Wednesday, 2010-12-22

dabotr3buchet: why is a method that returns a dict named '_entity_list'?00:04
*** Mussi has joined #openstack00:12
*** dfg_ has quit IRC00:18
*** joearnol_ has quit IRC00:25
*** maplebed has joined #openstack00:28
*** ar1 has joined #openstack00:33
*** kevnfx has joined #openstack00:35
*** sirp1 has quit IRC01:00
*** jesse_ has joined #openstack01:20
*** maplebed has quit IRC01:31
*** jesse_ has quit IRC01:39
*** Mussi has quit IRC01:49
*** reldan has quit IRC01:53
*** lvaughn_ has quit IRC01:57
*** jbryce_ has joined #openstack01:58
*** reldan has joined #openstack02:00
*** jbryce has quit IRC02:01
*** jbryce_ is now known as jbryce02:01
*** dragondm has quit IRC02:09
*** gaveen has joined #openstack02:14
winston-dHi, anyone there?02:18
winston-dI have a dump question about 'euca-bundle-images'02:18
winston-dhow can I bundle a Windows VM image?02:20
winston-dOr how can I bundle a image for HVM?02:26
*** reldan has quit IRC02:39
*** EdwinGrubbs has quit IRC02:46
*** EdwinGrubbs has joined #openstack02:50
*** jdurgin has quit IRC02:51
*** ccustine has quit IRC03:07
*** kashyapc has joined #openstack03:33
*** Xenith has quit IRC03:44
*** Xenith has joined #openstack03:48
*** jbaker has joined #openstack03:48
*** gaveen has quit IRC03:54
*** winston-d has quit IRC03:58
*** gaveen has joined #openstack03:58
*** Ryan_Lane has joined #openstack04:02
*** hadrian has quit IRC04:09
*** daleolds has quit IRC04:09
*** sirp1 has joined #openstack04:23
mtaylorttx: I have installed the claverifier plugin ... fingers crossed that it works04:29
*** joearnold has joined #openstack04:42
*** jbaker has quit IRC04:56
*** miclorb_ has quit IRC05:00
*** joearnold has quit IRC05:03
*** AimanA is now known as HouseAway05:08
*** joearnold has joined #openstack05:20
*** sirp1 has quit IRC05:22
*** joearnold has quit IRC05:35
*** f4m8_ is now known as f4m805:42
*** f4m8 is now known as f4m8_05:42
*** nijaba has quit IRC05:53
*** nijaba has joined #openstack05:54
*** nijaba has joined #openstack05:54
*** jbaker has joined #openstack05:56
*** kevnfx has quit IRC06:04
*** littleidea has joined #openstack06:10
*** joearnold has joined #openstack06:14
*** zaitcev has quit IRC06:17
littleideado services need to be restarted after you push out new rings or they will use whatever is in the filesystem?06:23
*** littleidea has quit IRC06:28
*** krish has joined #openstack06:30
*** krish has left #openstack06:31
*** f4m8_ is now known as f4m806:32
*** jbaker has quit IRC06:35
*** winston-d has joined #openstack06:44
*** anticw has quit IRC06:47
*** anticw has joined #openstack06:52
*** littleidea has joined #openstack06:57
*** guigui has joined #openstack07:02
*** ramkrsna has joined #openstack07:02
*** Abd4llA has joined #openstack07:09
*** littleidea_ has joined #openstack07:24
ttxzykes-: is planned for bexar. Might not make it though.07:27
*** _skrusty has quit IRC07:27
*** littleidea has quit IRC07:27
*** littleidea_ is now known as littleidea07:27
*** _skrusty has joined #openstack07:31
*** rcc has joined #openstack07:41
*** ArdRigh has joined #openstack07:46
*** miclorb has joined #openstack07:54
*** miclorb has quit IRC07:57
*** larstobi has quit IRC08:03
*** littleidea has quit IRC08:11
*** ibarrera has joined #openstack08:12
*** larstobi has joined #openstack08:12
*** doude has joined #openstack08:27
*** joearnold has quit IRC08:27
*** littleidea has joined #openstack08:28
*** littleidea has quit IRC08:28
*** miclorb_ has joined #openstack08:37
*** bolapara has quit IRC09:06
*** anticw has quit IRC09:20
*** irahgel1 has joined #openstack09:36
*** allsystemsarego has joined #openstack09:53
*** kashyapc has quit IRC09:54
fabiandzykes-: yes, kvm uses qemu ...10:00
fabiand(if the question wasn't answered yet :) )10:00
*** miclorb_ has quit IRC10:14
*** wedjat has joined #openstack10:22
wedjathello world10:23
*** anticw has joined #openstack10:24
wedjati followed the documentation "howtoinstallmultinode" but i can't get the last step working : uploading files to a container10:27
wedjati have a 404 Error not found, just as here :
wedjatseems that i need to create a container first in Swift, right ?10:27
*** littleidea has joined #openstack10:28
*** _skrusty has quit IRC10:28
*** littleidea_ has joined #openstack10:30
*** doude has quit IRC10:32
*** littleidea has quit IRC10:32
*** littleidea_ has quit IRC10:32
*** littleidea has joined #openstack10:33
olivier_Hi wedjat, seems there are all preparing christmas :-)10:36
wedjati should too10:41
*** _skrusty has joined #openstack10:42
*** Podilarius has left #openstack10:43
*** arcane has quit IRC10:45
*** calavera has joined #openstack10:46
*** littleidea has quit IRC10:46
*** arcane has joined #openstack10:50
*** reldan has joined #openstack11:05
*** kashyapc has joined #openstack11:07
*** doude has joined #openstack11:08
*** brd_from_italy has quit IRC11:19
*** larstobi1 has joined #openstack11:25
*** brd_from_italy has joined #openstack11:26
*** ArdRigh has quit IRC11:27
*** reldan has quit IRC11:35
*** krish has joined #openstack11:40
*** littleidea has joined #openstack11:52
*** littleidea has quit IRC11:56
*** krish has quit IRC12:15
*** reldan has joined #openstack12:17
*** ctennis has quit IRC12:18
*** larstobi has quit IRC12:30
*** ctennis has joined #openstack12:37
*** ar1 has quit IRC12:40
*** littleidea has joined #openstack12:45
*** hggdh has quit IRC12:56
*** Abd4llA has quit IRC13:06
*** WonTu has joined #openstack13:07
*** WonTu has left #openstack13:07
notmynamelittleidea: the proxy servers look at the mtime of the file and reload if necessary13:23
alekibangozykes-: kvm and qemu are very closely related, so the answer is yes13:25
alekibangokvm has sheepdog13:25
fabiandalekibango: do you know how stable/reliabe sheepdog is?13:27
alekibangoi didnt test it in production, but from the method used i would say it will be reliable13:28
alekibangoit makes sense for me13:28
alekibangoand the project is there for some time, its not fresh new13:28
alekibangoi would be very happy to have sheepdog for bexar13:29
alekibangoi dont believe there will be better solution for reliable storage soon13:29
*** hggdh has joined #openstack13:30
alekibango(talk to xen people to implement support...)13:30
*** kevnfx has joined #openstack13:31
fabiandMh yes. The design looks sane.13:31
alekibangohaving support for kvm/qemu, in future maybe for xen or uml -- thats good for me13:32
fabiand:) yep.13:33
*** westmaas has joined #openstack13:40
*** doude has quit IRC13:42
*** doude has joined #openstack13:43
*** doude has quit IRC13:44
*** doude has joined #openstack13:45
*** doude has quit IRC13:46
*** doude has joined #openstack13:46
*** doude_ has joined #openstack13:50
*** kevnfx has quit IRC13:52
*** doude_ has quit IRC13:52
*** hadrian has joined #openstack13:58
*** lvaughn has joined #openstack14:03
ttxjaypipes: another day, another i18n merge try.14:07
wedjathi everyone, did somebody succeed in applying the swift documentation for using multiple storage nodes ?14:07
jaypipesttx: heh, thx for the encouragement :)14:08
wedjatyes that is that one14:08
ttxjaypipes: I proposed my first branch for merging. Figured I had to contribute to the big pile of review-needed stuff14:08
wedjati succeeded in installing swift and did all the steps14:08
jaypipeswedjat: you asking whether that works?14:08
wedjatbut when i try to upload a file as indicated, i get an 404 error14:09
jaypipesttx: :P14:09
jaypipescreiht: around?  some help for wedjat  :)14:09
wedjatjaypipes: i want to know if it could be an error in the last step. Seems that it's needed to create a container before creating an object14:09
jaypipeswedjat: creiht, notmyname, redbo are all swift devs. they should be able to help :)14:10
*** mattt_ has quit IRC14:10
wedjatjaypipes: ok thanks :)14:10
fabiandis libivrt used in nova to controll guests?14:11
*** DigitalFlux has joined #openstack14:11
fabiand? :)14:11
jaypipesfabiand: it can be. so can xenapi.14:12
jaypipesannegentle: ping14:12
fabiandjaypipes, but xenapi can just be used for xen guests, right?14:13
fabiandis there a recommended hypervisor for nova?14:13
jaypipesfabiand: yes14:13
jaypipesfabiand: and no. :)14:13
fabiandso what does it mean?14:13
fabiandwhat do you mean14:13
jaypipesfabiand: lots of hypervisors are supported. we try not to make any "recommendations"14:13
jaypipesfabiand: in an effort to be vendor-neutral as much as possible.14:14
*** DubLo7 has joined #openstack14:14
fabiandah okay. but there is a (historical bias) towards ...?14:14
fabiandjaypipes, wonderful :)14:14
jaypipesfabiand: I responded "yes" to your question about xen guests for xenapi...14:14
fabiandah oaky. thanks for yes and no.14:15
jaypipesfabiand: :) no probs14:15
jaypipesfabiand: we also have msft hyper-v support contributed from
fabiandah interesting ...14:15
jaypipesfabiand: it should be going into trunk in the next few days.... if we ever get the code review backlog scaled down ;)14:15
fabiandso is there an internal "common hypervisor management interface abstarction layer" something like deltacloud?14:16
fabiandjaypipes, :) i've not heard about big hyer-v deployments ...14:16
fabiandbut good to know ..14:17
jaypipesfabiand: there is a layer within nova that abstracts the hypervisor layer.  not sure about deltacloud, perhaps jdarcy or pzaitcev could comment when they're on the chanel (both RHatters)14:17
*** daleolds has joined #openstack14:17
*** jdarcy has joined #openstack14:18
jaypipesjdarcy: ah, what nice timing :)14:18
jaypipesfabiand: feel free to ask jdarcy about deltacloud ;)14:18
fabiandjdarcy: good day jdarcy :) I just wondered if there is a hypervisor abstraction within nova and if it is comaprable to deltacloud.14:19
gholtwedjat: Several folks have run through those multi node swift instructions, but that doesn't mean something can't be improved. ;) Also, st will create the container for you with those upload commands, so probably it's the account that is not found. Check your syslog and see if you see errors.14:20
jdarcyfabiand: I'm not really the right person for that (I'm purely a storage guy) but my impression is that the answer is no.14:23
jdarcyfabiand: To clarify, I don't mean there isn't a hypervisor abstraction within nova, but that I wouldn't expect it to be comparable to deltacloud.14:24
*** krish has joined #openstack14:24
fabiandjdarcy: yes. I also think now that there is a difference14:24
jdarcyI'm trying to find out the status of a Nova back end for Deltacloud.14:26
fabiandjdarcy: it seems as if the nova api will be compatible with racksapces api, so the deltacloud rackspace driver can be used ..14:27
fabiandthere is an ml post about that14:27
jdarcyOK, I guess that makes things a lot easier.14:27
*** sirp1 has joined #openstack14:28
jdarcyThe down side of this API layering is that things tend to start at a "lowest common denominator" kind of level, but that should improve over time.14:28
*** DubLo7 has quit IRC14:29
fabiandyep. I also wonder how much features get lost when stacking up apis ...14:30
*** rogue780 has joined #openstack14:31
fabiandjdarcy: as you are a pure storage guy :), can you say something about a nice way to store guest images in a scalable and redundant fashion?14:32
jdarcyfabiand: Are you talking about "at rest" storage, or storage while actually in use?14:33
fabiandwhile actually in use ..14:33
*** reldan has quit IRC14:34
jdarcyI'm a pretty big fan of GlusterFS for that kind of thing.14:34
fabiandmh okay ...14:34
alekibangojdarcy: isnt sheepdog better?14:34
*** aliguori has joined #openstack14:35
jdarcyalekibango: Sheepdog and RBD are interesting alternatives and would probably be better for certain (common) use cases, but they're much more specialized.14:35
fabiandjdarcy: i see first commits for cloudfs :)14:35
alekibangobut i think thats what we ask for - live strorage for virtual guests14:35
alekibangofabiand: do you mean that kind of storage, right_14:36
jdarcyalekibango: In particular, they're both implemented at the qemu level, so they're not readily shareable with non-qemu code.  Similarly, since they're block devices you can't re-use the same infrastructure for shared file service (though RBD is part of Ceph which can do that).14:36
fabiandalekibango: yes, I am also talking about that.14:37
alekibangoso you think rbd/cepth will be having big future when it will be stable?14:37
alekibangoi would tend to think so14:37
jdarcyalekibango: Absolutely.14:37
alekibangoeven if i love simplicity of glusterfs too....14:38
jdarcyI think Ceph will be a better choice for those chasing raw performance, which is part of why I'm pushing CloudFS in other directions.14:39
alekibangojdarcy: some people here are testing rbd and will work on it as a backend for storage14:39
alekibangothey say its pretty stable... for them14:39
jdarcyalekibango: Here=Openstack, or here=where you are?14:39
alekibangohere - openstack14:39
alekibangohe is not here atm14:40
jdarcyWe definitely have customers who are using RBD now and are pushing us for official support.14:40
alekibangointeresting... when the ugly warning about experimental software will go away then?14:40
jdarcyNot for a while.  There's a large and often unappreciated gap between most people being able to run something without problems and the developers actually seeing the number/severity of bugs drop off.14:42
*** krish has left #openstack14:43
alekibangojdarcy: thanks for your insights.. you always help a lot14:43
jdarcyIt's what I come here for.  :)14:44
alekibangoso, would u use sheepdog for larger mass of virtual hosts? how reliable that might be?14:45
*** littleidea has quit IRC14:45
fabiandYes, thanks :)14:45
alekibangoi tend to think it should be the right tool atm14:46
alekibango(when i dont need xen)14:46
jdarcyalekibango: I really haven't been able to evaluate the reliability of sheepdog.  What I do know is that its scale is somewhat limited.14:47
alekibangoimho the size is still ok if you will use  clusters of <100 nodes14:48
alekibangoand we can divide cloud into clusters of that size14:48
alekibangoso that might not be a (big) problem for us14:49
jaypipesholy crap we have a lot of active code reviews to do...14:49
alekibangoyes, review my one please, its easy one14:49
*** ppetraki has joined #openstack14:54
wedjatgholt: thank you for your answer. But my command "stat" seems to work. I just have a problem with the command upload14:57
*** schisamo has joined #openstack14:58
gholtwedjat: Ah, okay. Anything telling in the syslogs?15:01
wedjatPUT /v1/AUTH_53de2aebc8cb4884a65497bf834bd398/myfiles HTTP/1.0 40415:04
wedjatjust the same 404 erros15:04
wedjatit tries to make a PUT to create a container, right ?15:04
gholtYet stat works? That's weird. That put is the container put.15:04
*** gondoi has joined #openstack15:05
wedjatyeah, stat works. My command upload gives me this :
wedjatmy command stat gives this :
wedjatwhich i guess is a correct output15:07
gholtThose are different users, but the same account. Hmm.15:07
gholtCan you try using the same user for both commands?15:08
wedjatyeah, sorry15:08
wedjatyes, i used the same user for the same commands15:08
gholtWell, the account string is different on those two pastes is why I wonder what's up?15:08
gholtOne starts with AUTH_ce and the other AUTH_5315:09
wedjati recreated the accounts since that morning :(15:12
wedjatbut i still have this problem15:12
wedjatwith the same account15:12
*** gaveen has quit IRC15:13
wedjathere are my logs when i try "upload" :
jaypipesalekibango: do a quick merge with trunk...15:16
jaypipesalekibango: there was a conflict in process_unittest.py15:16
alekibangook thats why you should merge fast... i will look on it15:17
alekibangokeeping it in limbo for week will not help15:17
*** hggdh has quit IRC15:18
jaypipesalekibango: yeah, understood. :) but we've had some very big patches to review...15:18
jaypipesalekibango: just be thankful you're not me, and having to re-merge trunk a dozen times for my i18n branch ;)15:18
alekibangoi know, np15:19
alekibangoi am :)15:19
*** hggdh has joined #openstack15:21
gholtwedjat: I'm just not sure what went wrong where. When you do a stat right now on the account, it works? You can see the successful HEAD in the logs? Yet you see the failed 404 PUT on the same account? Without being able to log in myself and poke around, I'm just not sure what went wrong.15:22
*** jbaker has joined #openstack15:23
*** JuanPerez has joined #openstack15:23
gholtwedjat: I hate to give that kind of response. I hate getting that kind of response worse. ;) But I'm at a loss.15:23
wedjati understand you, don't worry :)15:24
wedjatwell, i am not pretty sure of what i should really see after the curl commands you know15:24
*** DigitalFlux has quit IRC15:25
gholtWell, that st stat command just does the same HEAD as the above step (steps 3 and 4)15:25
wedjati have a 204 http code, so i guess that's ok15:27
alekibangojaypipes: the file is gone in trunk.  also those tests...15:28
*** aliguori has quit IRC15:29
jaypipesalekibango: when you bzr merged, did it say "conflicts"?15:29
alekibangoi didnt use merge yet, just looking15:29
alekibangobtw nova should switch to git :)  bzr is so strange to me still15:30
alekibangoyes 1 conflict...15:31
*** DigitalFlux has joined #openstack15:33
jaypipesalekibango: just because you aren't familiar with bzr doesn't mean git is better... *sigh*...15:34
alekibangoheh, there are some clues that it might be... esp. for larger projects15:34
jaypipesin other news, /me finally found the problem with i18n-strings...15:34
alekibangobut both are still awkward to operate15:34
* jaypipes slaps soren...15:34
dendrobatesjaypipes: what was it?15:34
jaypipesdendrobates: utils.runthis(_("Generating root CA: %s", "sh"))15:35
jaypipesdendrobates: should be utils.runthis(_("Generating root CA: %s"), "sh")15:35
dendrobatesthat is it, just a missing parens!15:35
jaypipesdendrobates: wasn't showing up locally because it was stubbed out with a fake on my box.15:36
jaypipesdendrobates: yes, just a missing parens. :)15:36
jaypipesdendrobates: or rather, a parens in the wrong place :)15:36
dendrobatesI feel stupid for not catching that especially because the ca tests were failing15:36
dendrobatesI just assumed it was more complex15:37
alekibangomissing parents is sad15:37
ttxjaypipes: \o/15:37
* ttx celebrates by having some food.15:37
*** smaresca is now known as icarus90115:38
wedjatshould a proxy server be installed on a storage node ?15:38
wedjator at least have a mount point too ?15:38
*** abecc has joined #openstack15:39
*** dfg_ has joined #openstack15:39
gholtWell, should is a loaded word. ;) But yes, they can exist on the same server no problem.15:39
jaypipessoren: and btw, it wasn't your fault of course, I just felt like slapping you. :)15:39
wedjatthanks for your answer :)15:39
jaypipesttx, dendrobates: ok, I'm going to approve the i18nstrings branch... again. fingers crossed :)15:40
*** rogue780_ has joined #openstack15:44
*** rogue780 has quit IRC15:44
jaypipesalekibango: got that merge fixed up?15:44
*** johnpur has joined #openstack15:45
*** ChanServ sets mode: +v johnpur15:45
alekibangonono, got kids here :),  but i will soon15:45
jaypipesalekibango: hehe, no worries :)15:47
wedjatin the doc, "swift-ring-builder account.builder add z<ZONE>-<STORAGE_LOCAL_NET_IP>:6002/<DEVICE> 100", what should be equal DEVICE ? /dev/sdb1 for example, or just sdb1 ?15:47
*** Ryan_Lane has quit IRC15:47
sandywalshany citrix people in the house?15:49
jaypipesttx, dendrobates: i18n-strings is merged! \o/15:50
jaypipessandywalsh: typically, they are not, no...better able to get in touch with them on the ML.15:51
jaypipessandywalsh: something about their VPN and IRC/freenode IIRC...15:51
gholtwedjat: Just the sdb1 part.15:51
wedjatthanks gholt :)15:51
wedjatbut sdb1 is just a mount point into the directory specified by the "devices" resource, right ?15:52
sandywalshthx jaypipes15:52
wedjatand you could use for example /srv/node/sdb1 as a mount point for a filesystem ?15:52
gholtYes, that's true. So the object server, when it gets a request for sdb1 will prepend your devices conf value.15:52
*** irahgel1 has quit IRC15:52
gholtSo, in the case you mention, the conf should be devices = /srv/node15:53
openstackhudsonProject nova build #315: FAILURE in 12 sec:
*** irahgel1 has joined #openstack15:54
*** kevnfx has joined #openstack15:56
dendrobatesttx jaypipes: so now we need .pots files right?15:59
dendrobatesI assume there is a tool for pulling out the strings.15:59
jaypipesdendrobates: ya. I'll work on that later with monty...16:01
*** aliguori has joined #openstack16:01
jaypipesdendrobates: see email to ML I just sent..16:01
ttxjaypipes: openstackhudson> Project nova build #315: FAILURE in 12 sec: ?16:03
dendrobatesI fixed it.  The last change needed an approval16:03
dendrobatesthe bot is delayed16:03
*** ramkrsna has quit IRC16:04
*** f4m8 is now known as f4m8_16:05
*** schisamo has quit IRC16:06
jaypipesttx, dendrobates: ya, fixing that right now...16:06
*** seshu has joined #openstack16:06
jaypipesttx: have to make a change to the doc builder...16:06
sirp1ttx: ill be adding a status update to xs-snap blueprint shortly16:13
*** sophiap has quit IRC16:13
*** sophiap_ has joined #openstack16:13
ttxsirp1: cool, thanks16:13
* ttx is about to abndon the idea of setting up LP locally to fix it16:14
jeremyblp == launchpad?16:15
ttxjeremyb: yes16:15
jeremybmaybe, you're talking about a new ssh key16:15
jeremyb(i guess you're not deploying your own lp clone)16:16
*** joearnold has joined #openstack16:16
ttxjeremyb: I am trying to run a LP clone locally to test some of my fixes.16:16
jeremybyou're patching lp itself?16:16
ttxjeremyb: unsurprisingly, it's a bit difficult.16:16
* jeremyb heard lp was dfsg free but not the build scripts or some other auxilliary stuff needed to set up or run it16:17
ttxjeremyb: yes, it fails to expose some data I need :)16:17
ttxjeremyb: that's not true.16:17
* jeremyb heard in person from someone that wasn't entirely sure16:18
jeremyband wasn't looking at it right then16:18
*** dragondm has joined #openstack16:18
*** joearnold has quit IRC16:18
jeremybanyway, need to head to work16:18
*** schisamo has joined #openstack16:22
*** Ryan_Lane has joined #openstack16:23
*** aliguori has quit IRC16:23
uvirtbotNew bug: #693464 in nova "Double dots in the API docs are causing broken links on" [Undecided,New]
*** kashyapc has quit IRC16:34
*** calavera has quit IRC16:36
jk0would someone mind reviewing this for me please?
dubsa branch i'm working on contains a few scripts that do not seem appropriate for any of the existing top level nova directories, i'm wondering where to put them.16:40
dubsi have a python script that needs to be installed on xenserver hosts to apply networking restrictions when instances are started and stopped, and a init script for xenserver hosts.16:41
dubsalso, there is a small patch file that has to be applied to one of the pre-installed xenserver init scripts16:41
edayvishy: so, looking at your project-vpn branch... there is no rest API into cloudpipe now?16:41
jk0could probably go in the plugins dir16:41
uvirtbotNew bug: #693467 in swift "swift i18n conversion" [Undecided,New]
*** hadrian has quit IRC16:44
antonymjk0: yeah, that's where we put them now16:47
antonymjk0: they are xenserver specific, so we created a xenserver directory under plugins, they didn't exactly pertain to xenapi so we left them in a seperate directory16:48
jk0yeah, I figured that would be the place for them16:48
dubsdoesn't seem ideal though, nothing about any of the scripts are "plugins"16:49
jk0it's the most ideal place in the existing tree, and already being used16:50
*** kashyapc has joined #openstack16:50
uvirtbotNew bug: #693471 in glance "SQLAlchemy requires date column to be a datetime object" [High,New]
jk0dubs: maybe we could rename the plugins dir to something more generic16:51
jk0probably have to get approval first though16:52
antonyma hypervisor specific directory might make sense, as these are things that need to be installed to add additional value to them16:53
dubsor even a subdir in contrib?  contrib/xenserver/ maybe16:53
antonymthen under than we can put kvm,xen,etc specific items16:53
jk0I'm thinking just renaming the plugins dir to something more generic, and sticking all of the xenserver stuff in the xenapi dir16:53
jk0that way each hv can have it's own dir inside of <whatever it's renamed to>16:54
antonymwould probably make more sense to rename xenapi to xenserver then16:54
jk0whatever happens, I think we should keep everything under one subdir and not accross multiple (plugins, contrib, etc)16:54
jk0one subdir that has one tree under it for easy scping16:54
antonymyeah, i'd rather not spread it out all over the place16:55
*** nelson__ has quit IRC16:55
*** nelson__ has joined #openstack16:55
*** JuanPerez_ has joined #openstack16:55
*** brd_from_italy has quit IRC16:55
jk0in theory the plugins dir is a contrib, so maybe we could just move plugins/xenapi to contrib/xenserver and rid of plugins16:55
antonymwell the xenapi plugin is required for nova to work, whereas the networking protections aren't required16:57
*** JuanPerez has quit IRC16:57
*** JuanPerez_ is now known as JuanPerez16:58
jaypipesantonym: since when is the xenapi plugin required for nova to work? :)16:58
jk0yeah, it's not deployed automatically16:58
ttxjaypipes: reading your i18n mail... should we try to catch missing _()s in pre-merge testing ?16:58
antonymnova w/ xenserver ;)16:58
jk0still has to be done manually16:58
jaypipesttx: as much as possible...but it's not something that will break things if they are missed.16:59
jk0ttx: I did for mine, otherwise there are conflicts16:59
ttxjaypipes: is it something we can automate in any way ?16:59
jaypipesttx: eh, well, we could...given the time to write the script, sure.16:59
ttxjaypipes: ok17:00
*** ibarrera has quit IRC17:02
sandywalshjk0, any tests for that branch?17:05
*** ccustine has joined #openstack17:05
jk0sandywalsh: not yet, waiting on the Citrix guys' branch to merge17:06
jk0there have a huge test suite lined up17:06
wedjatbtw, i solved my problem with the 404 error : it came from a misusing of ports in the ring construction17:06
sandywalshjk0, the new xenserver tests?17:06
wedjatthanks to people who helped me :)17:06
wedjatand sorry :p17:06
*** hadrian has joined #openstack17:07
sandywalshjk0, and how about if _poll_task() throws an exception? Does it still get logged?17:07
sandywalshjk0, does it make it past send_exception() ?17:07
openstackhudsonProject nova build #316: STILL FAILING in 10 sec:
jk0dunno, I'd have to check17:08
*** hadrian has quit IRC17:08
gholtwedjat: Hehe, awesome that you figured it out! :)17:12
sandywalshjk0, InstanceDiagnostics was pre-existing I guess?17:13
jk0it was from a previous merge17:13
sandywalshjk0, are all operations async? (that is, do they all go through _poll_task)? If not, how do non-async tasks get logged?17:14
jk0anything that doesn't go through _poll_task does not get logged at the moment17:15
jk0baby steps17:15
jk0most of the existing ones do though17:15
sandywalshjk0, but is InstanceDiagnostics part of your bp, or was it done separately?17:15
jk0it's part of the BP17:16
*** cw_ is now known as cw17:16
sandywalshI don't see it getting called anywhere?17:16
jk0it's not yet17:16
jk0I'm submitting these in small chunks17:16
sandywalshso is this a merge prop or just a review?17:17
jk0it's a merge prop17:17
*** maplebed has joined #openstack17:17
sandywalshdo we want to merge if it's not complete?17:17
vishyeday: correct17:17
jk0sandywalsh: yes17:17
vishyeday: it was a huge security hole to be signing stuff that way17:17
jk0I asked a while back if this was acceptable, and it is17:17
sandywalshI'm surprised17:17
edayvishy: ahh, ok17:17
edayvishy: conflicts on it during merge, btw :)17:18
vishyoh the i8n doh17:18
vishyprobably have to remerge all of my branches17:18
wedjatgholt: yeah, it is perfectly working now :)17:19
wedjatbtw, is it possible to create an object without any container associated ?17:19
gholtwedjat: Ah no. it kinda need sto be there /v1/account/container/object is the url pattern.17:19
wedjatbecause i uploaded 3 files, i can download these 3 files but a stat command give me only two objects :s17:20
*** cclien has quit IRC17:21
jk0sandywalsh: I see what you mean about _poll_task, maybe the logging should happen outside of the try17:21
jk0is that what you were referring to?17:21
sandywalshjk0, just updated ticket with comment on it17:23
jk0cool, thanks17:23
vishyand all of the help text for defines needs _() as well17:24
openstackhudsonYippie, build fixed!17:24
openstackhudsonProject nova build #317: FIXED in 1 min 13 sec:
jk0sandywalsh: I'll get these worked in17:24
sandywalshjk0, make sense?17:24
jk0yeah, definitely17:24
jk0I get what you mean about the test17:24
jk0I just didn't want to roll my own xenserver test suite when the citrix guys  already have it going17:25
sandywalshand if the logging fails, should the op fail?17:25
jk0my thoughts were to just add them to their existing tree when it lands17:25
sandywalshI think just stub xenapi out for your test17:25
jk0no, that shouldn't happen17:25
*** sophiap_ has quit IRC17:29
*** Ryan_Lane is now known as Ryan_Lane|food17:32
*** fabiand has quit IRC17:33
*** Ryan_Lane|food has quit IRC17:35
*** fabiand has joined #openstack17:36
jk0eday: I'm thinking we might want to create a new blueprint for the distributed data stuff17:38
vishyeday: did you see my comment on ec2-lockout branch? Any ideas?17:40
edayjk0: we have one for the core data.. the bigger thing here is defining which interface this should be going through17:41
vishyeday: also fixed conflicts in project-vpns17:41
edayie, core operational data vs logging/audit/monitoring data17:41
*** reldan has joined #openstack17:43
edayjk0: also, the blueprint for the diagnostics stuff doesn't go into much detail, so it's hard to approve stuff in small chunks when you can't picture the end result.17:43
edayjk0: describing things like what other systems/people will be accessing this data. That will help answer what type of interface it should be going through17:44
reldanvishy: Hi, vishy. We found commented class Host in We want to create Availability_Zone and Host model in and link them together.17:46
jk0well, I know how Rackspace wants to do it, but that might not apply to others17:46
jk0I'm open to anything but my only concern is getting the blueprint done in time for Bexar17:46
reldanvishy: Is it the right way?17:46
edayjk0: I'm sure this is good for the project is a whole, I just don't know the best way to go about it yet17:47
vishyreldan: debatable, we have the Service class as well17:47
Adiantumhello gents17:48
vishyreldan: is there a reason you need Availability_Zone to be a whole model?17:48
jk0eday: I can keep going with it for now (making it accessible via API and wrapping up other loose ends) and maybe we can revisit the data store when all of that is in place? might help give the bigger picture17:49
vishyreldan: as a hacky first version I would probably just add 'zone' to the Service Model17:49
vishyjk0: if you keep a clean abstraction for the api for getting the data, it will make it easier to pull out the collection and aggregation layer17:50
Adiantumvishy: and is there any way to determine that zone is not available if it will be in service model?17:51
reldanvishy: We have several options and try to choose the best one.17:51
vishyAdiantum: services can be disabled/enabled17:51
*** anthrax03 has joined #openstack17:51
vishyAdiantum: do you think zones should be able to be marked as well?17:51
*** jesse_ has joined #openstack17:52
vishyAdiantum: it seems like you could just do the equivalent of: update services set disabled=1 where zone='xxxx'17:52
Adiantumvishy: i mean that we may be crashed. is there any mechanism to determine that service is not working (but still enable)?17:53
edayjk0: well, then we may have built something that isn't the correct way. :)  If there are more design details in the spec that we can discuss first, it might be easier to model it correctly the first time17:53
jk0the good news is there isn't much for specs, bad news is we need to get this figured out right away17:54
*** mtaylor is now known as mtaylor|afk17:56
*** sophiap has joined #openstack17:56
vishyAdiantum: yes, service report in every 10 seconds17:56
Adiantumvishy: aha, thank you. seems it's good for us17:57
reldanvishy: Adiantum: Thanks. Looks goog!17:57
jk0eday: I don't know enough about sqlalchemy to answer this, so would it make sense to maybe just create a separate db, nova-something, and start putting these tables in there instead of the core nova db?17:58
vishythe branch here:
vishyadds a nova-manage command to show health of services17:58
perestre1kanvm sorry ;)17:58
vishyand a euca hack for admins: euca-describe-availability-zones verbose17:58
edayjk0: it's hard to answer any of these questions without more info on the bigger picture, and it's more about interfaces than where to actually store the data17:59
jk0eday: okay, taking a step back, what do we need to do to get the bigger picture? I'm not sure I know it well enough myself (at least at the level you are probably thinking)18:01
edayjk0: in the spec, just a few use cases.. if it's going to be available via a rest API, just a mockup of what that may look like to make sure things are being stored properly18:04
jaypipes_cerberus_: you see you have a small merge conflict detected on os_api_refactor?18:04
edayjk0: also, the big Q is does anything else inside of nova need to use this in the future, or is it really just logging data18:04
jk0ok, I'll update that with my point of view and see what you think18:04
jaypipes_cerberus_: in /nova/api/openstack/__init__.py18:04
edayjk0: if not, I'd argue introducing another interface, and not going directly into the core db18:04
edayfor a dployment, it may end up in the same place, but at least have a different API so it could be stored somewhere else too (like auth api works)18:05
jk0I see what you mean - I'll write up my thoughts on it and have you take a look18:05
*** reldan has quit IRC18:06
*** jdurgin has joined #openstack18:09
_cerberus_jaypipes: I did not18:10
_cerberus_I'll take a look. Thanks18:10
*** JuanPerez has quit IRC18:14
*** aliguori has joined #openstack18:15
*** johnpur has quit IRC18:22
*** reldan has joined #openstack18:22
jk0eday: want to have a look at ?18:23
openstackhudsonProject nova build #318: SUCCESS in 1 min 15 sec:
*** Ryan_Lane|food has joined #openstack18:26
openstackhudsonProject nova build #319: SUCCESS in 1 min 17 sec:
*** rogue780_ is now known as rogue78018:31
*** guigui has quit IRC18:37
*** anthrax03 has quit IRC18:38
*** reldan has quit IRC18:38
*** kashyapc has quit IRC18:40
*** reldan has joined #openstack18:45
*** Ryan_Lane|food is now known as Ryan_Lane18:45
*** Xenith__ has joined #openstack18:46
*** dendrobates is now known as dendro-afk18:48
*** dendro-afk is now known as dendrobates18:48
*** tmarble has quit IRC18:49
jaypipessirp1: hola :)18:55
*** joearnold has joined #openstack18:55
sirp1jaypipes:  howdy18:55
jaypipessirp1: I decided to do some doc cleanup on glance today...brain was frying this morning with auth stuff :)18:56
jaypipesmtaylor|afk, annegentle: when you get a chance, would be great to setup with doc published from Hudson like swift/nova :)18:57
sirp1jaypipes: cool, will check it out18:57
*** deshantm has quit IRC18:58
*** deshantm has joined #openstack19:00
* jaypipes loudly applauds antonym and dubs for actually including quality documentation for multi-tenant networking protections. w00t.19:01
dubshaha, thanks.  that's mostly ant work :)19:04
*** sophiap has quit IRC19:04
*** sophiap has joined #openstack19:04
*** dendrobates is now known as dendro-afk19:05
*** dendro-afk is now known as dendrobates19:07
antonymit probably still needs additional docs, so i may expand more later on19:10
antonymcould use some items on recompiling the xenserver kernel which i'm sure citrix would love me to add heh19:10
*** Icarus81 has joined #openstack19:12
*** kevnfx has quit IRC19:12
*** hadrian has joined #openstack19:17
*** Icarus81 has quit IRC19:20
rogue780Hi. I followed the directions to install nova from from December 9th. When I try to run the instance, I can't get it to go past "pending" status.19:24
*** jesse_ has quit IRC19:27
*** chmouel has left #openstack19:28
*** chmouel has joined #openstack19:28
*** jesse_ has joined #openstack19:32
*** jesse_ has joined #openstack19:33
*** sophiap has quit IRC19:36
*** tmarble has joined #openstack19:38
vishyrogue780: that is usually a hypervisor related problem, are there any errors in the logs?19:40
*** gondoi has quit IRC19:41
rogue780vishy, I'll tell you in a second, the box is rebooting19:42
devcamcarhey all19:43
devcamcari have a packaging question relating to the django dashboard that we are releasing for bexar19:44
devcamcarit has a dependency on nova/adminclient.py19:44
devcamcarthe adminclient is a bit of an odd beast, given that it is currently part of the project but has no dependencies on nova itself19:44
devcamcarit really is its own standalone piece and should be distributed as such19:44
vishyrogue780: you will need 64bit proc with virtualization support turned on in the bios and 64bit os for the default instructions to work.19:45
devcamcari essentially need to do this in the django bits:  from nova import adminclient19:45
devcamcarhowever in order to do that now, you have to have all the dependencies from nova satisfied, including gflags, etc, even though the adminclient needs none of it19:45
devcamcarwe've been manually copying this file around in our prod environment, which works but is hackish19:46
devcamcarwe need a more formal way to distribute the adminclient, anyone have thoughts on the best way to do that?19:46
rogue780vishy, all the log files in /var/log/nova are empty19:47
vishyapparently logging hasn't been fix since we switched to upstart, not so useful19:48
vishycan you check to see if you have anything in /var/lib/nova/instances19:48
*** gondoi has joined #openstack19:48
annegentledevcamcar: so there are a lot of pre-reqs, plus we need to find a good way to distribute the file itself? (files?)19:49
vishyrogue780: the three comments I made above were supposed to be directed to you19:49
rogue780vishy, instance-614007856 is in there19:50
vishycool, so first check /var/log/libvirt/qemu/nstance-614007856.log and see if it exists and has any errors in it19:51
vishyer that was supposed to be instance-19:51
vishymissed an i19:51
annegentledevcamcar: we've got an install script (production-style) for Nova, maybe we can add an option to copy the adminclient into the environment with that script?19:54
*** sophiap has joined #openstack20:00
devcamcarannegentle: actually i think the adminclient should be its own isolated module20:01
*** sophiap has quit IRC20:01
devcamcarannegentle: it is essentially a tiny version of admin only api boto20:01
devcamcarit is a client to nova's server20:02
*** sophiap has joined #openstack20:02
devcamcari'd love to just be able to pip install novaadminclient20:02
rogue780vishy, /var/log/libvirt/qemu/ is empty20:05
annegentledevcamcar: ah, okay. Is that just a packaging task then?20:05
vishyso it broke before the hypervisor20:06
annegentledevcamcar: I find it to be similar to Mike Mayo's iOS offering, and the install experience there could be better20:06
devcamcarannegentle: yes, so far I think the nebula dashboard is the only thing that is really using it, but since we're about to release that we'll need to deal with this soon20:06
rogue780i'm going to reboot one more time20:06
vishycan you do a file /var/lib/nova/instances/instance-614007856 and post results in a pastie?20:07
rogue780in just a minute or two20:07
openstackhudsonProject nova build #320: SUCCESS in 1 min 15 sec:
*** olivier_ has quit IRC20:10
rogue780vishy, which file do you want from that directory?20:12
vishyi just want to no types and sizes20:12
*** olivier_ has joined #openstack20:18
*** HouseAway is now known as AimanA20:18
vishythat appears to be correct20:19
vishycan you try a virsh create libvirt.xml from inside that folder?20:19
openstackhudsonProject nova build #321: SUCCESS in 1 min 14 sec:
rogue780$ virsh create libvirt.xml  \\ error: Failed to create domain from libvirt.xml \\ error: internal error no supported architecture for os type'hvm'20:21
*** rcc has quit IRC20:21
*** sophiap has quit IRC20:22
devcamcarso who is our resident packaging expert?20:22
vishythere you go20:22
vishyno virtualization for you20:22
vishyyou either have virtualization turned off in the bios or your proc doesn't support it20:23
edaydevcamcar: soren and mtaylor|afk20:25
devcamcareday: thanks20:25
annegentledevcamcar: Hm. I want to say ttx can coordinate it, and Soren and Monty have the know-how20:25
devcamcari will track them down soon20:25
annegentledevcamcar: heh I type slower than eday20:25
*** reldan has quit IRC20:26
devcamcarannegentle: whats best way for me to queue this? shall i open a new bug on lp?20:26
rogue780vishy, would loading something like esxi potentially fix this problem? please note, I have almost no idea what I'm talking about when it comes to virtualization... I was just kind of thrown into this project by my boss20:26
annegentledevcamcar: I'm not sure, it's not another blueprint, but it's a task for the existing blueprint. I'd say, update the spec for now, to add packaging as a task, and also check with ttx tomorrow.20:27
edayrogue780: run: egrep -c '(vmx|svm)' /proc/cpuinfo20:28
devcamcarannegentle: thanks, i'll follow up with ttx tomorrow20:29
rogue780eday, it returns 020:29
*** pothos has quit IRC20:30
edayrogue780: ok, that confirms vish's comment. might want to make sure your cpu model has virt extensions, and if they do, check your bios to make sure it's enabled20:30
*** jesse_ has joined #openstack20:30
edayrogue780: modern cpus have extra instructions that can optionally be enabled for hardware virtualization, and you don't have them currently (they would show up in flags under /proc/cpuinfo20:31
rogue780eday, thanks. Now I need to find out if it is a bios setting or an actual hardware limitation20:33
vishyrogue780: in the mean-time, you can also test using qemu20:33
edayrogue780: cat /proc/cpu and google your cpu module #20:33
edayerr, /proc/cpuinfo20:34
vishyby editting the flagfile to have --libvirt_type=qemu20:34
rogue780will do. we just had a power surge and may have lost some equipment here. I'll be back later. Thanks for the help, gentlemen/gentleladies.20:35
vishyit will be very slow though :)20:35
annegentlerogue780: doh! weather or squirrel?20:36
annegentlerogue780: good luck, thanks for trying out OpenStack :)20:37
rogue780ok. nothing fried20:38
*** dendrobates is now known as dendro-afk20:38
annegentlerogue780: whew20:39
jesse_rogue780: surprisingly many vendors send boxes with hardware virtualization disabled by default20:40
*** joearnold has quit IRC20:40
rogue780looks like the cpu on this box is a celeron 2.66ghz...i doubt that has any awesomeness in it20:41
rogue780vishy, where is the flagfile where I can change it to use qemu?20:44
*** brd_from_italy has joined #openstack20:44
vishyrogue780: good question, I'm not sure where the packages install it currently20:44
vishyperhaps /etc/nova/nova.conf ?20:44
rogue780yup. just found it20:44
rogue780what services do I need to restart now?20:45
*** dendro-afk is now known as dendrobates20:46
*** miclorb_ has joined #openstack20:49
rogue780it says it's running now, but when I try to ssh to it it says no route to host20:52
vishyso apparently launchpad isn't smart enough to look for new conflicts in proposed branches when trunk changes20:55
rogue780well, folks, I've gotta run. I'll be back tomorrow to pick your brains some more. thanks so much for the help!20:55
vishyrogue780: probably a dnsmasq issue20:55
vishytry a killall dnsmasq then restart nova-network20:56
vishythen do a euca-get-console-output to see20:56
vishyif the instance is reporting getting an address20:56
vishyoh, also possible that it is due to metadata server being broken20:57
vishypatch went in today for that20:57
rogue780how soon should it hit the ubuntu repo?20:58
rogue780it just hangs on euca-get-console-output20:59
*** ctennis has quit IRC21:01
rogue780well, I do really have to go now. I've got an appointment up north of baltimore in an hour. I'll be back tomorrow.21:01
rogue780thanks again for the help21:01
*** westmaas has left #openstack21:01
*** westmaas1 has joined #openstack21:05
*** rogue780 has quit IRC21:06
*** gondoi has quit IRC21:06
*** joearnold has joined #openstack21:10
*** larstobi has joined #openstack21:16
*** daleolds has quit IRC21:19
*** daleolds has joined #openstack21:20
*** enigma has joined #openstack21:27
*** MarkAtwood has quit IRC21:29
openstackhudsonProject nova build #322: SUCCESS in 1 min 16 sec:
*** herki has quit IRC21:48
*** herki has joined #openstack21:48
*** kevnfx has joined #openstack21:50
*** ctennis has joined #openstack21:51
*** ctennis has joined #openstack21:51
*** kevnfx has quit IRC22:02
openstackhudsonProject nova build #323: SUCCESS in 1 min 16 sec:
*** jordandev has joined #openstack22:08
*** brd_from_italy has quit IRC22:12
tr3buchetanyone know if the xenserver power states are documented somewhere? trying to find a good list22:16
*** allsystemsarego has quit IRC22:16
*** jordandev_ has joined #openstack22:19
*** jordandev_ has quit IRC22:20
*** jordandev has quit IRC22:20
*** Alex111 has joined #openstack22:20
*** MarkAtwood has joined #openstack22:24
*** enigma has quit IRC22:37
*** littleidea has joined #openstack22:37
*** enigma has joined #openstack22:39
*** jkakar has quit IRC22:39
*** enigma has quit IRC22:41
*** DubLo7 has joined #openstack22:43
*** MarkAtwood has quit IRC22:52
*** littleidea_ has joined #openstack22:54
*** MarkAtwood has joined #openstack22:54
*** jkakar has joined #openstack22:55
*** sophiap has joined #openstack22:56
*** littleidea has quit IRC22:58
*** littleidea_ is now known as littleidea22:58
*** joearnold has quit IRC23:02
*** schisamo has quit IRC23:04
*** jkakar has quit IRC23:05
vishyI really appreciate the code reviews guys.  I still have two bugfix branches outstanding if anyone has time23:07
*** seshu has left #openstack23:08
*** joearnold has joined #openstack23:08
*** Alex111 has quit IRC23:13
edayvishy: stop already :)23:17
jesse_he needs his karma23:18
vishyeday: gotta pass jaypipes23:18
jesse_he keeps wanting me to write patches that adds bugs so he can fix them23:19
*** jkakar has joined #openstack23:19
vishyeday: looks like you need to either change hudson or make call nose in order for your nosetest branch to merge23:19
edayvishy: the trick is writing more blueprints, not code23:19
*** hggdh has quit IRC23:19
edayyeah, I need to wait for soren/mtayler for that23:19
edayboth seem to be missing today23:20
*** hggdh has joined #openstack23:23
*** dfg_ has quit IRC23:24
dabovishy: I reviewed and approved both of those.23:28
vishydabo: thx!23:28
*** joearnold has quit IRC23:36
*** nelson__ has quit IRC23:40
*** nelson__ has joined #openstack23:41
*** jbaker has quit IRC23:43
*** joearnold has joined #openstack23:47
*** Ryan_Lane is now known as Ryan_Lane|away23:56
*** Ryan_Lane|away has quit IRC23:56
openstackhudsonProject nova build #324: SUCCESS in 1 min 15 sec:

Generated by 2.14.0 by Marius Gedminas - find it at!