czajkowski has quit IRC19:28
*** alekibango has quit IRC19:28
*** jimbaker has joined #openstack19:28
rykeri'm wondering, because i'm comparing cloudstack to openstack right now19:29
*** klumpie has quit IRC19:29
*** czajkowski has joined #openstack19:29
*** klumpie has joined #openstack19:30
*** czajkowski has quit IRC19:30
*** czajkowski has joined #openstack19:30
*** KenD has joined #openstack19:30
*** Ryan_Lane has quit IRC19:33
vishyany reviewers around?  FlatDHCP is broken in trunk:
zulcouldnt the libvirt_conn IptablesFirewallDriver so that other virt technologies use it as well?19:40
*** Ryan_Lane has joined #openstack19:42
kpepplevishy: did we change the behavior of / when running individual test files ? i'm trying to test a single test file with './ -N nova/tests/' and it's bombing with 'ImportError: No module named test_nova_manage" ... it was running yesterday19:43
*** nijaba has quit IRC19:44
*** nijaba has joined #openstack19:45
*** nijaba has joined #openstack19:45
ironcamelkpepple: i usually run something like: nova.tests.blah to run just 1 test19:46
openstackhudsonProject nova build #573: SUCCESS in 1 min 44 sec:
openstackhudsonTarmac: Add some methods to the ec2 admin api to work with VPNs.  Also implements and properly documents the get_hosts method.19:47
kpeppleironcamel: where blah == ClassName ... so my test class is "NovaManageTestCase" and my command line would be " nova.tests. NovaManageTestCase" ?19:48
vishykpepple: the syntax is nova.tests.test_nova_manage19:49
vishyand you can omit nova.tests19:49
vishyyou can also do classes well19:49
kpepplevishy ironcamel : thx .. running now19:49
ironcamelcool, didn't know you could omit nova.tests19:49
vishy./ -N test_nova_manage:NovaManageTestCase.some_test_name19:50
vishyto run an individual test (note the colon between module and class name)19:50
vishyironcamel: that was a recent fix that went in to run_tests19:50
*** citral has quit IRC19:51
vishykpepple: running individual tests is a little dangerous without this patch though:
*** daveiw has joined #openstack19:53
openstackhudsonProject nova build #574: SUCCESS in 1 min 42 sec:
openstackhudsonTarmac: Removes processName from debug output since we aren't using multiprocessing and it doesn't exist in python 2.6.119:53
kpepplevishy: yeah, i actually implemented some stray record reaper code in my tests to get around it ...19:53
*** daveiw has quit IRC19:54
*** daveiw has joined #openstack19:54
*** daveiw has left #openstack19:55
*** daveiw has joined #openstack19:55
*** zul has quit IRC19:59
*** abecc has quit IRC20:04
vishyttx, soren, mtaylor: any packagers around?20:05
vishylvdisplay needs to be added to nova_sudoers20:06
*** joearnold has joined #openstack20:06
*** fabiand_ has left #openstack20:08
*** daveiw has quit IRC20:08
*** h0cin has quit IRC20:08
*** hadrian has quit IRC20:09
*** abecc has joined #openstack20:10
mtaylorvishy: whazzup?20:10
*** abecc has quit IRC20:11
justinsbAnyone else having problems with the bazaar server at the moment?20:12
mtaylorjustinsb: lemme check20:12
mtaylorjustinsb: seems _REALLY_ slow20:16
justinsbmtaylor: Yes.. that's what I'm seeing....20:16
justinsbmtaylor: OK, not just me then.  Thanks!20:16
*** jbryce has joined #openstack20:18
*** lamar has joined #openstack20:21
vishymtaylor: lvdisplay needs to be added to nova_sudoers (not sure where to report bugs related to packaging)20:21
edaysandywalsh: from your queue service email, by "broadcast queues", do you mean something different than the multi-cast queues?
*** brd_from_italy has joined #openstack20:24
*** Ryan_Lane has quit IRC20:25
*** deepheat_ has quit IRC20:25
devcamcarcreiht: got a sec for a swauth question?20:27
sandywalsheday, sort of. TTL is risky since I need to know that each worker got a copy of the message.20:29
*** hazmat has quit IRC20:30
sandywalsheday, consider the zone problem. Each service needs to report status back to each scheduler instance. Ideally these messages don't get lost.20:30
sandywalsheday, arguably we could say "well, one will eventually get there"20:30
sandywalsheday, so, in that case, does having a longer TTL really help? I suppose TTL could be a large number.20:31
*** nijaba has quit IRC20:31
*** lamar has quit IRC20:31
sandywalsheday, (hours or more)20:32
*** _cerberus_ has quit IRC20:32
*** hazmat has joined #openstack20:32
*** cw has quit IRC20:33
*** nijaba has joined #openstack20:33
*** nijaba has joined #openstack20:33
*** cw has joined #openstack20:33
*** _cerberus_ has joined #openstack20:34
*** miclorb has joined #openstack20:34
edaysandywalsh: Do the zone status messages really matter that they get there? If you send them out every minute or something, you just use the latest, no?20:35
edaysandywalsh: If you need guaranteed delivery, then you have two options. Longer TTL like you mention, or setup a worker that can replication to other queues (get from one, put to X). This is a common behavior so as a phase two we can add a "router api" hat basically configures these workers to do this type of replication (kind of like exchanges in amqp)20:36
sandywalsheday, Most likely it's fine. I was just wondering: how do I know that it's getting through at all?20:36
sandywalsheday, i.e. could it be that Scheduler 2 never gets updates from Compute 820:37
edaysandywalsh: broadcast has two distinct meanings in queues usually, anyone who can see it for a time period vs real replication20:37
*** dendrobates is now known as dendro-afk20:37
sandywalsheday, right, we don't want to clog the queues with messages no one is going to consume20:38
edaysandywalsh: well, this is something you would see from the scheduler monitoring, no? You could say "who are you active compute nodes", and if you don't see compute 8 on the list, it could be the queue (or compute 8 is down)20:38
*** rcc has quit IRC20:39
sandywalsheday, yup, I see your point. Not as big a deal as I first thought. If broadcast messages aren't getting to worker consistently we have bigger issues.20:40
*** dendro-afk is now known as dendrobates20:40
edaysandywalsh: think of it more as an event notification queue than a broadcast. more like an rss feed with TTLs that a replicated queue20:41
edaysandywalsh: but, if you suddenly have the ability to 'hide' messages and delete them, you get a traditional worker queue too20:42
sandywalsheday, right, it's not like a transaction where missing the message could mess up a back balance. Each event is an atomic snapshot of the state.20:42
edaysandywalsh: yup. and you can push deltas, and if a failure scenario means you're lost in the delta (have a way to detect), you can always request the full thing in a separate 1-1 queue20:43
*** KenD has quit IRC20:43
*** KenD has joined #openstack20:45
*** brd_from_italy1 has joined #openstack20:46
sandywalsheday, k, I'm sold. :) thx20:46
edaysandywalsh: cool, that was easy! :)20:48
*** brd_from_italy has quit IRC20:48
*** mgoldmann has joined #openstack20:49
*** lamar has joined #openstack20:55
*** KenD has quit IRC20:57
*** miclorb has quit IRC21:02
*** KenD has joined #openstack21:03
*** mahadev has quit IRC21:06
*** aimon has quit IRC21:07
*** mahadev has joined #openstack21:07
edayjaypipes: <321:07
*** dprince has joined #openstack21:07
*** jesse__ has joined #openstack21:08
*** te has joined #openstack21:09
*** jesse__ is now known as anotherjesse21:09
*** hadrian has joined #openstack21:11
*** miclorb has joined #openstack21:14
*** dubsquared has quit IRC21:14
uvirtbotNew bug: #724558 in swift "Memcache clients not handling memcache restarts correctly" [Undecided,New]
jarrodIOError: [Errno 13] Permission denied: '/dev/xvdb' <-- ughh21:18
*** KenD has quit IRC21:19
*** hvaldivia has joined #openstack21:19
jarrodit looks like the code i just trying to execute open() on that device name... since its running as user nova, how does it execute open as a superuser?21:19
*** KenD has joined #openstack21:20
jarrodbecause it's flags are wb for write access21:20
jarrodnova is in sudoers with NOPASSWD21:21
*** markwash has joined #openstack21:21
*** pothos has joined #openstack21:23
*** RobertLaptop has quit IRC21:24
jarrodunless nova-compute needs to run as root21:24
*** pothos_ has quit IRC21:25
vvuksanjarrod: i don't think so21:25
vvuksani am not sure what the xvdb stuff is21:26
jarrodi believe its the xen volume db21:26
jarrodbut its bombing trying to open the device with write privileges21:26
jarrodthe error is specifically Permission denied21:27
*** sandywalsh has quit IRC21:27
*** husimon has quit IRC21:27
jarrodwhy it would request write privileges on a device as a non-user is beyond me21:29
jarrodprobably a bug21:31
openstackhudsonProject swift build #206: SUCCESS in 28 sec:
openstackhudsonTarmac: ratelimiting does not handle memcache restart21:31
antonymjarrod: think you had to run it as root to get around it, jordanrinke ran into that prob21:31
*** husimon has joined #openstack21:32
jarrodok let me try that21:32
*** perestrelka has quit IRC21:33
*** perestrelka has joined #openstack21:33
jaypipesannegentle: one more branch to keep an eye on for docuemntation:
annegentlejaypipes: ok, good-o21:34
annegentletalking to tr3buchet now :)21:35
jarrodnova.conf needs --run_user=21:35
jarrodor i guess --nova_user to be consistent with rabbit_user21:36
*** jimbaker has quit IRC21:37
*** Dumfries has joined #openstack21:37
*** RobertLaptop has joined #openstack21:38
Dumfrieshaving  a issue with my scheduler in that it is unable to find a host. I get the error "raise driver.NoValidHost(_("No hosts found"))". I have manually attempted to run the scheduler in the foreground and specified the sql db to use but still get the same error. (all in one dev box).21:39
vishyDumfries: I noticed the same issue21:40
Dumfriesvishy: well glad I am not going crazy. Did you manage to find a workaround?21:40
vishynova-manage service list21:40
vishycheck to see if it is enabled with a :)21:41
vishyunfortunately the error is intermittent for me21:41
vishydoes it happen every time for you21:41
vishybattery is dying I'll try and look later21:41
*** sandywalsh has joined #openstack21:41
DumfriesAt the moment yeah, well at least since updating.21:41
*** miclorb has quit IRC21:41
Dumfries nova-scheduler enabled21:41
jarrodantonym: i ran as root and that fixed it21:43
*** miclorb has joined #openstack21:44
*** te has quit IRC21:45
*** te has joined #openstack21:46
jarrodnevermind, xen is getting No bootable device21:46
jarrodopenstack is not made to work with xenserver yet21:47
Dumfrieskvm that mutha21:48
jarrodive got a multinode kvm setup deployed and working GREAT21:49
jarrodi just wanted VM snapshots21:49
jarrodbut screw this it aint worth it21:49
jarrodatleast not with openstack at this point21:49
*** mgoldmann has quit IRC21:50
*** hazmat has quit IRC21:51
*** ctennis has quit IRC21:51
*** brd_from_italy1 has quit IRC21:51
*** miclorb has quit IRC21:53
*** hadrian has quit IRC21:53
*** anotherjesse has quit IRC21:54
*** ctennis has joined #openstack22:02
*** ctennis has joined #openstack22:02
*** hadrian has joined #openstack22:03
dprinceTwo approvals on this:22:04
dprinceCan we get that merged?22:04
*** miclorb_ has joined #openstack22:04
*** miclorb_ has quit IRC22:04
Dumfriesvishy: got it working22:13
kpepplejk0: i had to set lp:~ken-pepple/nova/dynamicinstancetypes back to "work in progress" as it no longer merged cleanly (migrate script numbering, testing issues,etc). will re-propose for merge later tonight.22:16
jk0cool, thanks for the heads up22:16
devcamcarjaypipes: pong!22:17
jaypipesdevcamcar: hey, any change of making django-nova support the openstack API?22:18
devcamcarjaypipes: no change yet22:18
devcamcarstill on the roadmap eventually22:19
devcamcari think there's still a bunch of stuff missing22:19
jaypipesdevcamcar: k, was just wondering. nothing more :)22:19
devcamcarlike being able to create projects/users22:19
jaypipesdevcamcar: missing from the OS API or ..?22:20
devcamcarif you look at nova/ you'll get an idea of what needs to be there22:20
devcamcarjaypipes: yes missing from OS API22:20
jaypipesdevcamcar: OS API will never create projects b/c projects make no sense in the OS API, though.22:20
*** MarcMorata has joined #openstack22:20
jaypipesdevcamcar: perhaps it is best to create a django-openstack-api instead?22:20
devcamcarjaypipes: well it has to be exposed somewhere22:21
devcamcarjaypipes: no, this stuff isn't really django specific22:21
*** vvuksan has quit IRC22:21
jaypipesdevcamcar: on the Openstack API side, there will always be only one project, because the concept of a project simply doesn't fit into the OpenStack API.22:21
devcamcarjaypipes: maybe the admin API is exposed using termie's API then22:22
devcamcarjaypipes s/is/should be22:22
edayjaypipes: for *rackspace* it may not make sense, but the openstack API should contain all functionality22:23
*** miclorb_ has joined #openstack22:23
edayjaypipes, devcamcar: plus I thought we were abstracting the idea of projects nd just making nested "accounts"?22:23
jaypipeseday: openstack API has no concept at all of "projects"22:23
edayjaypipes: understood, but it doesn't mean it can't :)22:24
jaypipeseday: yes, there was a ginormous thread on the ML about nested accounts.22:24
edayjaypipes: or some abstract construct that could be used in some deployments22:24
devcamcari didn't get the sense that any consensus was ever reached22:24
jaypipesI think I'll just make a django control panel for the openstack API side of things.22:24
dragondmjaypipes: projects will == account22:24
dragondmthus there will be 1 project per customer account22:25
*** MarkAtwood has quit IRC22:25
devcamcarjaypipes: most of what is in openstack-dashboard will convert quite well to openstack api, but there are things that have to fall into place first.  i'd rather see us be patient than to start a whole nother project22:26
jaypipesdragondm: when is that happening?22:26
*** hvaldivia has quit IRC22:26
jaypipesdevcamcar: what is openstack-dashboard? different from django-nova?22:26
dragondmI am adding the create projects (accounts)/ users apis right now22:26
dragondmjaypipes: right now22:26
jaypipesdragondm: ah, ok, sorry.22:27
devcamcarjaypipes: openstack-dashboard is a sample site built using django-nova.  i'm actually going to combine them (django-nova will still be a separate module, but it makes more sense to have these two projects in the same repo)22:27
dragondmit's part of the multitenant blueprint22:27
jaypipesdragondm: yes, I see that now. thx.22:27
*** omidhdl has joined #openstack22:28
devcamcarjaypipes: so whether we call it projects or accounts or whatever, there still needs to be an api endpoint that exposes the ability to create/delete accounts(projects?), start VPNs, etc.  and all that is currently exposed only in EC2 based admin api22:28
jaypipesdevcamcar: k, understood.22:28
*** MarcMorata has quit IRC22:29
devcamcarjaypipes: so if we get an OS or Direct api implementation of those methods, as well as support for things like sec groups and volumes in OS api, then we can switch over22:29
*** MarcMorata has joined #openstack22:29
jaypipesdevcamcar: not looking to switch over, just to have *something* useful for the OS API side of things...22:30
devcamcarjaypipes: i *am* looking to switch over though :)22:30
justinsbdevcamcar: Basic volume support for OS is already in one of my merge proposals.  But, guarded by a 'testing-only' flag until the spec includes volumes22:31
devcamcarcreiht: ping! :)22:34
*** RobertLaptop has quit IRC22:37
notmynamedevcamcar: creiht is out of the office for the rest of the day22:38
devcamcarnotmyname: thanks. i'm trying to get a bit of help with swauth.  when i do swauth-prep -K mykey it splodes22:39
kpepplea question about smoke-tests: i need to test nova-manage for some of the changes on the dynamic-instance-types branch. should these go in ?22:39
notmynamewhat error do you see?22:39
devcamcarnotmyname: Auth subsystem prep failed: 500 Server Error22:40
devcamcarand i'll pastie log bit to you in a sec22:40
*** tripson has joined #openstack22:40
*** mdomsch has quit IRC22:41
devcamcarfew things in there that look strange to me22:42
devcamcari'm not sure where/why its getting (though that should still work)22:42
devcamcari have provided my own value for default_cluster_url22:43
notmynamegholt: ^?22:44
notmynamedevcamcar: looking22:44
gholtThat's the default admin url.22:44
devcamcargholt: ahh ok22:45
gholtUse -A to set what you actually want if you want something different. ( swauth-prep --help )22:45
devcamcargholt: as far as i can tell that should be ok22:45
gholtOh, method not allowed22:46
devcamcarnotmyname, gholt: how do you guys smoketest a swift installation?22:46
gholtYou need allow_account_management turned on for the admin host.22:46
notmynameI knew he'd know the answer :-)22:46
devcamcargholt: # If set to 'true' any authorized user may create and delete accounts; if22:47
devcamcar           # 'false' no one, even authorized, can.22:47
*** ppetraki has quit IRC22:47
devcamcarthat doesn't sound like what i want22:47
devcamcari just want to be able to create accounts as the superuser22:47
gholtFor smoke testing, we usually just run the functional tests.22:47
devcamcargholt: the nose tests?22:48
gholtWell, they run via nosetests, but not the unit tests.22:48
notmynamedevcamcar: ./.functests22:48
gholtAnd that's what you want, allow_account_management = true. That let's .super_admin and any reseller admins (if you make any) create accounts.22:48
devcamcarok awesome22:49
gholtWe usually just run two (failover) machines with it turned on, with limited routing.22:50
uvirtbotNew bug: #724622 in nova "Openstack API Content Type negotiation " [Undecided,New]
*** bludadept_work` has quit IRC22:52
devcamcarnotmyname: where does .functests live?22:53
gholtThe root of the repo22:53's in the root of..yeah, what he said22:53
gholtYou can configure those tests via the /etc/swift/func_test.conf22:54
*** omidhdl has quit IRC22:54
gholtBut they are kinda tuned to how we setup SAIO dev environments.22:54
*** omidhdl has joined #openstack22:54
gholtYou have to pay special attention to the different users and their different access levels, as the tests will test that. :)22:54
uvirtbotNew bug: #724623 in nova "Creating servers with metadata fails" [Undecided,In progress]
*** omidhdl has quit IRC23:02
devcamcarnotmyname: hah, found a big problem, account-server was listening on port 602 instead of 6002.   oh, the little joys in life23:03
*** mahadev_ has joined #openstack23:06
*** mahadev has quit IRC23:06
devcamcarnotmyname, gholt: some progress. now i'm getting a different error:
devcamcara 40023:08
devcamcarand some rsync unhappiness: Feb 24 22:55:27 (hostname) proxy-server: ERROR with Account server (internal ip):6000/10 re: Trying to PUT to /v1/AUTH_.auth: Connection refused (txn: tx8c9e7760-39d4-4e8d-abb9-9217396f3eeb)23:08
*** glenc_ has joined #openstack23:09
*** jbryce has quit IRC23:09
*** gondoi has quit IRC23:10
gholt6000? Didn't you say your account servers were supposed to be running on 6002?23:10
devcamcargholt: account-server port is commented out, its using defaults23:12
*** glenc has quit IRC23:12
devcamcargholt: same for all of the services, all are using default ports23:12
gholtWell, that line above show it trying to connect to (internal ip):6000? Maybe the ring is whack?23:12
*** jc_smith has joined #openstack23:13
devcamcargholt: could be, checking that now23:15
devcamcargholt: oh yea, that's it bigtime. my fancy script builder is using 6000 there instead of 600223:15
uvirtbotNew bug: #724633 in nova "Incorrect method naming scheme for Auth Token database interactions" [Undecided,New]
gholtdevcamcar: I'm out for a bit, but I'll be on later, so feel free to post questions and I'll try to help when I get back. :)23:22
*** jc_smith has left #openstack23:26
*** acmurthy has quit IRC23:32
vishyDumfries: what was it?23:33
Dumfriesvishy looks like when the update was run compute was still running but had not been configured to use the new data store.23:34
*** mahadev_ has quit IRC23:36
*** mahadev has joined #openstack23:37
openstackhudsonProject nova build #575: SUCCESS in 1 min 43 sec:
openstackhudsonTarmac: adds live network injection/reconfiguration. Some refactoring.23:37
vishyah ok23:38
*** glenc_ has quit IRC23:39
devcamcargholt, notmyname: however when I curl I am getting a 412 precondition failed:23:39
devcamcarany ideas?23:41
*** glenc has joined #openstack23:41
*** allsystemsarego has quit IRC23:44
*** ovidwu_ has quit IRC23:48
*** vvuksan has joined #openstack23:52
*** dendrobates is now known as dendro-afk23:56
*** bcwaldon has quit IRC23:57
*** MarcMorata has quit IRC23:59
*** bcwaldon has joined #openstack23:59

Generated by 2.14.0 by Marius Gedminas - find it at!