Friday, 2011-02-18

*** jmckenty has quit IRC00:03
kpepplejk0: ok, running tests for last time and I'll push/propose for merge00:04
kpepple* thinks NetworkTestCase takes a looooonnnggg time00:05
*** freeflying has quit IRC00:07
*** freeflying has joined #openstack00:07
kpepplejk0: okay, how do we do this ... do i just propose for merge on the branch page or do i set the blueprint to "needs code review" ?00:09
*** arcane has quit IRC00:09
jk0kpepple: go ahead and request a merge prop thru the branch (my flavors BP was already accepted, so we can juse use that, it points to your wiki spec)00:10
kpepplejk0: done00:12
jk0got the URL handy to the merge prop?00:13
*** rnirmal has joined #openstack00:14
jk0excellent, now we just sit back and let the reviews come in :)00:15
*** rnirmal has quit IRC00:15
kpepplejk0: s/reviews/hate mail/00:15
*** arcane has joined #openstack00:17
*** dendrobates is now known as dendro-afk00:18
*** aliguori has quit IRC00:18
justinsb_jk0, kpepple: Migration # 004 has already gone to zones... I'm hoping I'll grab #005, but #004 is definitely gone :-)00:19
jk0cool, kpepple: mind changing that to #006?00:20
justinsb_jk0: Or #005 if you beat me to it!00:20
kpepplejk0 , justinsb_ : no problem ... i was worried this would happen00:20
jk0yeah, no way around it00:21
*** msassak has quit IRC00:22
kpepplejk0: repushed ... do i need to re-propose ?00:22
jk0nope, all good :)00:22
jk0that's one thing LP does really nice00:22
kpepplejk0: cool00:23
justinsb_jk0: I saw your python bindings to OpenStack... what's the plan with those?  Are those going to live in lp:nova?  (BTW, they're great, I use them for my testing.)00:24
jk0I don't think there are any plans to include them in Nova, but they are our "blessed" bindings if you want to call them that00:25
jk0we're working on feature parity in the API right now00:25
*** ctennis has joined #openstack00:25
jaypipeskpepple: documentation ftw. nice work.00:25
justinsb_jk0: I think it would make lots of sense to have an API binding that could move in parallel with API changes, along with integration tests, for every patch where it's relevant.  But maybe that's just a dream...00:26
jk0that's essentially what we're doing with it now, only it's just a separate project00:26
jk0we make updates to python-novatools in parallel to trunk merges00:27
justinsb_jk0: One project would mean everything's in lockstep, but I'm just happy someone's building API bindings!  One step at a time!00:27
openstackhudsonProject nova build #537: SUCCESS in 1 min 36 sec:
openstackhudsonTarmac: Correctly pass the associate paramater for project_get_network through the IMPL layer in the db api.00:27
kpepplejaypipes: thx. i've had to take a bunch of "flyers" on assumptions of coding style within this branch ... please comment early and often, it's a bit of a learning experience00:29
*** desai has joined #openstack00:32
*** freeflying has quit IRC00:32
*** freeflying has joined #openstack00:32
openstackhudsonProject nova build #538: SUCCESS in 1 min 37 sec:
openstackhudsonTarmac: Makes FlatDHCPManager clean up old fixed_ips like VlanManager.00:33
*** adiantum has quit IRC00:37
*** dfg has quit IRC00:40
Code_Bleuwhat am i doing wrong:
*** adjohn has joined #openstack00:43
vvuksanit's not -k mykey.priv but -k mykey00:43
*** adiantum has joined #openstack00:44
Code_Bleuwuksan: ok, i will try that00:46
Code_Bleuvvuksan: that worked...thanks00:46
vvuksanalthough if you can log in00:46
vvuksanyou may need to do this00:47
vvuksanssh-keygen -f nova_key && chmod 600 nova_key ;  euca-add-keypair nova_key > nova_key.priv00:47
vvuksani forget00:47
*** bsdninja has joined #openstack00:47
*** eikke has quit IRC00:49
Code_Bleuvvuksan: ok i must have done something else wrong.  my vm instance has a address, but my openstack server is on the 10.0.0 i cant ssh to the vm00:52
*** bsdninja has left #openstack00:53
*** adiantum has quit IRC00:53
*** msassak has joined #openstack00:53
*** littleidea has joined #openstack00:57
*** adiantum has joined #openstack00:58
*** freeflying has quit IRC01:00
*** freeflying has joined #openstack01:01
vvuksanCode_Bleu: what kind of networkign are you using ?01:03
Code_Bleuvvuksan: default VLAN config01:04
*** vvuksan has left #openstack01:04
*** vvuksan has joined #openstack01:05
Code_Bleui see the bridge interface and am able to now ping the gateway for that vm segment ( but i still cant ping the vm or ssh...i have verified that i have the authorize rules in also01:05
vvuksanyou need routing01:05
*** vvuksan has left #openstack01:05
*** vvuksan has joined #openstack01:05
vvuksantry something like that01:05
*** adiantum has quit IRC01:08
*** adiantum has joined #openstack01:13
*** littleidea has quit IRC01:18
Code_Bleuvvuksan: i just noticed that after i try to run the instance..the status goes from "launching" to "...failed to spawn mykey.."01:21
vvuksanCode_Bleu: redo the key01:22
vvuksanwith ssh-keygen01:22
Code_Bleudoes it have to have the name "nova_key" like you mentioned?01:23
*** adiantum has quit IRC01:23
*** alekibango has joined #openstack01:26
rds__hi guys, I have same problems running nova-volume, after installed nova lp:/nova/bexar via novascript01:26
rds__this is the error
rds__can someone help me please01:27
Code_Bleuvvuksan: still failed to spawn key01:28
kpepplerds__: you need to create a volume group called nova-volumes on your server ...01:28
*** adiantum has joined #openstack01:29
vvuksanCode_Bleu: odd01:29
*** pvo has quit IRC01:29
vvuksanCode_Bleu: type euca-describe-keypairs01:30
vvuksanrds__: sudo vgcreate nova-volumes /dev/sdb01:31
Code_Bleuvvuksan: that looks jacked there a way to clear that out and redo my keypairs?01:31
Code_Bleuvvuksan: i have 3 different ones in there.01:32
Code_Bleuvvuksan: on that isnt labeled, on with the "mykey" label, and the last one with the "nova_key" label01:32
Code_Bleuvvuksan: the one without any labels, appears to be the correct keypair number...i have seen that number earlier when also has the project name appended to the key.  The other ones listed do not01:33
Code_Bleuvvuksan: is there a file that i can edit and clean that up?01:34
vvuksanCode_Bleu: you will need to clean up the table in mySQL01:35
rds__vvuksan, kpepple thank for answer and link01:35
Code_Bleuvvuksan: what about the euca-delete-keypair?01:36
vvuksanCode_Bleu: you could try that as well01:36
*** Ryan_Lane|dc has quit IRC01:36
*** vvuksan has quit IRC01:36
Code_Bleuvvuksan: should the EC2_ACCESS_KEY match the same number as the euca-describe-keypairs number?01:39
Code_Bleuvvuksan: using the euca-delete-keypair, i removed everything from there except the one i just generated with the ssh-keygen01:43
Code_Bleuvvuksan: do i need to do nova-manage project zipfile.... again and re source the novarc?01:44
*** MarkAtwood has joined #openstack01:44
kpeppledevcamcar: you around ? got db-migrate issue with our branch01:45
*** MarkAtwood has quit IRC01:45
*** miclorb_ has quit IRC01:45
*** MarkAtwood has joined #openstack01:45
kpepplejustinsb_: you around ? i can't migrate forward because your 005- migration doesn't exist yet01:45
*** sebastianstadil has quit IRC01:47
*** jc_smith has quit IRC01:49
*** adiantum has quit IRC01:51
Code_Bleuhow do you do the opposite of this: uec-publish-tarball ubuntu1010-UEC-localuser-image.tar.gz dub-bucket x86_6401:54
Code_BleuI think i chose the wrong hardware type and want to remove the image and start over01:55
*** adiantum has joined #openstack01:56
creihtmtaylor: much better, thanks!01:58
*** desai has quit IRC02:02
*** littleidea has joined #openstack02:03
*** creiht has quit IRC02:04
*** desai has joined #openstack02:05
*** creiht has joined #openstack02:06
*** ChanServ sets mode: +v creiht02:06
Code_Bleuhow do you clean up images without just going to the directory and deleting them?02:08
*** littleidea_ has joined #openstack02:10
*** mray has joined #openstack02:10
Code_Bleunevermind..found it "euca-deregister <image identifier>02:12
*** sebastianstadil has joined #openstack02:13
*** miclorb__ has joined #openstack02:14
*** littleidea has quit IRC02:14
*** littleidea_ is now known as littleidea02:14
*** gregp76_ has joined #openstack02:18
*** dragondm has quit IRC02:18
*** littleidea has quit IRC02:20
Code_Bleudoes anyone know what the hardware-arch syntax are?  64 bit is the x86_64.  What is it for non 64bit? i386,i486..etc...or just x86?02:21
Code_Bleuthats for the uec-publish-tarball command02:21
kpeppleCode_Bleu: for ubuntu images, it's i386 ...
Code_Bleukpepple: i downloaded the one listed on this page...and in they list it as x86_64, but my hardware is not can i just use that same image and change the hardware-arch to i386?02:24
Code_Bleusorry forgot link to page:
*** Ryan_Lane has joined #openstack02:25
*** MarkAtwood has quit IRC02:25
kpeppleCode_Bleu: what virtualization are you using ? UML ?02:25
kpeppleCode_Bleu: you won't be able to just change it to i386 ... it's got a 64bit kernel and won't boot without it02:26
Code_Bleukpepple: thats what i was thinking my problem was.  this doc is assuming 64bit..and that is a 64bit image right?02:27
kpeppleCode_Bleu: i haven't used that image but i would assume yes.02:27
Code_Bleukpepple: im using whatever the default is for the manual install of openstack.  Im following everything on this site.02:27
Code_Bleukpepple: why is it even an option to change the hardware-arch in the publish tarball then?02:28
*** desai has quit IRC02:28
jaypipeskpepple: will do! (in the morning) :)02:29
*** littleidea has joined #openstack02:29
*** littleidea has quit IRC02:31
annegentleCode_Bleu: heh. good point re: hardware arch, but those are the two params you enter for the command02:32
kpeppleCode_Bleu: ah, your using publish tarball ... this creates a image, ramdisk and kernel ... untar the tar.gz file and see what the kernel that is in there ... you might be able to replace it with a different one02:32
kpeppleannegentle: so that tarball is the 64bit version ? do we have a 32 bit version ?02:32
kpeppleannegentle: the 32bit vs 64bit questions come up a lot here on IRC02:33
annegentlethat tarball is one that we have tested and know it works, we could make a tested 32bit one as well, sounds like a good idea02:33
*** msassak has quit IRC02:33
Code_Bleumakes sense why i was getting a "failed to spawn <key>" when trying to run the instance ;-)02:34
*** gregp76_ has quit IRC02:37
*** adiantum has quit IRC02:40
*** adiantum has joined #openstack02:41
*** littleidea has joined #openstack02:44
*** littleidea has quit IRC02:46
*** miclorb__ has quit IRC02:48
*** miclorb_ has joined #openstack02:49
*** sebastianstadil has quit IRC02:50
*** adiantum has quit IRC02:50
Code_Bleuso can i just download an image from and publish tarball it?02:56
*** sebastianstadil has joined #openstack03:01
*** adiantum has joined #openstack03:03
*** mray has quit IRC03:05
*** olivier__ has joined #openstack03:06
*** olivier_ has quit IRC03:07
*** mdomsch has quit IRC03:07
*** olivier__ has quit IRC03:09
*** mray has joined #openstack03:10
*** winston-d has joined #openstack03:11
*** olivier_ has joined #openstack03:12
*** miclorb_ has quit IRC03:15
*** desai has joined #openstack03:24
*** miclorb has joined #openstack03:30
*** mdomsch has joined #openstack03:31
*** desai has quit IRC03:51
*** dragondm has joined #openstack03:52
*** adiantum has quit IRC03:54
*** adiantum has joined #openstack03:58
*** rds__ has quit IRC04:03
*** miclorb has quit IRC04:09
jarrodcode_bleu: yes04:14
*** kashyapc has joined #openstack04:15
dsockwellall right -- attempting to install nova yet again.  I've followed the procedure at , referencing the ubuntu 10.04 procedure to get the code.04:18
dsockwellwhen I begin step 3, i encounter the following error:04:18
dsockwellroot@nova0:~/nova# python build04:18
dsockwellNo handlers could be found for logger "nova.root"04:18
*** bird has joined #openstack04:20
dsockwelli'd like to complete the setup process -- could there be some dependency i'm missing?04:20
jk0have you ran the unit tests? those usually make missing dep easier to track down04:21
*** adiantum has quit IRC04:21
Code_Bleuim still getting a "failed to spawn key" even after cleaning old instance and image up, and downloading 32bit image from ubuntu and publish tar balling it.  Here is the output from the comppute.log  Anyone know what im doing wrong?
dsockwelljk0: i have not.  what's the best way to do that?04:22
jk0dsockwell: cd to the nova code base and just ./ -N04:22
dsockwellImportError: No module named nose04:23
jk0there you go :)04:23
dsockwelln/m i will complete the installation on ubuntu and then worry about taking it to centos04:25
Code_Bleumy server does not support the vmx or svm cpu flags that i know is needed for kvm, but i thought that openstack works with other hypervisors that dont require those specs?04:25
dsockwellsupposedly you can use kqemu or xen04:26
dsockwellthe fire-and-forget script from launchpad implements a qemu environment04:27
dsockwelljk0: looks like there's a whole dependency session missing from the guide I mentioned.  Thanks for showing me the test script.04:27
*** adiantum has joined #openstack04:28
Code_Bleuhow do i use qemu instead of kvm?04:32
dsockwellCode_Bleu: good question.  i will be in the same position once I install the core packages.04:38
Code_Bleudsockwell: i believe i got it running by adding --libvirt_type=qemu to the nova.conf file04:43
Code_Bleunow my prob is i still cant ping or ssh04:43
dsockwelli was never able to get networking working on the demo install04:43
*** mray has quit IRC04:45
dsockwellwoo, executed04:45
*** mray has joined #openstack04:57
*** kashyapc has quit IRC04:58
*** Ryan_Lane has quit IRC04:59
dsockwellOK, got a little further.  I'm installing nova as per on ubuntu 10.04.  I've resolved some dependencies, and gone up to step 3, but when I create my user it doesn't generate a file05:05
dsockwellhow can I make nova spit out some credentials?05:05
*** adiantum has quit IRC05:08
kpeppledsockwell: what happens when you do the nova-manage project zipfile command ? can you use the nova-manage project environment command ?05:11
dsockwellkpepple: one moment, i will try it05:11
dsockwelli gave: nova-manage project zipfile novacat dsockwell05:13
dsockwelland it returned:05:13
dsockwell(nova.root): TRACE: OSError: [Errno 2] No such file or directory: '/usr/local/lib/python2.6/dist-packages/nova-2011.2-py2.6.egg/nova/..//CA'05:13
dsockwellso it seems like I need to create a CA?05:13
dsockwellnova-manage project environment novacat dsockwell05:15
*** adiantum has joined #openstack05:15
dsockwellseems to work OK, but returns no output05:15
kpeppledsockwell: look for a file called novarc in your directory and then source it05:15
dsockwellgot it05:16
kpeppledsockwell: for the CA, you should see a directory called "CA" in your nova directory ... there should be a bunch of files there05:16
dsockwellfound those05:16
kpeppledsockwell: once you source novarc, you should be able to use a bunch of euca-* commands05:17
*** miclorb has joined #openstack05:17
dsockwellok.  once I have users, I will want to give out credentials in zipfiles.  is there some other userator outside of nova-manage that can generate them, or will I need to point nova-manage to the CA?05:19
kpeppledsockwell: you need nova-manage to point to the CA ... we need to copy the nova/CA/ directory somewhere ... just looking where ... i remember seeing the instructions, but can't remember where05:20
dsockwellit wouldn't be                    '/usr/local/lib/python2.6/dist-packages/nova-2011.2-py2.6.egg/nova/..//CA'05:20
kpeppledsockwell: i don't think it's the dist-packages ...05:21
kpeppledsockwell: did you install from repo (via bzr trunk) or via apt-get packages05:21
dsockwellfrom repo05:22
dsockwelli did 'bzr clone lp:nova' .05:23
kpeppledsockwell: did you do the "cd CA; ./" command ?05:24
dsockwellunless did it for me05:24
kpeppledsockwell: it's on bottom of fourth step at
dsockwellroot@nova0:~/nova/CA# ./genrootca.sh05:27
dsockwellNot installing, it's already done.05:27
kpeppledsockwell: you're on ubuntu 10.10 ?05:28
dsockwellnope, 10.0405:28
kpepplehmmm ... okay. let me boot my server and see where it's installed.05:28
dsockwellroot@nova0:~/nova/CA# find / -name CA |grep nova05:28
*** kashyapc has joined #openstack05:29
dsockwellthat's got some files dated at about install time05:32
kpeppledsockwell: can you run the nova-manage project zipfile command as root ?05:32
dsockwellsure, what arguments is that supposed to take?05:33
dsockwelli tried nova-manage project zipfile <projectname> <username>05:33
kpeppledsockwell: that's it ... but sudo it05:33
dsockwellright, i've done all this in a root shell05:34
dsockwellperhaps that was wrong05:34
dsockwellhm, most of that tutorial is sudo'd anyway05:35
dsockwellit still gives me the erorr i pasted before, it's looking in /usr/local/lib/python2.6/dist-packages/nova-2011.2-py2.6.egg/nova/..//CA05:35
kpeppledsockwell: where is the paste ?05:35
dsockwelllet me pastebin the whole trace, one moment05:36
kpeppledsockwell: what is in /usr/local/lib/python2.6/dist-packages/nova-2011.2-py2.6.egg/CA ?05:40
dsockwelli found some certificate-like files in /usr/local/lib/python2.6/dist-packages/nova-2011.2-py2.6.egg/nova/tests/CA05:42
kpeppledsockwell: you should see a openssl.cnf file in there05:42
dsockwelllinking the path I just pasted to /usr/local/lib/python2.6/dist-packages/nova-2011.2-py2.6.egg/CA didn't seem to work either05:43
dsockwelli'll try linking the CA directory from my install location05:44
kpeppledsockwell: do this ... cp your nova/CA/* (your bzr repo) directory to /usr/local/lib/python2.6/dist-packages/nova-2011.2-py2.6.egg/CA05:44
kpeppledsockwell: yeah, i think that would work too05:44
dsockwellroot@nova0:~/nova# nova-manage project zipfile novacat0 dsockwell05:45
dsockwell2011-02-17 22:44:46,852 WARNING nova.auth.manager [-] No vpn data for project novacat005:45
dsockwellroot@nova0:~/nova# ls -lh nova.zip05:45
dsockwell-rw-r--r-- 1 root root 5.7K 2011-02-17 22:44 nova.zip05:45
dsockwellawesome :)05:45
kpeppledsockwell: okay ... the file is what you want ... don't worry about the vpn stuff05:45
dsockwellok -- so i take it I want to give that zip file to my unprivileged user, extract it into ~/.nova and then source ~/.nova/novarc whenever I want to use euca2ools?05:46
kpeppledsockwell: yes. from there you can generate keys for uploading images and launching instances.05:48
dsockwellgot it05:49
dsockwellnow, i've tried euca-describe-availability-zones, and it hangs05:49
dsockwell[Errno 111] Connection refused05:50
dsockwelldenhac@nova0:~$ curl
dsockwellcurl: (7) couldn't connect to host05:51
kpeppledsockwell: hmmm ... do other euca-* commands work ? if not, you need to open a port on your firewall ... probably 8773 (nova-api default)05:51
dsockwelliptables -L tells me there are no filters in place05:51
kpeppledsockwell: also check that nova-api is running05:52
dsockwellmaybe I should start nova-api05:52
dsockwellwhat a concept05:52
kpeppledsockwell: from here on out, you'll need nova-scheduler, nova-compute, nova-network, nova-api and nova-objectstore (or glance) up05:53
dsockwellok.  it's indicating that now I need to provide a configuration.05:54
*** hadrian has quit IRC05:55
dsockwelllooks good enough to me, i'll try just copying that into /etc05:55
birddsockwell:  Check nova.conf too05:55
kpeppledsockwell: i use only one conf file for all nova-* binaries which i put in /etc/nova/nova.conf05:56
dsockwelloh, it wants a whole directory to itself, doesn't it05:57
*** gregp76_ has joined #openstack05:57
dsockwelli started the service.  i think i need to give it some logging options, logging to the terminal is pretty confusing.05:59
kpeppledsockwell: in your nova.conf file, put in "--verbose", "--daemonize=1" and "--logdir=/var/log/nova"06:01
*** miclorb has quit IRC06:01
*** dragondm has quit IRC06:01
dsockwellah -- there's a skeleton file on one of the tuts, i'll start with that06:03
dsockwellif i don't specify a database, will it default to sqlite?06:06
dsockwelli'll just set up mysql06:07
kpeppledsockwell: yes, it defaults to sqlite3 without a --sql_connection flag (with username:password) in /etc/nova/nova.conf06:08
*** mdomsch has quit IRC06:14
*** mray has quit IRC06:14
dsockwellok, do I need to explicitly pass /etc/nova/nova.conf to nova-api?06:23
kpeppledsockwell: you usually start all of the nova binaries with something like "nova-api --flagfile=/etc/nova/nova.conf"06:24
*** pothos has quit IRC06:29
dsockwellit's not giving me any indication that it's sourcing that file.06:31
kpeppledsockwell: look in /var/log/nova/nova-api.log ... you should see it dump "flags" when it starts, which are the configuration options from that file06:32
dsockwell2011-02-17 23:30:24,823 AUDIT nova.root [-] Starting /usr/local/bin/nova-api on
dsockwellthat's the only line, repeated a few times06:34
kpeppledsockwell: hold on ... i forgot. nova-api is different .. it uses paste ... look in your nova/etc directory for a nova-api.conf file06:34
dsockwellfound it06:34
kpeppledsockwell: put that file in the /etc/nova directory06:37
dsockwellalready there06:37
*** maplebed has joined #openstack06:37
dsockwelli compared the two as well, they're identical06:37
dsockwellin the tutorial for 10.10 they stuff it into a screen -- is it really supposed to run there in production?06:38
kpeppledsockwell: they shouldn't be identical ... nova.conf should have a bunch of "--flags" directives while nova-api.conf looks more like a MS .ini file (paste)06:38
kpeppledsockwell: no, they just use screen to demo it and easily change between log files06:38
dsockwell/root/nova/etc/nova-api.conf is the same as /etc/nova/nova-api.conf06:38
dsockwellsorry for the confusion there06:39
dsockwell/etc/nova/nova.conf is some command line stuff06:39
kpeppledsockwell: oh ... hmmmm  .. you should see something like this - - when you start nova-api (nova-api.log)06:40
dsockwelli do see that06:41
kpeppledsockwell: okay ... but your euca-* tools don't work ?06:41
*** pothos has joined #openstack06:41
dsockwellthey seem to work, actually, i just want to suppress the vomitus from nova-api when I use them06:42
dsockwellit gives me about 2 screens of helpful debug when I try to use euca2ools06:42
kpeppledsockwell: ah ... i run my install from pa:trunk packages and it starts nova-api with "su -c nova-api --flagfile=/etc/nova/nova.conf nova"06:44
kpeppledsockwell: however, that might be the upstart script06:44
dsockwellok, when i pass the directives --verbose, --daemonize=1, and --logdir=/var/log/nova from the command line, it seems to work OK06:48
dsockwellbut when i try --flagfile, it acts as though I passed no args at all06:48
dsockwelltriple checking the syntax06:50
kpeppledsockwell: hmmm ... i haven't really looked at nova-api and it's a bit different than the other binaries ...06:50
dsockwellok, so.  what's the postcondition for ' install' ?  should it have given me some working daemons?06:52
kpeppledsockwell: if you pulled from launchpad, your nova/bin directory has all your daemons but i don't think install actually runs them06:54
*** pothos has quit IRC06:58
*** pothos has joined #openstack06:58
dsockwellwill the daemon ignore --flagfile if it's incomplete?06:59
kpeppledsockwell: what do you mean ?07:02
dsockwelln/m, i got it07:02
dsockwellwhat I want to do now is use my terminal without nova-api puking on it07:03
dsockwelli guess i could just start another one07:03
dsockwellso, wait07:08
dsockwellthis PPA business means there's packages i could have used?07:08
dsockwelli need a drink07:08
kpeppledsockwell: yes ... we don't do a good job of telling people that. there are two ppas: one nightly trunk and another just releases. i have the nightly trunk on my machine (so i almost always have an install that i can fire up) and then use trunk source via bzr to develop against07:10
*** maplebed has quit IRC07:11
dsockwelllet me nuke this installation real quick, then i'll try the installation procedure from the official manual07:12
kpeppledsockwell: you don't need to nuke the installation to install from packages ...07:13
*** naehring has joined #openstack07:14
dsockwellit's no trouble, and i dirtied up /usr/local pretty good.07:14
*** 30BAAUZIF has joined #openstack07:14
*** omidhdl has joined #openstack07:17
*** adiantum has quit IRC07:18
dsockwellkpepple: so riddle me this -- if I install openstack on Ubuntu LTS now, and migrate to Centos6 when it drops, what kind of trouble am I asking for?07:21
kpeppledsockwell: the good people at grid dynamics have created a repo for RHEL 6 already ... see ... it seems pretty straight forward07:23
dsockwellall right.  think moving the object stores, etc, would be a pain?07:27
dsockwelli know the configuration database and files are pretty well self contained07:27
*** pothos has quit IRC07:28
*** pothos has joined #openstack07:29
kpeppledsockwell: that's a wild card ... from the notes, RHEL/CentOS doesn't support qcow images, which i don't think i need. i haven't played with a lot of the imagestore options yet ... i need to get glance running (metadata server) .07:30
*** adiantum has joined #openstack07:30
*** naehring has quit IRC07:31
dsockwellok.  i'm planning on running the live images off a solaris box, i'm just concerned about the machine images, etc.07:31
dsockwelli suppose it wouldn't be too hard to just rebundle them07:31
dsockwellso the solaris machine would handle all the cow business07:32
dsockweller, that is using the solaris box as a file server, and executing the images on ubuntu (and then centos) machines which will read the images through a SAN07:33
kpeppledsockwell: okay ... that's probably fine07:33
dsockwelli suppose the worst that could happen is i'd have to repackage the running images and store them off the cluster as amis, then rebundle them07:34
dsockwellor wreck everything when centos6 comes out and start over, i could do that07:35
dsockwellthat'd be fine07:35
kpeppledsockwell: either way would work07:35
kpepplegood luck with everything07:36
*** ramkrsna has joined #openstack07:37
*** ramkrsna has joined #openstack07:37
*** Code_Bleu has quit IRC07:39
*** CloudChris has joined #openstack07:39
*** berendt has joined #openstack07:41
dsockwellthanks again, kpepple.  i think i'm going to call it a night, actually.  i need to at least pretend to keep daylight hours.07:43
*** littleidea has joined #openstack07:56
*** naehring has joined #openstack07:58
*** jc_smith has joined #openstack07:58
*** perestre1ka has quit IRC08:01
*** jc_smith has quit IRC08:01
*** gregp76_ has quit IRC08:03
*** Code_Bleu1 has joined #openstack08:08
*** rcc has joined #openstack08:08
*** fabiand_ has joined #openstack08:13
*** gasbakid has joined #openstack08:14
*** yamahata_ has joined #openstack08:26
*** rchavik has joined #openstack08:27
*** skiold has joined #openstack08:32
*** zokr has joined #openstack08:33
*** miclorb_ has joined #openstack08:33
*** filler has quit IRC08:34
*** MarcMorata has joined #openstack08:36
*** miclorb has joined #openstack08:44
*** miclorb_ has quit IRC08:44
*** adiantum has quit IRC08:44
*** calavera has joined #openstack08:45
*** miclorb has quit IRC08:47
*** rcc has quit IRC08:47
*** Nacx has joined #openstack08:52
*** filler has joined #openstack08:55
*** adiantum has joined #openstack08:57
*** rcc has joined #openstack09:05
*** ramkrsna has quit IRC09:07
*** littleidea has quit IRC09:14
uvirtbotNew bug: #721141 in nova "empty logfiles in /var/log/nova/ after logrotating with logrotated" [Undecided,New]
sorenttx: It's doing *something*
*** ramkrsna has joined #openstack09:20
*** adiantum has quit IRC09:20
sorenberendt: Oh, darn.09:22
openstackhudsonProject nova-bexar build #3: SUCCESS in 1 min 34 sec:
openstackhudsonTarmac: Fixes tarball contents by adding missing scripts and files to / MANIFEST.in09:23
*** johnh has joined #openstack09:25
*** adiantum has joined #openstack09:27
ttxsoren: looks good, though it didn't autotrigger nova-bexar-tarball yet09:28
*** ramkrsna_ has joined #openstack09:30
ttxsoren: did you have specific plans to SRU something on bug 658257... or should I just FixReleased it ?09:30
uvirtbotLaunchpad bug 658257 in nova "get_console_output fails" [Medium,Fix released]
ttx(talking about the ubuntu/nova task)09:31
sorenttx: Misconfig. Fixed.09:31
uvirtbotNew bug: #721144 in openstack-manuals "swift docs incorrectly reference CDN" [Undecided,New]
sorenttx: I did. Now I don't.09:32
ttxsoren: ok09:32
*** kashyapc has quit IRC09:33
*** ramkrsna has quit IRC09:33
*** gasbakid has quit IRC09:36
*** kashyapc has joined #openstack09:40
berendtsoren: very fast solution.. :)09:40
*** allsystemsarego has joined #openstack09:41
*** allsystemsarego has joined #openstack09:41
ttxsoren: looks like the hudson toolchain on nova/bexar is working. Approve away.09:43
ttxsoren: I should probably push a version update in the queue as well, what do you think ?09:45
*** adiantum has quit IRC09:46
sorenttx: Erk. Yeah. That's going to suck.09:46
sorenttx: But yes.09:46
ttxsoren: why is it going to suck ?09:47
sorenttx: Because our versioning thingamajig only has: NOVA_VERSION = ['2011', '2']09:47
sorenttx: There is no micro-version.09:48
sorenErk indeed.09:48
*** metoikos has joined #openstack09:52
*** adiantum has joined #openstack09:52
CloudChrisAnyone from Sofia in here?10:02
sorenSofia, Turkey?10:03
johnhis there a guide on howto configure vlan-networking for the novaproject ? Howto setup the nic's, cloudpipe and so on?10:03
CloudChrisno Sofia, Bulgaria10:03
sorenErk, geography fail.10:04
sorenCloudChris: Right, Bulgaria.10:04
sorenI don't think so, no. why?10:04
*** 30BAAUZIF has quit IRC10:04
CloudChriswaiting for someone ... dunno name ... standby10:04
*** gasbakid has joined #openstack10:07
*** gasbakid_ has joined #openstack10:07
*** exitdesc1iption has quit IRC10:08
*** adjohn has quit IRC10:08
*** bird has quit IRC10:11
*** bird has joined #openstack10:12
*** guigui has joined #openstack10:15
openstackhudsonProject nova-bexar build #4: SUCCESS in 1 min 35 sec:
openstackhudsonTarmac: Checks whether the instance id is a list or not before assignment. This is to fix a bug relating to nova/boto. The AWK-SDK libraries pass in a string, not a list. The euca tools pass in a list.10:17
*** kashyapc has quit IRC10:19
*** guigui1 has joined #openstack10:21
*** guigui has quit IRC10:21
*** adjohn has joined #openstack10:23
openstackhudsonProject nova-bexar build #5: SUCCESS in 1 min 33 sec:
openstackhudsonTarmac: Get a fresh connection in rpc.cast rather than using a recycled one.10:28
*** guigui1 has quit IRC10:37
*** guigui has joined #openstack10:37
*** eikke has joined #openstack10:45
*** londo__ is now known as londo_10:46
openstackhudsonProject nova-bexar build #6: SUCCESS in 1 min 34 sec:
openstackhudsonTarmac: Fixes undefined reference to volume_ref that is used for a log message in init_host in volume/ The variable was supposed to be just volume.10:47
*** ramkrsna__ has joined #openstack10:49
*** perestrelka has joined #openstack10:51
*** ramkrsna_ has quit IRC10:52
rcchey soren, do we honour error codes from ec2 api?
rcci don't think so, but wanted to double-check10:57
*** naehring has quit IRC11:01
*** adiantum has quit IRC11:04
sorenrcc: What do you mean?11:07
rcclet's say11:08
rccyou create a group11:08
rccthat's already existing11:08
*** skiold has quit IRC11:08
rccso, we should return InvalidGroup.Duplicate plus the 400 HTTP error code11:08
rccthat InvalidGroup.Duplicate, we don't return it, right?11:08
sorenIf thats11:08
sorenwhat EC2 does, that's what we should do.11:09
johnhis it possible to install nova compute and swift on one system ? and if it works, how to setup swift ?11:10
*** adiantum has joined #openstack11:11
rccto your knowledge, we don't return any of those error codes, do we?11:12
*** guigui has quit IRC11:16
*** ramkrsna__ has quit IRC11:18
berendtjohnh: i didn't try to run swift and nova on one node, but should be no problem. you can find a deployment guide here:
sorenrcc: Don't think so, no.11:23
rcck, fair enough, looking for ideas to fix one of my bugs, thx11:24
berendtshould be disapproved merges be totally removed (delete branch) or what to do with them?11:24
rccor you could evaluate the feedback you got for that merge disapproval, fix the code and re-propose for merging11:25
*** h0cin has joined #openstack11:26
berendtrcc: sure ;) but it's a merge that was disapproved because of an other upcoming merge11:26
rccthen i'd wait for the upcoming merge to happen, then merge with your branch and check conflicts . If the two branches have fixes for same stuff, then yeah, i'd say just get rid of the branch11:27
rccmy .2 cents11:27
*** yt_ has quit IRC11:27
*** ramkrsna has joined #openstack11:28
*** ramkrsna has joined #openstack11:28
*** HugoKuo_ has joined #openstack11:34
*** HugoKuo__ has quit IRC11:35
*** kashyapc has joined #openstack11:38
sorenberendt: Yeah, I miss a "not going to be merged, but kept around for hysterical raisins" sort of state of merge proposals.11:39
sorenberendt: For now, I just delete them.11:39
*** adiantum has quit IRC11:40
berendtsoren: is it possible to add such a state for branches?11:40
sorenBranches themselves have an "abandonded" or "obsolete" state.11:40
sorenFind the branch on launchpad ("bzr lp-open" from your working dir should do the trick), edit details, set it to one of those states.11:41
* soren heads lunchwards11:41
*** adiantum has joined #openstack11:46
*** adjohn has quit IRC11:53
*** adiantum has quit IRC11:54
*** adiantum has joined #openstack12:00
*** masumotok has joined #openstack12:13
*** omidhdl has quit IRC12:14
*** adiantum has quit IRC12:19
*** berendt has quit IRC12:21
*** adiantum has joined #openstack12:25
*** ctennis has quit IRC12:32
*** skiold has joined #openstack12:33
*** adiantum has quit IRC12:40
*** alekibango has quit IRC12:42
*** skiold has quit IRC12:43
*** adiantum has joined #openstack12:46
*** skiold has joined #openstack12:46
ttxsoren: if you're ok with the bexar version of it, you should probably approve too, I'd like the cactus version to land first.12:50
*** j05h has left #openstack12:54
*** j05h has joined #openstack12:54
*** gasbakid_ has quit IRC12:56
*** guigui has joined #openstack13:01
*** adiantum has quit IRC13:03
*** hggdh has quit IRC13:07
*** hggdh has joined #openstack13:08
*** adiantum has joined #openstack13:10
sorenttx: Good point.13:11
sorenttx: I'd like to see it in action, too.13:11
*** eikke has quit IRC13:16
*** eikke has joined #openstack13:18
*** calavera has quit IRC13:19
uvirtbotNew bug: #721254 in nova "nova returns no status code for state scheduling" [Undecided,New]
*** ctennis has joined #openstack13:22
*** ctennis has joined #openstack13:22
johnhcan someone help me on setting up the nova swift on one system ?13:23
*** cyonyx has joined #openstack13:23
johnhi have 600gb of unmounted space, how can i use it for the storage node ?13:25
wedjathi johnh13:25
wedjatdo you want to use swift ?13:25
johnhkind of :)13:25
johnhi dont know if its really necessary13:26
wedjati d'ont know. swift is really useful for what it is designed for : REST storage service13:26
*** gasbakid has quit IRC13:26
*** adiantum has quit IRC13:26
*** gasbakid has joined #openstack13:27
*** adiantum has joined #openstack13:39
*** guigui has quit IRC13:40
*** adiantum has quit IRC13:46
*** Ryan_Lane has joined #openstack13:48
*** adiantum has joined #openstack13:53
*** guigui has joined #openstack13:59
*** jbryce has joined #openstack14:02
*** Ryan_Lane has quit IRC14:02
*** lvaughn_ has quit IRC14:06
*** lvaughn has joined #openstack14:06
*** aliguori has joined #openstack14:08
*** hadrian has joined #openstack14:09
*** mimustafa has joined #openstack14:09
mimustafaHi, I have a very quick question about Swift14:10
*** ppetraki has joined #openstack14:10
*** guigui has quit IRC14:10
*** guigui3 has joined #openstack14:11
*** gasbakid has quit IRC14:11
mimustafaIf say we have spare hdd capacity in the form of empty drive bays or let's say we added a new storage server, would auto replication begin once the server has been joined to a zone/ring?14:11
*** vvuksan has joined #openstack14:14
*** dprince has joined #openstack14:16
*** mimustafa has left #openstack14:18
dabomimustafa: it's been a while since I worked with the swift code, but if memory is correct, yes, replication would begin once the new node has been joined.14:19
*** grapex has joined #openstack14:24
uvirtbotNew bug: #721285 in nova "nova-api unconditionally logs at DEBUG level" [Undecided,New]
*** vvuksan has quit IRC14:32
*** abecc has joined #openstack14:34
*** johnh has quit IRC14:34
*** vvuksan has joined #openstack14:35
*** adiantum has quit IRC14:35
pikensoren or vishy, if your on let me know, I have a few questions about the rabbit-rpc code and how it is used. I am thinking about canabilizing the class to use in another project because I like how you guys did it.14:39
*** ramkrsna has quit IRC14:41
*** adiantum has joined #openstack14:42
*** imsplitbit has joined #openstack14:44
*** omidhdl has joined #openstack14:51
uvirtbotNew bug: #721297 in nova "nova returns wrong date format in AWS API Calls" [Undecided,New]
*** adiantum has quit IRC14:53
*** Ryan_Lane has joined #openstack14:55
*** hub_cap has joined #openstack14:55
*** Ryan_Lane has quit IRC14:57
*** abecc has quit IRC14:58
*** adiantum has joined #openstack14:58
*** rchavik has quit IRC15:01
*** hub_cap has quit IRC15:05
*** kashyapc has quit IRC15:07
*** uksysadmin has joined #openstack15:08
uksysadminHi, I'm having problem with a multi-node set up under vbox and access to meta data on  I came across a discussion by vishy and soren on routing funnies and is exactly the issue I'm seeing.15:09
zulsoren: ping15:09
uksysadminWhat I want to know is what needs to run on the compute node?15:09
uksysadminis it *just* nova-compute15:09
*** mray has joined #openstack15:13
sorenuksysadmin: Yes.15:18
sorenzul: Make it quick.15:18
sorenpiken: Make it quick :)15:18
imsplitbitsoren: what is the meaning of life?15:18
imsplitbitsoren: make it quick15:18
sorenYou answered your own question.15:19
uksysadminI like it when I'm right15:20
*** clayg has quit IRC15:21
zulsoren: i just noticed that you have a patch in the natty ppa for libvirt can you send that to hallyn if its approiate15:21
*** clayg_ has joined #openstack15:21
*** adiantum has quit IRC15:22
sorenzul: If I thought it was appropriate, I'd upload it to Ubuntu myself.15:22
zulsoren: k thanks15:22
sorenWhy would I sent it to Serge?15:22
zulbecause he kind of looks after libvirt15:22
soren$ grep -c Serge changelog15:24
soren$ grep -c Soren changelog15:24
sorenNevertheless.. It's Ubuntu. There's no such thing as a maintainer.15:24
* soren wanders off15:24
sorenvishy: by the way, please don't change anything that involves iptables today. I'm reworking it completely.15:25
sorenI'm 75% done. All the cool parts are done, now I'm just porting over the existing rules to the new world order.15:25
*** adiantum has joined #openstack15:26
uksysadminif you're looking at iptables stuff, can I ask why nova-network creates the PREROUTE to the address on the default interface IP on start up (should it be configurable? does it make sense to?) and when stopping nova-network it doesn't remove the rule.15:27
*** clayg_ is now known as clayg15:29
*** fsaad has quit IRC15:35
*** vvuksan has quit IRC15:36
creihtmtaylor: ping!15:44
*** rds__ has joined #openstack15:44
*** adiantum has quit IRC15:45
*** armando has joined #openstack15:45
*** ewanmellor has joined #openstack15:47
*** photron has joined #openstack15:49
uksysadminOK - I've seen that my issue with accessing isn't a routing issue - the instance isn't getting an IP address on my compute node.  When this was all running under one host this was fine. my br100 bridge doesn't have an IP assigned on my compute node as it did when it was a single instance. Any pointers?15:50
*** cyonyx has quit IRC15:50
uksysadmindnsmasq on the controller is setting the correct mac, instance and an IP to assign so traffic's just not getting to my instance...15:51
*** adiantum has joined #openstack15:52
*** guigui3 has quit IRC15:52
uksysadmin(think I've got it... thanks for being great listeners ;-)... dnsmasq set to listen on private IP address which existed as part of the bridge - so was set on start up and exists as part of nova-network... my remote compute node doesn't have a real IP that's part of this private range.15:56
*** gondoi has joined #openstack15:59
*** dragondm has joined #openstack16:00
*** rnirmal has joined #openstack16:00
uksysadminI give up.  It's the weekend.16:00
*** mray has quit IRC16:00
*** uksysadmin has quit IRC16:01 Well, at least that rolls off the tongue...16:05
*** Daviey has quit IRC16:06
*** adiantum has quit IRC16:09
ttxnova-core: one more review needed on -- this kinda blocks 2011.1.1 (since we want it to land in Cactus first)16:10
*** annegentle-offto has joined #openstack16:11
*** Daviey has joined #openstack16:11
*** vvuksan has joined #openstack16:11
*** CloudChris has quit IRC16:11
jaypipesmtaylor: around?16:12
*** littleidea has joined #openstack16:14
*** adiantum has joined #openstack16:15
*** bird has quit IRC16:15
*** aliguori has quit IRC16:15
*** MarkAtwood has joined #openstack16:17
*** MarkAtwood has joined #openstack16:17
*** ewanmellor has quit IRC16:19
imsplitbitjaypipes: hey16:22
creihtmtaylor: when you get a moment, it doesn't seem like merges are being made for the swift 1.2 series16:22
photroncan someone point me documentation about the relation between users, projects and networks?16:22
creihtmtaylor: oh... hrm... looks like it might have been merged, but I didn't get an email?16:23
creihtOh I have to subscribe to it16:24
*** flashn has quit IRC16:25
sandywalshjaypipes, hey, that "in progress" / "needs review" thing works great16:25
*** adiantum has quit IRC16:25
*** skiold has quit IRC16:26
annegentle-offtophotron: this page may help but also look at
annegentle-offtophotron: it's a good question for its own page16:28
*** HugoKuo_ has quit IRC16:29
jaypipessandywalsh: very nice work on phase 2 zones.16:29
sandywalshthx ... many miles to go before I sleep :)16:30
jaypipessandywalsh: :)16:30
jaypipesimsplitbit: hey there :)16:30
*** flashn has joined #openstack16:30
*** adiantum has joined #openstack16:30
*** howardroark has joined #openstack16:30
*** howardroark has quit IRC16:32
jaypipesanyone else know what "burwell" is?16:32
ttxjaypipes: no. A typo, maybe.16:33
jaypipesttx: I was wondering if it was an ESXi release name or somehting...16:33
ttxjaypipes: don't you think that hudson when rjecting a branch should set it back to "Work in progress"16:34
ttxlooks like in 90% of the cases it needs some OP intervention, so it clogs the "needs review" list16:34
jarrodwill cactus support kvm snapshots?16:35
photronannegentle-offto: thanks16:35
jaypipesttx: I do.16:36
jaypipesttx: I blame mtaylor. :)16:37
*** aliguori has joined #openstack16:39
*** adiantum has quit IRC16:39
ttxjaypipes: about blame... could you push your glance-admin branch to LP and link it on ?16:42
jaypipesttx: hehe, subtle. I like that.16:43
ttxsubtle, paranoid, and coughing. That defines me.16:43
jaypipesttx: well, you are French, after all :P16:44
* jaypipes prepares for obligatory British remark..16:44
*** adiantum has joined #openstack16:44
*** fabiand_ has quit IRC16:46
vishysoren: sweet16:48
*** adiantum has quit IRC16:52
*** adiantum has joined #openstack16:55
jaypipestr3buchet: like how I "volunteered" you on that bug? ;)17:02
*** iRTermite has quit IRC17:02
*** Pentheus has quit IRC17:03
*** littleidea has quit IRC17:03
*** dsockwell has quit IRC17:03
*** iRTermite has joined #openstack17:04
*** littleidea has joined #openstack17:04
*** maplebed has joined #openstack17:04
*** dsockwell has joined #openstack17:04
jk0jaypipes: I chuckled :)17:07
jaypipesjk0: hehe, until I volunteer *you* for the next one ;)17:07
* jk0 has selective vision ;)17:07
*** eikke has quit IRC17:12
*** sleepsonthefloor has quit IRC17:12
*** sleepsonthefloor has joined #openstack17:12
*** dendro-afk is now known as dendrobates17:13
tr3buchetjaypipes: yes17:13
tr3buchetjaypipes: that's how you get stuff done17:13
*** adiantum has quit IRC17:15
*** littleidea has quit IRC17:17
*** gregp76 has joined #openstack17:18
*** dprince has quit IRC17:19
*** MarcMorata has quit IRC17:21
*** adiantum has joined #openstack17:22
*** dendrobates is now known as dendro-afk17:22
mtaylorjaypipes: what?17:24
mtaylorttx: and tarmac is supposed to set a branch back to work in progress if it rejects it, is this not happening?17:24
jaypipesmtaylor: well good morning sir.17:24
jaypipesmtaylor: it sets it back to Needs Review, but that's not what I wanted to talk to you about :)17:25
mtaylorah. cool17:25
jaypipesmtaylor: wanted to get your advice on this:
mtaylorjaypipes: ok... which thing?17:27
*** Failbaitr has quit IRC17:27
jaypipesmtaylor: the non-existence of glance.version when running the test in non-virtualenv is killing it. But glance.version gets BUILT during install, so not sure how to handle this...17:28
mtaylorah. hrm. ok. um...17:29
*** adiantum has quit IRC17:29
mtaylorjaypipes: ok, I'm going to need coffee first for that one17:29
jaypipesmtaylor: kind of a chicken and egg thing maybe?17:30
*** omidhdl has quit IRC17:30
*** dendro-afk is now known as dendrobates17:31
mtaylorjaypipes: glance.version seems to do the right thing - it tries to import glance.vcsversion, and if that doesn't exist, it uses some default values17:35
mtaylorjaypipes: so I don't think there is any need to try to run in your tests17:36
*** adiantum has joined #openstack17:36
*** Failbaitr has joined #openstack17:36
jaypipesmtaylor: perhaps it's just a path issue?17:37
*** RJD22 is now known as RJD22|away17:37
*** RJD22|away is now known as RJD2217:40
*** eikke has joined #openstack17:45
*** iammartian has joined #openstack17:53
*** zokr has quit IRC18:02
*** joearnold has joined #openstack18:09
*** Nacx has quit IRC18:10
jaypipesmtaylor: hmm, so I cannot reproduce the ImportError on glance.version on my local machine. :( Any suggestions on what might be going on there?18:10
mtaylorjaypipes: no! very confusing - nothing looks ostensibly incorrect18:11
*** dprince has joined #openstack18:12
* jaypipes is a sad panda.18:13
*** ccustine has joined #openstack18:15
*** photron has quit IRC18:19
*** dragondm has quit IRC18:24
*** abecc has joined #openstack18:25
*** jc_smith has joined #openstack18:30
*** greenisus has joined #openstack18:30
*** burris has quit IRC18:32
kpepplevishy , xtoddx : did we come to agreement on a logging approach ala lp718360 ? i see that we had another related bug logged today. just wondering if i need to withdraw my branch or revise ...18:33
kpeppleor 718360 so that uvirtbot wakes up18:34
*** pvo has joined #openstack18:36
*** Ryan_Lane has joined #openstack18:46
*** adiantum has quit IRC18:46
*** greenisus has quit IRC18:48
*** hvaldivia has joined #openstack18:48
hvaldiviaI installed nova in 4 machines, two of them are nova-compute18:49
hvaldiviabut executing euca-describe-availability-zones verbose18:50
hvaldiviaonly shows one machine18:50
*** armando has quit IRC18:52
hvaldiviaI looked at /var/log/nova/nova-compute.log and this is the output
*** adiantum has joined #openstack18:53
kpepplehvaldivia: assume you are using trunk ... did you migrate your database to current version ? your database is lacking a column that was introduced this week ...18:55
hvaldiviakpepple: Last week I downloaded  nova-deployment-tools script and yesterday I installed the cloud18:58
hvaldiviathe script uses apt-get to perform the installation18:59
*** troytoman-away is now known as troytoman18:59
pvoheya troytoman. hows the teefs?19:00
*** adiantum has quit IRC19:00
kpepplehvaldivia: let me look at the script (i don't use that) ... there are trunk and release ppa19:00
hvaldiviaso I'am confused. If I use add-apt-repo…. etc and apt-get. Which relese should I get?19:00
hvaldiviakpepple: thank you19:01
*** burris has joined #openstack19:02
kpepplehvaldivia: this page describes the different ppas -
troytomanpvo: I survived19:02
vvuksani am unclear on the network creation19:05
vvuksani did this19:05
vvuksannetwork create 1 519:05
vvuksansince I want to be able to hand out 5 IPs19:05
*** adiantum has joined #openstack19:05
vvuksanhowever it shows only .0 to .3 in fixed IPs19:05
vvuksanand in networks it changed the cidr to /3019:06
imsplitbitwhen I try to gen a zipfile from my project I get this19:08
kpepplehvaldivia: to fix this problem, do this: backup your database (just in case), then run "# nova-manage db sync"19:09
imsplitbitI am following
imsplitbitand am stuck between steps 7 and 819:09
hvaldiviakpepple: I dont understant that info. I mean, I know that someone needs to execute sudo add-apt-repository ppa:nova-core/ppa19:09
hvaldiviasudo apt-get update19:09
kpeppleimsplitbit: if you are on trunk, copy or symlink your nova/CA/* to /usr/local/lib/python2.6/dist-packages/nova-2011.1-py2.6.egg/CA directory19:10
hvaldiviabut I dont know how this is related to the trunk19:10
vishykpepple: i put a couple of suggestions in for alternate approaches, let me try to find todd19:10
imsplitbitkpepple: kk I will try that19:11
vishykpepple: looks like he is at lunch19:11
*** vvuksan has quit IRC19:13
kpepplehvaldivia: yes, someone needed to add the repository (one of the two: "stable"/release or "unstable"/trunk) -- that is in the instructions for the nova deployment tool ( in step 1.19:14
*** adiantum has quit IRC19:14
kpepplevishy: ok, no rush ... but logging issues seem to be adding up :( and i'd like to get a consistent approach (xtoddx has submitted a patch to the other nova-api log issue)19:14
*** retr0h has joined #openstack19:16
*** retr0h has joined #openstack19:16
kpeppleannegentle: you around ? have a question about ...19:17
*** adiantum has joined #openstack19:19
hvaldiviakpepple: What is the difference between:19:21
hvaldiviasudo add-apt-repository ppa:nova-core/release19:21
hvaldiviasudo add-apt-repository ppa:nova-core/trunk19:21
hvaldiviasudo add-apt-repository ppa:nova-core/ppa19:21
hvaldiviaif I modify release to trunk, I will get the latest modifications?19:22
Ryan_Lanehvaldivia: yes19:22
kpepplehvaldivia: different versions of the packages ... release should have version 645 and trunk has version 70019:23
kpepplehvaldivia: but trunk changes every day and may break more often19:24
hvaldiviaBut I executed sudo add-apt-repository ppa:nova-core/releas on my four machine. I dont understand why one of them is ask for the new column 'network.label'19:24
hvaldiviane of them is asking for the new column 'network.label'19:24
*** pvo has quit IRC19:25
kpepplehvaldivia: do a "# sudo apt-cache showpkg nova" which will show you the package version ...19:26
kpepplehvaldivia: or "# sudo apt-cache showpkg nova-compute" would be better19:26
*** jaypipes is now known as jaypipes-away19:26
*** arreyder has quit IRC19:30
*** vvuksan has joined #openstack19:31
hvaldiviaYou are right19:31
hvaldiviakpepple : You are right.19:32
hvaldiviaone is the trunk version and the other one is a dist version19:32
kpepplehvaldivia: that is bad :)19:33
hvaldiviaIn fact the cloud controller has the nova components from the release19:33
hvaldiviaThe the other machines are running trunk-nova-componenets.19:34
kpepplehvaldivia: you should standardize on the cloud controller's version ... it has the database, which is the most important19:34
hvaldiviaI think that the problem is the nova-deployment-script, I will try to fixed19:35
kpepplehvaldivia: ok19:35
*** arreyder has joined #openstack19:36
hvaldiviakpepple: You are good, I was struggling with this issues all day19:36
kpepplehvaldivia: no, i've just made all these mistakes before :)19:36
*** hvaldivia has left #openstack19:44
*** gondoi has quit IRC19:45
*** allsystemsarego has quit IRC19:47
*** adiantum has quit IRC19:49
*** hggdh has quit IRC19:52
*** hggdh has joined #openstack19:52
*** allsystemsarego has joined #openstack19:53
*** allsystemsarego has joined #openstack19:53
*** aliguori has quit IRC19:54
*** arreyder has quit IRC19:55
*** arreyder has joined #openstack19:55
*** adiantum has joined #openstack20:02
*** masumotok has quit IRC20:07
*** h0cin has quit IRC20:08
*** joearnold has quit IRC20:10
*** Mussi has joined #openstack20:11
*** MarcMorata has joined #openstack20:12
*** mdomsch has joined #openstack20:15
*** aliguori has joined #openstack20:16
*** adiantum has quit IRC20:18
*** Mussi has quit IRC20:25
*** adiantum has joined #openstack20:31
*** rds__ has quit IRC20:31
*** if01 has joined #openstack20:32
*** dendrobates is now known as dendro-afk20:32
sorenIt sure is quiet in here...20:33
if01Anyone has any news on dual NICs on openstack?20:33
*** dendro-afk is now known as dendrobates20:33
*** pothos_ has joined #openstack20:34
if01I gave it a try using virt-manager to add a NIC and it seemed to "partially" work...20:35
*** pothos has quit IRC20:35
*** eikke has quit IRC20:36
*** pothos_ is now known as pothos20:36
if01Well, I guess no news (= bad news in this case...)20:37
*** if01 has left #openstack20:38
*** joearnold has joined #openstack20:41
*** joearnold has quit IRC20:44
*** adiantum has quit IRC20:48
*** j05h has quit IRC20:50
*** Vek has joined #openstack20:53
Vekso this is where everyone hangs out.20:53
*** adiantum has joined #openstack20:53
kpeppleVek: sometimes ... it a bit slow today20:54
Vekyeah, well, we just got to the coffee shop.20:54
*** Seoman has joined #openstack20:56
*** MarcMorata has quit IRC20:56
*** dprince has quit IRC20:56
*** westmaas is now known as westmaas_away21:02
openstackhudsonProject nova build #539: SUCCESS in 1 min 39 sec:
openstackhudsonTarmac: Use WatchedFileHandler instead of RotatingFileHandler.21:03
*** dendrobates is now known as dendro-afk21:03
*** mdomsch has quit IRC21:03
*** pvo has joined #openstack21:05
*** Seoman has quit IRC21:12
*** MarcMorata has joined #openstack21:13
*** mdomsch has joined #openstack21:18
*** imsplitbit has quit IRC21:20
*** if_ has joined #openstack21:22
*** if_ has left #openstack21:22
*** zul has quit IRC21:27
*** dragondm has joined #openstack21:27
sorenErr... Heh :)21:36
sorenI meant to say: ah21:36
vishysoren: are you familiar with window scaling?21:38
vishysoren: in tcp packets21:38
*** drico__ has quit IRC21:39
*** drico__ has joined #openstack21:40
sorenvishy: mmm... Vaguely.21:40
vishysoren: I'm having a strange issue where metadata isn't responding to a syn packet21:41
vishybut then later it responds perfectly21:41
*** KenD has joined #openstack21:41
vishyand the only difference between the two packets (that i can see in tcpdump is: one is wscale 2 and the other is wscale 821:41
sorenvishy: What makes you suspect window scaling? SYN packets are rather small :)21:41
vishyi'm not convinced that is the problem, but I can't see anything else that might be doing it21:42
sorenAny chance you're hitting some other limit? conntrack_max or something?21:42
vishyit seems unlikely because it happens every time21:42
vishytrying to access metadata from windows21:42
vishythe only other difference21:42
sandywalshquick question: are there broadcast (fanout) exchanges already set up per service type? I want to broadcast messages to all schedulers from, let's say, a compute service. Or will I need to add that?21:42
vishyis that the second time (the one that works), it sends out an ARP first and gets a response21:43
sorenThat sounds much more like the right place to look to me.21:43
jk0vishy: is the source behind NAT?21:44
sorenIt doesn't send an ARP packet the first time at all?21:44
jk0or some sort of weird bridge?21:44
jarrodis there anyway to assign the public IP to the instance like rackspace does instead of NAT/private?21:44
*** joearnold has joined #openstack21:44
*** adiantum has quit IRC21:44
sorenvishy: Actually, I'll just focus on finishing this iptables thing, if that's ok. Too much brain power assigned to it to really be useful here anyway.21:45
*** KenD has left #openstack21:45
vishyjk0 the source is a vm inside of nebula21:48
retr0hvishy: did you say the source was windows?21:49
vishyretr0h: yes21:50
*** adiantum has joined #openstack21:51
*** ctennis has quit IRC21:55
*** joearnold has quit IRC22:04
*** joearnold has joined #openstack22:07
termiesoren: ah, perfect solution (re roll into one right after release for migrations)22:07
*** adiantum has quit IRC22:07
*** ctennis has joined #openstack22:07
*** ctennis has joined #openstack22:07
*** KenD has joined #openstack22:12
*** adiantum has joined #openstack22:14
*** allsystemsarego has quit IRC22:21
*** vvuksan has quit IRC22:23
*** jbryce has quit IRC22:25
*** ppetraki has quit IRC22:26
jarrodis there a way to bypass nova-network NAT and utilize the public ip on the instances directly?22:26
*** adiantum has quit IRC22:27
sandywalshthere's a new pip-requires for hudson22:32
sandywalshdoes hudson not pick it up automatically?22:32
sandywalshpip python-novatools22:32
termiesandywalsh: as it stands hudson is being manually updated22:32
termiesandywalsh: in the testing discussion we're also proposing adding tests for the pip-based versions22:33
sandywalshtermie, gotcha ... is that something I can do?22:33
sandywalshtermie, my merge prop is failing due to it not being present22:33
termiesandywalsh: not right now unles you have access to that box, but you can read my most recent email to the mailing list to see the beginnings of some testing strategy22:34
*** adiantum has joined #openstack22:34
termiesandywalsh: aye, i've had the same happen, it is sort of a blocking-on-soren being around around thing right now, i think maybe mtaylor or jaypipes might have similar access22:34
sandywalshtermie, ok, thanks ... I'll wait until someone has the cycle.22:35
*** olivier_ has quit IRC22:38
sorentermie: I'm here. What's up?22:38
*** olivier_ has joined #openstack22:38
termiesoren: heya :) sandywalsh needs another library installed on the hudson box22:38
*** mdomsch has quit IRC22:38
sorenOh, novatools?22:38
termiesoren: is there any process for that right now besides asking somebody specific to do it?22:39
sorensandywalsh: How urgent is it?22:39
termiesoren: yup22:39
sorenAny chance at all it can wait until Monday? A big part of the value of Hudson is reproducability. It's like to install novatools from a package, but I've not yet gotten off my arse and made package of it.22:39
*** GasbaKid has joined #openstack22:41
termiesoren: i have a reasonable package development dev environment set up, btw, so maybe when you have a chance you can walk me through some of your process for managing the packages so that i can start contributing to that code22:42
*** dendro-afk is now known as dendrobates22:45
*** blueadept has quit IRC22:46
*** joearnold has quit IRC22:52
*** elastic_dog has quit IRC22:57
*** elasticdog has joined #openstack22:57
*** elasticdog has joined #openstack22:57
*** dendrobates is now known as dendro-afk22:58
mtaylortermie: what is needed?22:59
mtaylortermie, sandywalsh: we do not run tarmac/jenkins tests using pip - we run them using ubuntu package installed versions ...23:00
termiemtaylor: aye, i'm aware of that, just trying to get sandywalsh's merge to get in23:00
termiemtaylor: soren wants to wait an build a package first23:00
mtaylortermie: cool. lemme see what I can do23:00
termiemtaylor: sandywalsh didn't respond as to amount of urgency23:00
sandywalshmtaylor, termie I guess it's only as urgent as the merge is :)23:00
mtaylorah - gotcha - actually reading the full scrollback now23:00
sandywalshbe nice to get it under my belt before truck changes radically23:01
sorenmtaylor: If you're up for it, the packaging from the python-cloudservers package should be easily portable.23:01
termiei accidentally a whole lot of coffee23:01
sorenmtaylor: I believe novatools is a fork of that.23:01
mtaylorsoren: I'm a bit backlogged at the moment... but if I get a free moment over the weekend I'll hit it23:02
mtaylorsoren: ping me if you start working on it, and I'll do the same - you know, so we don't both independently do it. :)23:02
sandywalshtermie, soren I have it on PyPi23:03
termiesandywalsh: they use ubuntu packages and a PPA to do hudson stuff right now23:03
sorenmtaylor: Will do.23:04
termiesandywalsh: we expect to set up an additional hudson that is doing it via the pip pieces, though this failure would still block the merge23:04
sandywalshtermie, gotcha ... well, thanks for looking into it for me!23:04
sandywalshtermie, if there's something I can do scream.23:04
sandywalshtermie, have to run right now23:04
*** dendro-afk is now known as dendrobates23:05
termiesandywalsh: i just made a note on the TestingBrainstorm wiki page about it23:05
*** rnirmal has quit IRC23:05
sandywalshthanks termie23:08
*** adiantum has quit IRC23:10
*** adiantum has joined #openstack23:15
*** alekibango has joined #openstack23:19
*** GasbaKid has quit IRC23:22
*** grapex has left #openstack23:23
*** Pentheus has joined #openstack23:24
*** greyfinch has joined #openstack23:25
sorenvishy: Is there a method that gets called on shutdown (inverse of init_host sort of thing)?23:26
vishyi think so23:26
vishybut i don't think it is passed to manager23:26
greyfinchHowdy, folks. I'm looking at building a small cluster using openstack. Are there well defined ways of keeping VM images in a centralized location, instead of using the Swift component?23:26
vishysoren: service.stop()23:27
vishysoren: it isn't passed to manager23:27
vishysoren: and i don't think it traps a signal either23:28
sorenvishy: ?!?23:28
sorenvishy: So when does it get called?23:28
vishysoren: haha, looks like testing only23:29
soren"Awesome" :)23:29
sorenOk, /me punts on that for now, then.23:30
*** Ryan_Lane has quit IRC23:30
*** drico_ has joined #openstack23:31
*** drico__ has quit IRC23:31
*** Failbaitr has quit IRC23:31
*** hadrian has quit IRC23:31
*** dirakx has quit IRC23:31
*** ironcamel2 has quit IRC23:31
*** Failbaitr has joined #openstack23:32
*** hadrian has joined #openstack23:33
*** ironcamel2 has joined #openstack23:33
*** dirakx has joined #openstack23:33
sorenAwesome. I think my new iptables hotness is done.23:36
*** lvaughn has quit IRC23:36
*** MikeZ2 has joined #openstack23:37
*** lvaughn has joined #openstack23:37
*** dendrobates is now known as dendro-afk23:37
soreniptables really wasn't where I expected to see race conditions, but there you go.23:37
retr0hsoren: what did you run into?23:37
annegentlegreyfinch: Glance is a project of OpenStack that enables storage and retrieval of the VM images - see http://glance.openstack.org23:37
sorenretr0h: Running iptables too often would occasionally fail because of some (sound!) safe guards in the kernel.23:38
greyfinchannegentle: uncanny, I just pulled up a doc that mentioned Glance. It would be a effectively a full replacement for the standard distributed object store, yes?23:38
sorenretr0h: So when I tried spinning up a bunch of VM's at once, they'd almost always fail, except a single one.23:38
retr0hsoren: ohhh interesting.  "running too often" as in inserting,removing,updating rules?23:39
sorenretr0h: Yeah.23:39
soren...and while I was in there, I decided to stop using iptables and start using iptables-restore/iptables-save. iptables is racy anyway.23:39
sorenThe interface in the kernel doesn't let you add individual rules. It deals in full tables at a time.23:40
retr0hsoren: niiice.... yah I think -restore/-save are their recommended usage as is.23:40
retr0hsoren: ohhh - well nice find have a branch I can peek at to see your changes23:40
sorenSo if you add a rule, it'll fetch the entire table, add the rule, and shove the table back.23:41
annegentlegreyfinch: yup.23:41
sorenIf the timing is against you, you might end up losing rules.23:41
sorenI just want to take it for a quick spin to see if this actually works as well as I think it does.23:42
sorenIt's at lp:~soren/nova/iptables-concurrency23:42
retr0hsoren: finding all rules - adding new ones - and saving back wont be racy either?  or only one entry into reading/writting the rules?23:42
greyfinchannegentle: wonderful, thank you for your help.23:43
MikeZ2Could I trouble sombody to give me a link to doc on how to create a machine image for openstack.   I want to run something besides what came in images.tgz and everything I've tried doesn't generate 3 image files (running on ubuntu 10.04.02)23:44
*** lvaughn_ has joined #openstack23:45
sorenretr0h: I've added a manager that keeps track of all the rules Nova wants to have. There's still a chance we'll trip over other components that want to fiddle with iptables, but at least we won't trip over our own feet. Note that the race with other components was already there. This doesn't change that (and really can't, afaics).23:46
*** lvaughn has quit IRC23:47
*** adiantum has quit IRC23:48
sorenretr0h: Remember: finding all rules, adding a new one, and saving the whole thing back.... That's *exactly* what the iptables tool does.23:48
sorenretr0h: With my, new approach we can do this in bulk, and never more than once at a time.23:48
retr0hsoren: yeah, what you have done is quite hawt23:51
*** adiantum has joined #openstack23:52
*** pvo has quit IRC23:54
*** joearnold has joined #openstack23:56

Generated by 2.14.0 by Marius Gedminas - find it at!