Wednesday, 2012-01-04

*** lvaughn has joined #openstack00:00
*** deshantm has quit IRC00:00
*** deshantm_ is now known as deshantm00:00
*** clauden_ has quit IRC00:00
*** clauden_ has joined #openstack00:01
*** spiffxp has quit IRC00:03
*** bsza-II has quit IRC00:04
*** dolphm has joined #openstack00:09
*** bugfree has left #openstack00:11
*** ejat has quit IRC00:15
*** clauden__ has joined #openstack00:17
*** clauden_ has quit IRC00:17
*** dragondm has quit IRC00:18
uvirtbotNew bug: #911528 in tempest "nova tests dependent on specific ordering of services in keystone catalog" [Undecided,New]
uvirtbotNew bug: #911531 in horizon "PEP8 cleanup" [Undecided,New]
*** nati2 has quit IRC00:23
*** nati2 has joined #openstack00:24
*** willaerk has quit IRC00:24
*** clopez has joined #openstack00:30
*** dolphm has quit IRC00:34
*** spiffxp has joined #openstack00:36
*** freeflying has quit IRC00:38
*** freeflying has joined #openstack00:38
*** gyee has quit IRC00:40
*** rods has quit IRC00:42
*** rnorwood has quit IRC00:44
*** j3ll3 has joined #openstack00:48
*** j3ll3 has joined #openstack00:49
*** CrazyThinker has quit IRC00:53
*** dolphm has joined #openstack00:56
*** spackest has quit IRC00:56
*** Carmivore has quit IRC00:57
*** longhudou has joined #openstack00:57
*** livemoon has joined #openstack01:01
*** jakedahn has quit IRC01:02
*** Carmivore has joined #openstack01:02
*** po has quit IRC01:02
*** rustam has joined #openstack01:06
*** CrazyThinker has joined #openstack01:07
*** leifmadsen has quit IRC01:14
*** clopez has quit IRC01:15
uvirtbotNew bug: #911547 in glance "PEP8 cleanup" [Undecided,New]
*** ccc3 has joined #openstack01:17
*** spiffxp has quit IRC01:18
*** warik has quit IRC01:19
*** cheesecake has joined #openstack01:19
*** krow has joined #openstack01:21
*** tmichael has quit IRC01:21
*** smoser has joined #openstack01:22
*** warik has joined #openstack01:22
*** tungvs has joined #openstack01:22
*** daysmen has quit IRC01:24
tungvsHi all, can you show me how to configure nova-scheduler ? In the Diablo document, I'm interested in this "simple: In this method, hosts whose load is least are chosen to run the instance. The load information may be fetched from a load balancer.". Is a load balancer necessary ?01:24
*** rnorwood has joined #openstack01:25
*** ben_duyujie has joined #openstack01:28
*** ben_duyujie has left #openstack01:28
*** hadrian has quit IRC01:34
*** dendrobates is now known as dendro-afk01:35
*** dendro-afk is now known as dendrobates01:37
*** Ryan_Lane has joined #openstack01:40
uvirtbotNew bug: #911552 in nova "PEP8 python-novaclient cleanup" [Undecided,New]
*** devananda has quit IRC01:47
*** ejat has joined #openstack01:49
*** CrazyThinker has quit IRC01:53
*** jtanner has quit IRC01:54
*** mrjazzcat has quit IRC01:55
*** galthaus has joined #openstack02:00
*** dotdevops1 has quit IRC02:00
*** Carmivore has quit IRC02:01
uvirtbotNew bug: #911559 in nova "AMQPConnectionException on launch VM  after restart controller" [Undecided,New]
*** vladimir3p has quit IRC02:02
*** andrewbogott has quit IRC02:04
*** jdurgin has quit IRC02:05
*** CrazyThinker has joined #openstack02:05
*** Carmivore has joined #openstack02:06
*** rustam has quit IRC02:06
*** jakedahn has joined #openstack02:07
*** rustam has joined #openstack02:07
*** rustam has quit IRC02:08
*** rustam has joined #openstack02:08
*** rustam has quit IRC02:13
*** GonZo2000 has joined #openstack02:14
*** HugoKuo_ has joined #openstack02:15
*** GonZoPT has quit IRC02:16
*** HugoKuo__ has quit IRC02:19
uvirtbotNew bug: #911563 in keystone "PEP8 keystone cleanup" [Undecided,New]
*** westmaas has quit IRC02:22
*** ldlework has quit IRC02:22
*** ttx has quit IRC02:23
*** erkules|away has joined #openstack02:23
*** ironcamel2 has quit IRC02:23
*** erkules has quit IRC02:24
*** ttx has joined #openstack02:24
*** ttx has quit IRC02:24
*** ttx has joined #openstack02:24
*** jakedahn has quit IRC02:24
*** ironcamel2 has joined #openstack02:25
*** westmaas has joined #openstack02:25
*** Ryan_Lane has quit IRC02:31
*** oubiwann has quit IRC02:38
feiyuliucan't find the libvirt.h02:40
*** ccc3 has quit IRC02:47
*** dpippenger has quit IRC02:51
*** cheesecake has quit IRC02:53
*** dendrobates is now known as dendro-afk02:54
*** jakkudanieru has joined #openstack02:56
tungvsHi all, can you show me how to configure nova-scheduler ? In the Diablo document, I'm interested in this "simple: In this method, hosts whose load is least are chosen to run the instance. The load information may be fetched from a load balancer.". Is a load balancer necessary ?02:57
*** dolphm has quit IRC03:01
*** CrazyThinker has quit IRC03:01
uvirtbotNew bug: #911579 in swift "PEP8 swift cleanup" [Undecided,New]
*** krow has quit IRC03:07
*** micadeyeye has joined #openstack03:08
*** dolphm has joined #openstack03:08
*** micadeyeye_ has joined #openstack03:08
*** cheesecake has joined #openstack03:09
*** spiffxp has joined #openstack03:12
*** cheesecake has quit IRC03:14
*** CrazyThinker has joined #openstack03:16
*** CrazyThinker has joined #openstack03:17
*** shang has quit IRC03:29
*** ldlework has joined #openstack03:29
*** spiffxp has quit IRC03:30
*** dayou has quit IRC03:39
*** warik has quit IRC03:40
tungvsHas anyone run Windows instance with small or larger template ? THe storage which is added to the instance ruins partition order, so the instance cannot boot03:44
tungvsany work around now ?03:44
*** ecarlin has joined #openstack03:46
*** shang has joined #openstack03:46
*** ecarlin has quit IRC03:46
*** nati2 has quit IRC03:48
*** nati2 has joined #openstack03:49
*** dayou has joined #openstack03:52
mjforktungvs: not aware of that specific problem, did you check launchpad for similar issues?03:54
*** dolphm has quit IRC03:54
*** wariola has joined #openstack03:56
*** PiotrSikora has quit IRC03:56
tungvsThe cause is my guess, because I get an "a disk read error occur" when tried with small or larger template (which has storage). The tiny template works fine. mjfork, I'm searching03:57
*** PeteDaGuru has joined #openstack03:58
*** PiotrSikora has joined #openstack03:59
*** dolphm has joined #openstack03:59
mjforktungvs: if you don't find anything, open a new bug report with as many details as possible.04:10
*** pradeep has joined #openstack04:14
longhudoucan not connect to dashboard,04:16
longhudoumjfork: any tips for that?04:16
mjforklonghudou: is it running?04:16
longhudouI can see it from port 8004:17
longhudoubut whatever name and passwd I use,I can not connect04:17
mjforklonghudou: diablo keystone? devstack?04:18
longhudoudiablo keystone04:18
longhudouI install it manually04:19
mjforkstacktrace from dashbaord on login?04:20
mjforkanything in lgos04:20
longhudou WARNING [eventlet.wsgi.server] - - [04/Jan/2012 12:21:09] "POST /v2.0/tokens HTTP/1.1" 401 193 0.00710704:21
longhudouthis is from keystone.log04:21
*** deshantm has quit IRC04:21
mjforkkeystone-manage token list04:21
longhudounothing in it04:22
longhudoutoken   user    expiration      tenant -------------------------------------------------------------------------------04:23
mjforkyou need to create an admin token i am guessing04:23
mjforksomething liek04:23
mjforkkeystone-manage token add 999888777666 admin admin 2015-02-05T00:0004:24
longhudouERROR: 'NoneType' object has no attribute 'id'04:25
mjforklooks like all of keystone needs seutp04:25
mjforkdid this ever work?04:25
longhudouno,I am just installing04:26
longhudouhave not got it dowm04:26
mjforkreplace with your services IPs04:27
longhudouthis time ,it report"Unexpected error: list indices must be integers, not str"04:37
mjforklonghudou: what did?04:38
*** PeteDaGuru has quit IRC04:38
longhudouI enter the name as"admin"04:39
longhudoupasswd as"openstack"04:39
mjforkthe code i sent did admin/adin04:40
mjforkpaste the full stacktrace using stackbin04:40
longhudou WARNING [eventlet.wsgi.server] - - [04/Jan/2012 12:40:52] "POST /v2.0/tokens HTTP/1.1" 200 1088 0.19272504:42
longhudouin keystone.log04:42
mjforkno, dashboard log04:43
mjforkkeystone-manage endpointTemplates list04:48
mjforklonghudou: i need to leave04:49
longhudouok ,thanks ,I got to debug it by myself04:50
mjforki will be back online in ~8 hours if you need more.04:50
longhudounova    RegionOne glance  RegionOne keystone        RegionOne
mjforkkeystone should be identity04:50
mjforkdid you change any of the values in the paste I sent?04:51
mjforkother than IPs?04:51
mjforki would blow away keystone, start over04:52
mjforkand use this paste04:52
mjfork changing only IPs04:52
mjforkyo ucan also post in forums, i will respond there in the AM04:52
longhudouok , i will have a try04:52
longhudoumjfork: thanks04:52
*** CrazyThinker has quit IRC04:54
*** mjfork has quit IRC04:56
*** llang629 has joined #openstack04:59
*** llang629 has left #openstack04:59
*** abecc has quit IRC05:04
uvirtbotNew bug: #911599 in glance "Location field is wiped when updating metadata" [Undecided,New]
*** ejat has quit IRC05:08
*** FallenPegasus has joined #openstack05:09
*** tungvs has quit IRC05:09
*** CrazyThinker has joined #openstack05:09
*** tungvs has joined #openstack05:10
*** MarkAtwood has quit IRC05:11
*** ejat has joined #openstack05:14
*** ejat has joined #openstack05:14
*** dendro-afk is now known as dendrobates05:16
*** HowardRoark has quit IRC05:17
*** jeremy has quit IRC05:18
*** dotdevops has joined #openstack05:18
*** krow has joined #openstack05:21
*** FallenPegasus has quit IRC05:30
*** dolphm has quit IRC05:38
*** tungvs_ has joined #openstack05:40
*** tungvs has quit IRC05:40
*** lzyeval has joined #openstack05:42
lzyevalHi, does anybody know how to run the unittests on Swift?05:42
*** nati2_ has joined #openstack05:44
*** nati2 has quit IRC05:44
lzyevalgot it, nm05:45
*** rnorwood has quit IRC05:45
*** stoked has joined #openstack05:47
*** micadeyeye has quit IRC06:05
*** micadeyeye_ has quit IRC06:05
*** nati2_ has quit IRC06:09
*** nati2 has joined #openstack06:09
*** newopen has joined #openstack06:11
uvirtbotNew bug: #911616 in nova "windows image cannot start from template with storage" [Undecided,New]
*** jedi4ever has joined #openstack06:19
*** lzyeval has quit IRC06:21
*** dotdevops has quit IRC06:22
*** micadeyeye has joined #openstack06:24
*** nati2_ has joined #openstack06:24
*** nati2 has quit IRC06:25
*** ejat has quit IRC06:28
*** gasbakid has quit IRC06:29
*** krow has quit IRC06:30
*** HugoKuo__ has joined #openstack06:31
tungvs_hi, does anybody know how to configure nova-scheduler ?06:32
*** bepernoot has joined #openstack06:33
*** HugoKuo_ has quit IRC06:34
*** gasbakid has joined #openstack06:43
*** CrazyThinker has quit IRC06:46
*** bepernoot has quit IRC06:49
*** gasbakid has quit IRC06:51
*** nati2 has joined #openstack06:53
*** nati2_ has quit IRC06:56
uvirtbotNew bug: #911627 in swift "test.unit.proxy.test_server.TestController error" [Undecided,Invalid]
*** MarkAtwood has joined #openstack06:57
*** CrazyThinker has joined #openstack06:58
*** micadeyeye has quit IRC06:59
*** gasbakid has joined #openstack07:04
*** nyeates has quit IRC07:06
*** mindpixel has joined #openstack07:06
*** nati2_ has joined #openstack07:10
*** nati2 has quit IRC07:11
*** aculich has quit IRC07:15
*** micadeyeye has joined #openstack07:27
*** micadeyeye has quit IRC07:32
*** jeremy has joined #openstack07:33
*** tjikkun has quit IRC07:41
*** ldlework has quit IRC07:41
*** rocambole has joined #openstack07:41
*** martin has quit IRC07:42
*** martin_ has joined #openstack07:43
*** TREllis has quit IRC07:43
*** TREllis has joined #openstack07:43
*** tjikkun has joined #openstack07:44
*** tjikkun has joined #openstack07:44
*** bepernoot has joined #openstack07:47
*** gasbakid has quit IRC07:47
*** stoked has left #openstack07:48
*** marcuz has joined #openstack07:49
*** rasyid7 has quit IRC07:52
*** micadeyeye has joined #openstack07:58
*** reidrac has joined #openstack08:09
*** micadeyeye has quit IRC08:13
*** nati2 has joined #openstack08:13
*** nati2_ has quit IRC08:14
*** aculich has joined #openstack08:15
*** anpl has joined #openstack08:22
anplAnyone had installed Openstack using GridDynamics Repo on Sl 6.108:22
anplI have installed and not able to start keystone08:22
*** foexle has joined #openstack08:22
foexlehiho guys08:23
*** micadeyeye has joined #openstack08:25
*** bepernoot has quit IRC08:27
*** bepernoot has joined #openstack08:27
uvirtbotNew bug: #911655 in quantum "lxml missing in pip-requires" [Undecided,In progress]
*** arBmind has joined #openstack08:28
*** Remco_ has joined #openstack08:29
*** heyho has joined #openstack08:31
*** rocambole has quit IRC08:37
*** Remco_ has quit IRC08:38
*** rocambole has joined #openstack08:39
*** daysmen has joined #openstack08:43
*** eglynn has quit IRC08:43
uvirtbotNew bug: #911663 in quantum "PEP8 quantum cleanup" [Undecided,New]
*** oneiroi has joined #openstack08:49
*** tungvs_ has quit IRC08:49
*** tungvs_ has joined #openstack08:49
*** zul has quit IRC08:49
*** katkee has joined #openstack08:51
*** tungvs_ has quit IRC08:53
*** tungvs_ has joined #openstack08:53
*** uksysadmin has joined #openstack08:54
*** ton_katsu has joined #openstack08:54
*** cheesecake has joined #openstack08:56
*** cheesecake has quit IRC08:57
* uksysadmin is attempting the impossible today: getting Keystone working with Nova. ;-)09:01
uksysadminbtw - is it true that as soon as I use Keystone, I have to use the nova client? Euca2ools doesn't work?09:02
*** Hakon|mbp has joined #openstack09:07
*** slomkows has joined #openstack09:08
micadeyeyeplease advice the easiest way to start openstack (with all other services) on a linux box. Is there a command I can use?09:08
*** dnjaramba has joined #openstack09:10
micadeyeyefor svc in api objectstore compute network volume scheduler; do sudo service openstack-nova-$svc start; done09:10
foexledo any have a verry high load on a compute host if you work in an intance ? .... so i starte an instance (ubuntu11.10) and install packages in there. Works fine but my load on the host are fucking high but why ? It seems to be an I/O problme with kvm and qcow2 image ?09:12
micadeyeyefoexle, no09:12
*** zul has joined #openstack09:14
micadeyeyebut I would prefer a command that can start all services. i don't like using that long line.09:15
uksysadminfoexle, when a new instance is spawned it essentially does a copy of the image you've launched... if your disk is slow, cpu goes through the roof.09:15
uksysadminmicadeyeye, openstack is made up of many components - you can't just do "start openstack" unless you wrote a wrapper script that does what you want it to do.09:16
foexleuksysadmin: yes i know, but my disks are fast enough .... on a similar machine i started 20 instance with xen and no problems09:16
oneiroican glance provide an iscsi target? I seem to vaguely recall it can09:16
uksysadminit is designed to be highly scalable - not  everyone will install all OpenStack components on a single host (although I appreciate, for dev purposes its ideal)09:17
foexleuksysadmin: i start 1 instance with openstack and kvm => load is exploding09:17
heyhouksysadmin: I'm trying to use keystone with nova too.. and eucatools doesn't work anymore (at least for me)09:17
uksysadminwhat version of OpenStack?09:17
foexleuksysadmin: diablo09:18
uksysadminheyho, I noticed a conversation a while back that suggested you have to use Nova client and the Nova API... I'm not convinced that's the correct approach for OpenStack... but I'm biased - I like the world to live in an EC2-Api world.09:18
uksysadminfoexle, raise it as a bug - best way to get some investigation around it to see if its more widespread.09:19
*** jeremy has quit IRC09:19
foexlemaybe .... but first i'll try to find the problem09:20
heyhouksysadmin: yeah, It seems to me that EC2 Api are a lot easier..09:20
*** jeremy has joined #openstack09:21
uksysadminunfortunately for me I've been touting this great private cloud and how I can use the same tools to manage my internal cloud as well as my external cloud and suddenly Keystone changes that - breaks my vision of a hybrid cloud environment if I'm having to switch tools.09:22
heyhouksysadmin: I've installed keystone (but don't know how correctly), and horizon has very nice graphical style and approach..but as for now I've troubles with glance.. because the istances I created via glance are stuck in boot status and there are no specific instance folder in /var/lib/nova/instances .. And, recent new, glance-api services refuse to start :/09:26
uksysadminhow did you install keystone? what versions of OpenStack are you using?09:27
*** eglynn has joined #openstack09:27
heyhouksysadmin: I'm using 2011.3 and I've installed keystone following mjfork lineguides on his post here:
uksysadmincheers heyho, I'll give them a go.09:32
jakkudanieruHi, I have installed a openstack infra on ubuntu 11.1009:33
jakkudanieruand I am installing dashboard on the controller. I installed the distro package but whn I run syncdb I have a no modulenamed local.local_settings09:35
*** erkules|away has quit IRC09:35
*** erkules|away has joined #openstack09:35
*** erkules|away is now known as erkules09:35
jakkudanieruI did config the file and moved it from ../local/local_settings.py_example to ../local/local_settings.py09:36
jakkudanieruDoes somebody knows what the problem could be?09:37
*** rustam has joined #openstack09:39
heyhouksysadmin: you're welcome09:40
uksysadminheyho, although I'm trying the cloudbuilders packages first - I'm convinced someone out there has a usable apt-get install procedure! ;-)09:47
*** HugoKuo_ has joined #openstack09:48
*** koolhead17 has joined #openstack09:48
koolhead17hi all09:48
*** newopen has joined #openstack09:51
*** martin_ is now known as martin09:51
heyhouksysadmin: yes..I've also tried to follow this but something didnt work, maybe for a version mismatch09:52
*** HugoKuo__ has quit IRC09:53
*** eglynn has quit IRC09:53
uksysadminyeah giving up on cloudbuilders - I was running diablo from the ubuntu proposed packages and it has reverted a fix I needed regarding floating ips09:54
jakkudanieruI found the problem..a hardlink to ../local should be present in the dashboard directory09:54
uksysadminI keep hearing about installing from source is the best approach for keystone + horizon + glance09:55
uksysadminhello koolhead1709:55
koolhead17hola uksysadmin09:55
uksysadminjakkudanieru, probably best raise a bug regarding this so the package maintainers can update their process09:55
jakkudanieruThanks, I will09:56
*** rods has joined #openstack09:58
*** pixelbeat has joined #openstack09:58
uvirtbotNew bug: #911681 in glance "oparser missing in bin/glance-control" [Undecided,New]
*** eglynn has joined #openstack10:01
*** CrazyThinker has quit IRC10:05
*** darraghb has joined #openstack10:07
*** MarkAtwood has quit IRC10:09
*** wariola has quit IRC10:12
*** eglynn has quit IRC10:13
heyhoguys, how do you run openstack services? I mean, you put some commands in rc.local? I'm manually using "sudo service nova-x start" but I'd like an upstart job, maybe without requiring sudo10:15
*** dnjaramba has quit IRC10:16
*** CrazyThinker has joined #openstack10:17
*** livemoon has left #openstack10:18
*** dnjaramba has joined #openstack10:20
*** derekh has joined #openstack10:20
*** Ramonster has joined #openstack10:21
ttxheyho: Ubuntu packages come with upstart jobs.10:24
*** tungvs_ has quit IRC10:24
*** mnour has joined #openstack10:24
*** tungvs_ has joined #openstack10:24
*** micadeyeye has quit IRC10:25
heyhottx: when I shutdown and restart my cloud-controller node, nova-network, nova-compute, nova-scheduler, nova-volume don't startup automatically.. what am I missing?10:25
*** tungvs_ has quit IRC10:26
ttxheyho: what packages are you using ?10:26
*** tungvs_ has joined #openstack10:26
heyhottx: 2011.3 ones.. I simply installed via sudo apt-get install nova-common, nova-network, etc..10:29
ttxheyho: the ones in the main repository from Ubuntu 11.10 ?10:29
ttxheyho: or some PPA ?10:29
heyhottx: no ppa repositories added, the standard ones10:30
ttxheyho: those come with upstart scripts that should be started at boot10:30
ttxheyho: anything in the logs on a reboot ? Maybe there is some dependency (like DB or queue) unsatisfied at boot time10:31
ttxheyho: in all cases, if it works if you start it manually after boot, sounds like an Ubuntu/upstart-script issue, so you might have more luck asking in #ubuntu-cloud10:32
*** po has joined #openstack10:33
heyhottx: thanks for the info.. and do you know where are these scripts?10:34
ttxheyho: upstart scripts live in /etc/init/10:34
*** Emilien has joined #openstack10:35
ttxheyho: I bet they start at boot time, but fail for some reason that should be detailed in Nova logs.10:35
heyhottx: yeah they are all there.. I'll check again the nova logs10:36
uksysadminttx, do you know if the weekly essex openstack snapshots in precise are the ones that are just available in the main repos or is there something extra that needs adding to a precise install?10:39
*** shaon has joined #openstack10:42
ttxuksysadmin: there were some differences, but they were supposed to converge. zul should know10:43
ttxchecking that is somewhere down on my TODO10:43
uksysadmincheers - just happened to notice this info in there being some weekly snapshot packages.  I'm getting lazy in my old age ;-)10:43
*** jakkudanieru has quit IRC10:44
uksysadminthat will help a great deal with testing and reporting bugs10:44
ttxuksysadmin: ISTR they run it through some internal staging testing before publishing it.10:44
heyhottx: you were right again.. for example, nova-network didnt upstart and checking the log I received this ..10:45
uksysadmincool ttx10:45
ttxheyho: so it seems mysql is being a bit slow to start -- we should probably retry more aggressively10:46
uksysadminahh, heyho yeah I've seen that before10:46
ttxI think we have a bug filed on that10:46
uksysadminI think I've seen that too10:46
heyhoreally? that's conforting10:46
ttxheyho: actually the fix seems to be in oneiric-proposed10:47
ttxheyho: so enabling that and installing would let you confirm that the bug is gone in a future update.10:48
heyhoso, for now, I have to wait and start the failing services manually, right?10:48
uksysadminI'm running proposed - I'll add my $0.0210:48
heyhottx: what's oneiric proposed, a ppa? i should do add-apt-repositories?10:49
ttxheyho: oneric-proposed is an official ubuntu repository used to test updates before they hit ubuntu official updates (called oneiric-updates)10:49
uksysadminheyho, add:   deb oneiric-proposed main restricted universe multiverse to /etc/apt/sources.list, apt-get update, apt-get upgrade10:50
ttx(warning, every now and then you might catch a regression)10:51
heyhothanks guys, as you noticed I'm quite new in the linux world.. I appreciate your help10:51
* uksysadmin is more of a cowboy than ttx10:51
*** tungvs_ has quit IRC10:56
*** tungvs_ has joined #openstack10:56
*** pious has joined #openstack10:56
*** shaon has quit IRC10:56
*** longhudou has quit IRC10:56
*** dnjaramba has quit IRC10:56
piousanybody there?10:57
uksysadminheyho, I've just rebooted a oneiric host from the proposed repo running all services and it comes back fine.10:57
heyhoand do you know why I can't directly access /var/log/glance or /var/log/nova directly but only via sudo -s? I mean, should this be so restricted or do I have to use chown and chmod and change permissions on certain openstack directories..I think some of my issues depends on  access to openstack folder10:58
heyhouksysadmin: cool, I'm just trying it10:58
uksysadminthey're root owned10:58
heyhoand they should stay like that?10:58
*** dnjaramba has joined #openstack10:58
uksysadminfor security reasons yes10:58
ttxuksysadmin, heyho: it's always been a race, but if you reliably lost it and now it works, I guess you could confirm the fix on that bug10:59
newopenhi, can you explain me the network host10:59
uksysadminyour cloud host will be a haven for someone wanting to gain access - in a true prod environment you'd go out of your way to secure it as much you can11:00
heyhottx uksysadmin: yes, I got nova-network and the other services to upstart smoothly!11:00
uksysadminttx, I've not tested losing mysql and it reconnected - only a reboot where all services disappear anyway...11:00
heyhoexcept for glance-api and glance-registry.. but I think that's just me..I messed up them11:00
uksysadmincan do though11:00
piousfolks, I need some basic information about how how can I modify nova-scheduler, and test my changes into my experimental cluster.  I have already got the source, and have setup the machines.  I am not sure how can I make my changes into the nova schduler directory to take effect into my local setup.  I have looked around the Internet, but couldn't find how can I modify and test my changes.  Any pointers on this would be very helpful.11:01
uksysadmin==> /var/log/nova/nova-compute.log <==11:02
uksysadmin2012-01-04 11:02:00,064 ERROR nova [-] Recovered model server connection!11:02
uksysadmin==> /var/log/nova/nova-network.log <==11:02
uksysadmin2012-01-04 11:02:01,125 ERROR nova [-] Recovered model server connection!11:02
uksysadmin==> /var/log/nova/nova-scheduler.log <==11:02
uksysadmin2012-01-04 11:02:01,131 ERROR nova [-] Recovered model server connection!11:02
ttxuksysadmin: cool, please tag verification-done on that bug11:03
uksysadminwill do11:03
*** Remco_ has joined #openstack11:03
newopenwhat is the failure domain and why openstack speak about HA only in network?11:13
newopenwhich kind of problem new HA option can resolve11:17
*** feiyuliu has quit IRC11:18
*** katkee_ has joined #openstack11:18
*** feiyuliu has joined #openstack11:19
*** koolhead17 has quit IRC11:21
*** katkee has quit IRC11:22
*** tungvs_ has quit IRC11:24
*** micadeyeye has joined #openstack11:27
*** clopez has joined #openstack11:29
*** micadeyeye has quit IRC11:32
*** DuncanT has joined #openstack11:33
*** mikal has quit IRC11:37
*** eolo999 has joined #openstack11:41
*** mikal has joined #openstack11:41
*** micadeyeye has joined #openstack11:43
*** mikal has quit IRC11:45
*** MarcMorata has joined #openstack11:47
*** sidarta has joined #openstack11:53
*** pious has quit IRC11:55
*** clopez has quit IRC11:59
*** eglynn has joined #openstack12:01
*** ton_katsu has quit IRC12:02
micadeyeyeApiError: instance |1| has no fixed_ips. unable to associate floating ip12:03
micadeyeyeBut instance \1\ runs on my machine (localhost)12:04
*** pious has joined #openstack12:07
micadeyeye[micadeyeye@micadeyeye vmimages]$ euca-associate-address -i i-00000012
micadeyeyeInstanceNotFound: Instance i-00000012 could not be found.12:11
piousCan anyone here help me about how to modify the sources of nova-compute, especially, the sources in the scheduler directory of nova compute.  I want to modify the scheduler, and then test it on my experimental setup.  I am not a big python developer, and I am not sure how to run the modified sources on my experimental cluster.12:11
micadeyeyehow can i give instance 1 a fixed IP?12:12
micadeyeyekindly help.12:12
*** livemoon has joined #openstack12:14
*** dendrobates is now known as dendro-afk12:19
*** pradeep has quit IRC12:24
*** praefect has quit IRC12:25
*** clopez has joined #openstack12:26
*** Remco__ has joined #openstack12:26
*** praefect has joined #openstack12:26
uksysadminmicadeyeye, it is saying that instance doesn't exist - so it can't be found to allocate that ip12:26
*** micadeyeye has quit IRC12:28
*** Remco_ has quit IRC12:28
*** KAM has joined #openstack12:30
*** mjfork has joined #openstack12:31
*** sidarta has quit IRC12:32
*** dnjaramba has quit IRC12:33
*** dnjaramba has joined #openstack12:34
*** sidarta has joined #openstack12:40
*** paulmillar has joined #openstack12:42
*** markvoelker has joined #openstack12:42
*** paulmillar has quit IRC12:44
*** paulmillar has joined #openstack12:45
*** bsza has joined #openstack12:45
*** rgidron has joined #openstack12:48
rgidronsetup my openstack and everythi looks cool (describe_availability12:49
rgidronCan anyone help with newbie question?12:50
anplCan some one share me an updated doc for installing Openstack on Redhat variants12:50
*** livemoon has quit IRC12:50
uksysadminanpl, try this
*** nyeates has joined #openstack12:51
anplThanks uksysadmin I already started on that12:51
anplwe need to install Dashboard from source if I use EPEL right?12:51
rgidronanyone know this error: Unable to run euca-describe-images.  Is euca2ools environment set up?12:52
anplIt would be great if you can share some links on configuring Keystone as well12:52
rgidronI get this when running uec-publish-tarball12:52
mjforkanpl: for diablo + RHEL, I only had luck with dashbaord/keystone from source12:52
*** dayou has quit IRC12:53
anplThanks mjfork12:53
rgidronas far as I can tell my euca2ools environment is tip top..12:53
*** sidarta has quit IRC12:53
mjforkanpl: those are not sutable for production, but they will get keystone/dashboard working and usable\12:54
mjforkanpl you will also need to set up Keystone using this:
*** cloudgeek has joined #openstack12:55
mjforkrgidron: do you have keystone?12:55
anplThank you so much mjfork12:55
rgidronmjfork: newbie.. how do I find out?12:56
mjforkrgidron: how did you isntall openstack12:56
rgidronI followed the Diablo getting started guide12:56
mjforkthen you probably don't have keystone12:57
*** dayou has joined #openstack12:57
mjforkdid you create a user in nova?12:57
rgidronmjfork: yes12:57
rgidroni have a novarc file which works12:57
mjforkdid you use the command to get the zipfile?12:58
rgidronI run euca-describe-avaiability-zones and it worked fine12:58
rgidronI do have the zip file12:58
mjforkok, and unzip and then ". novarc" ?12:59
mjforkpastebin nova-api.log contents when you run euca-describe-images13:00
*** sidarta has joined #openstack13:03
anplopenstack-nova-db-setup is missing on EPEL13:03
rgidronmjfork: so execute euca-describe-images locally on the nova machine?13:04
*** slomkows has quit IRC13:04
*** leifmadsen has joined #openstack13:05
rgidronmjfork: 2012-01-04 15:06:49,742 ERROR nova.api [0fbf53cc-f9cf-4787-b6a3-1ee44b59164d novaadmin proj] Environment: {"CONTENT_TYPE": "application/x-www-form-urlencoded; charset=UTF-8", "SCRIPT_NAME": "/services/Cloud", "REQUEST_METHOD": "POST", "HTTP_HOST": "", "PATH_INFO": "/", "SERVER_PROTOCOL": "HTTP/1.0", "HTTP_USER_AGENT": "Boto/2.0 (linux2)", "SERVER_NAME": "", "REMOTE_ADDR": "", "wsgi.13:08
rgidron2012-01-04 15:06:49,743 ERROR nova.api [0fbf53cc-f9cf-4787-b6a3-1ee44b59164d novaadmin proj] UnknownError: An unknown error has occurred. Please try your request again. 2012-01-04 15:06:49,744 INFO nova.api [0fbf53cc-f9cf-4787-b6a3-1ee44b59164d novaadmin proj] 3.32393s POST /services/Cloud/ CloudController:DescribeImages 400 [Boto/2.0 (linux2)] application/x-www-form-urlencoded text/xml13:09
mjforkplease use paste.openstack.org13:09
rgidronmjfork - apologies, is that a dedicated web site?13:09
mjforkyes, http://paste.openstack.org13:10
mjforkyou can paste your logfile, it will give you a URL13:10
rgidronpaste the entire logfile or just the last bits?13:11
mjforklast lines relveant to your API call13:11
rgidronmjfork - got it thanks - link:
mjforkrgidron: looks like your glance server is not listening or confgiured wrong13:12
anplI had to use EPEL-testing repo to get the openstack-db-setup13:13
anplyum --enablerepo=epel-testing update openstack-nova13:13
mjforkwhat is the glance_api_servers setting in nova.conf13:13
*** derekh has quit IRC13:13
*** irahgel has joined #openstack13:14
*** pradeep1 has joined #openstack13:14
rgidronmjfork: glance_api_servers=
*** sandywalsh_ has joined #openstack13:15
mjforkrgidron: is that right server?13:16
*** redconnection has joined #openstack13:17
*** micadeyeye_ has joined #openstack13:17
*** micadeyeye has joined #openstack13:17
*** hadrian has joined #openstack13:17
rgidronmjfork: i"ll check on the guide I followed, I believe so13:17
mjforkthen SSH to that machine and make sure glance is running13:18
Birk_I'm having problems with nova volume when running this command: "sudo iscsiadm -m discovery -t sendtargets -p cpcaserver3" but i'm getting this error
mjforkrgidron: back in 45 mins or so13:18
*** olafont has joined #openstack13:19
mjforkBirk_: check this out in the forums
mjforkBirk_: i need to step away, but if I understand the solution at check the "host" field in the DB13:19
Birk_Ok, thanks !13:20
eolo999on step 10. of development_saio i get a 404; on 11. i get an "Account not found". What am I missing?13:20
anplmjfork: sorry to disturb you13:23
anplwhen I try to install the dependency for keystone13:23
anplyou have that line for Debian13:23
anplapt-get install -y python-pip gcc python-lxml libxml2 python-greenlet-dbg python-dev libsqlite3-dev libldap2-dev libssl-dev libxml2-dev libxslt1-dev libsasl2-dev python-httplib213:23
anplI could not get the python-greenlet-dbg from yum13:24
*** nyeates has quit IRC13:24
*** sidarta has quit IRC13:26
*** jtanner has joined #openstack13:27
*** Emilien has quit IRC13:27
*** dayou has quit IRC13:28
*** dayou has joined #openstack13:29
*** heut2008 has joined #openstack13:30
rgidronmjfork: Thanks for helping, it was a wrong glance server configuration as you thought. thank you very much!13:30
micadeyeyeis anyone help using boxgrinder? i asked it to push an image to openstack but i can't see the image in glance.13:30
micadeyeyeThis is a line from BG: I, [2012-01-04T15:22:22.536814 #7580]  INFO -- : Uploading and registering 'f16-basic-1.0-raw' appliance in OpenStack..13:30
micadeyeyeBut I can't find the image in /var/lib/nova/images/13:31
micadeyeyeand "glance index" doesn't show it13:31
kodapaAny progress on this: ?13:33
*** pious has quit IRC13:34
*** newopen has quit IRC13:34
uksysadminhey all - I've taken to trying the precise + essex w/keystone route and after configuring my api-paste.ini file I get the following error: ImportError: No module named metadatarequesthandler13:37
uksysadmincaused by this in my api-paste.ini I presume13:37
uksysadmin# EC2 #13:37
uksysadminuse = egg:Paste#urlmap13:37
uksysadminpipeline = logrequest totoken authtoken keystonecontext cloudrequest authorizer ec2executor13:38
uksysadminpipeline = logrequest totoken authtoken keystonecontext adminrequest authorizer ec2executor13:38
uksysadminpipeline = logrequest ec2md13:38
uksysadminpipeline = logrequest ec2ver13:38
uksysadminpaste.filter_factory = nova.api.ec2:RequestLogging.factory13:38
uksysadminpaste.filter_factory = nova.api.ec2:Lockout.factory13:38
uksysadminpaste.filter_factory = keystone.middleware.ec2_token:EC2Token.factory13:38
uksysadminpaste.filter_factory = nova.api.ec2:NoAuth.factory13:38
uksysadminpaste.filter_factory = nova.api.ec2:Authenticate.factory13:38
uksysadmincontroller =
uksysadminpaste.filter_factory = nova.api.ec2:Requestify.factory13:39
uksysadmincontroller = nova.api.ec2.admin.AdminController13:39
uksysadminpaste.filter_factory = nova.api.ec2:Requestify.factory13:39
uksysadminpaste.filter_factory = nova.api.ec2:Authorizer.factory13:39
uksysadminpaste.app_factory = nova.api.ec2:Executor.factory13:39
uksysadminpaste.app_factory = nova.api.ec2:Versions.factory13:39
uksysadminpaste.app_factory = nova.api.ec2.metadatarequesthandler:MetadataRequestHandler.factory13:39
*** vdo has joined #openstack13:39
uksysadmin# Openstack #13:39
uksysadminuse = egg:Paste#urlmap13:39
uksysadminpipeline = faultwrap authtoken keystonecontext ratelimit extensions osapiapp13:39
uksysadminpaste.filter_factory = nova.api.openstack:FaultWrapper.factory13:39
uksysadminpaste.filter_factory = nova.api.openstack.auth:AuthMiddleware.factory13:39
* uksysadmin apologises for that flood!13:39
uksysadminanyway - ifmy paste hadn't inadvertently grabbed all the file - the [app:ec2md] section mentions this metadatarequesthandler13:40
uksysadmincould've been worse - my copy buffer could've had something sensitive or worse in - like references to my abba library or somet ;-) [joke btw]13:40
* mnour Thought there is something error with my machine :D13:40
* uksysadmin gets his coat13:41
uksysadminanyhoo... so back to the issue... anybody else crazy enough to run the latest and greatest precise and essex-2 and when setting up keystone I now can't start my nova-api13:43
*** sandywalsh_ has quit IRC13:45
*** katkee_ has quit IRC13:46
foexleanyone have a solution for OSError: [Errno 2] No such file or directory: '/var/lib/nova/instances/Domain-0/console.ring' <--- xen ubuntu 11.1013:48
*** mgoldmann has joined #openstack13:52
*** hggdh has quit IRC13:52
*** Razique has joined #openstack13:53
foexleRazique: happy new year dude :)13:53
RaziqueHappy new year my friend :)13:53
RaziqueHappy new year to all openstackers too ^^13:53
Raziquefoexle: how did it went for u ?13:54
uksysadminhey Razique , thanks for answer the q  about euca2ools and keystone on the ML13:54
*** kbringard has joined #openstack13:55
Raziqueuksysadmin: you welcome :) don't hesitate if it doesn't work :)13:55
foexleRazique: the new year was started as the old one :D .... much beer and many problems hahahaha13:56
*** irahgel has left #openstack13:56
anplcan some one help me with the command to install python dependecies for keystone and dashboard on RHEL13:56
Raziquehere very calm, not much of booze, small party13:56
Raziqueand sorted :p13:56
uksysadminnow if I can just get Keystone to work I'd tell you! ;-)13:56
foexlenext time you come to berlin !13:57
Raziqueuksysadmin: sure I'll do my best, I don't use it u know :p13:57
tightworkWould PowerDNS be a good candidate for floating DNS in reference to
tightworkConsidering PowerDNS already is database backed13:57
Raziquefoexle: ahah sure looks like that was better than here ;)13:57
*** bcwaldon has joined #openstack13:57
foexleRazique: i think so ;) ...13:57
eolo999any example of keystone configuration file for swift?13:58
tightworkWhat is being used for official development? github, launchpad? bzr, git?13:59
*** sandywalsh_ has joined #openstack13:59
foexleRazique: i've verry verry big problems with my instances and his performance .... if i install a package in an instance my load will increse to 25 on the compute node .... i'm looking for a solution since many days ... but i cant find any issue. do you have the same problem in your cloud ?13:59
foexlewith kvm13:59
*** aliguori has joined #openstack14:00
Razique25 percent :o14:00
tightworkfoexle:  what process is the load? ksm or the kvm process itself?14:00
Raziquefoexle: the only instances -> hypervisor perf impact I've notice14:01
foexleits an I/O thing14:01
foexlebut dont know why14:01
Raziqueis when a huge network transfer happens on an instance, the load increases14:01
Razique(on the hypervisor)14:01
Raziqueexcept that, the isolation is rox stable14:01
foexle°° thats normal ?14:01
*** rnorwood has joined #openstack14:01
*** PotHix has joined #openstack14:02
foexletightwork: kvm14:02
Raziquefoexle: I'd say yes, must be a virtio stuff14:02
foexleomg .... ok i'll try xen ^^14:02
Raziquewe are talking about network saturation imho14:03
tightworkOveral perf issues, id definitely say get collectd going.. just taking shots in the dark without real trend analysis.14:03
Raziqueand the load increases is not that much14:03
*** kbringard has quit IRC14:03
*** rnorwood has quit IRC14:03
*** heut2008 has quit IRC14:04
Raziquefoexle: I've already noticed that behaviour on my previous farm14:05
Raziquewhich was using kvm + libvirt14:05
foexlebut its verry strange .... i start 5 instance on the same node and the node load will increase to 120 :D14:05
foexleWTF ^^14:05
foexlewhat do you use now ?14:05
Raziquewhat's the cpu on the node ?14:05
Raziquenow I use the openstack as upper layer :p14:06
foexle8 core ( 2x xeon)14:06
tightworkhmm I'm not seeing that, I just spun up 10 ubuntu's thing was fine14:06
tightworkquad xeon14:06
Raziquetightwork: same here14:06
foexleqcow2 images?14:06
Raziqueon a core quad (pre-prod) I'm able to launch around 20 instances @ the same time, with a lot which remains lower to urs14:06
tightworknot sure14:07
tightworkopenstack newb14:07
foexleAHHHHHH why i have only this fucking issues :(14:07
*** heut2008 has joined #openstack14:08
uvirtbotNew bug: #909568 in swift "The fallback json parser may not be safe" [Low,Confirmed]
*** HugoKuo__ has joined #openstack14:08
Raziquefoexle: can u run manually an instance ?14:08
Raziquevia virsh14:08
Raziquelike, create a new one14:08
Raziquecheck the load before, the running process14:09
mjforkfoexle: on the I/O issues, are you writing to a non-allocated drive that is contintually extending the disk?14:09
Raziquethe I/O14:09
*** kbringard has joined #openstack14:09
foexlemjfork: i'm starting an instance with an qcow2 image from glance.... nova compute create a "local" image on the disk14:09
foexleand this will be increase i think14:10
mjforkso you are writing to that second disk14:10
foexlei've 2 disk and disk.local14:10
foexleand i'll write to disk14:10
mjforkfoexle: are you on Ubuntu? RHEL?14:11
mjforkwhat do one of your running VMs look like in PS14:11
foexlebut yeah ur right .... it's not allocated. maybe 9gb ...14:12
*** HugoKuo_ has quit IRC14:12
mjforklook in nova/instances/_base14:12
foexleit will increase the image size permanently14:12
mjforkdo you see an ephemeral_* entries?14:13
foexleahhh hmmm ... ok there are files with 10G14:13
*** AlanClark has joined #openstack14:13
anplmjfork: I could not complete your steps to install Keystone and Dashboard on Scientfic Linux14:14
anplthe dependencies failed14:14
*** deshantm has joined #openstack14:15
foexlemjfork: it's only one ephemeral_0_20_None14:15
mjforkanpl: which?14:15
mjforkfoexle: my guess is they are sparse files?14:15
mjforkfoexle: compare du -h vs ls -h14:15
*** derekh has joined #openstack14:16
*** |miska| has left #openstack14:17
mjforkanpl: was that the full run?14:20
mjforkfoexle: du -h /var/lib/nova/instances/_base/ephemeral_0_20_None14:20
mjforkn/m, i know it is less than 2014:21
foexleroot@compute1:~# du -h /var/lib/nova/instances/_base/ephemeral_0_20_None14:21
mjforkyou could try making that a fully allocated disk to see if your IO problems go awawy14:21
anplmjfork: sorry I did not understand what does full run means14:21
mjforkit will be SLOOOOW to provision since you need to copy 20GB fiule14:21
mjforkfoexle: but would tell you if that was the issue14:21
mjforkfoexle: also need to check you are using virtio drivers14:21
mjforkfoexle: may not hurt to pop over to a KVM channel and ask14:22
mjforkanpl: that looks like partial log, not full log14:22
anploh ok14:22
*** dprince has joined #openstack14:22
anpllet me paste the full log14:22
foexlemjfork: hmmm ok i'll try that .... but i'm asking me ... why i have only this problem :( ... it's not a big deal to install a compute node.14:23
mjforkfoexle: check with #kvm14:24
*** ejat has joined #openstack14:24
foexlemjfork: but thanks .... i'll try your proposal14:24
anplmjfork: how can I get the full log14:24
anplthe screen is scrolling down14:24
*** foexle has quit IRC14:24
mjforkrun the "comand > <log file>"14:25
anplsorry for asking basic questions like this14:25
*** primeministerp has joined #openstack14:25
anpli did that14:25
anplbut it's not full14:25
anplpython tools/  > a.txt14:25
mjforkclose enough - pastebin that14:26
*** mchenetz has joined #openstack14:28
anplmjfork: do you need  /root/.pip/pip.log as well14:28
mjforkanpl: i did these steps on ubunut14:28
mjforkgive me a few14:28
mjforkont the phone14:28
anplmjfork: thanks14:29
mjforkanpl: pastebin it please14:29
*** zul has quit IRC14:32
*** foexle has joined #openstack14:32
anplmjfork: I have added pycrypto at the end of pip-requires14:32
*** zul has joined #openstack14:33
anplmjfork: I changed the line on pip-requires to -e git+
anplnow the installation is progressing I feel14:34
*** feiyuliu has quit IRC14:34
mjforkanpl: reported by me :-)14:35
anplmjfork: :)14:35
anplmjfork: now I have different error14:35
anplplease wait I will share pastie14:36
*** arBmind_ has joined #openstack14:36
*** paulmillar has quit IRC14:38
mjforkneed the python source14:38
anplmjfork: pip.log
mjforkyum install python-devel i think14:38
*** arBmind has quit IRC14:38
*** arBmind_ is now known as arBmind14:38
*** mdomsch has joined #openstack14:39
*** imsplitbit has joined #openstack14:40
micadeyeyewhere is the python-passlib lib?14:42
micadeyeyeError: Package: openstack-keystone-2011.3.1-2.fc16.noarch (updates)14:42
micadeyeye           Requires: python-passlib14:42
eolo999\me is sad for the state of documentation14:42
* eolo999 is sad for the state of documentation14:42
anplmjfork: that worked14:42
anplmjfor: : Thank you so much14:42
micadeyeyesudo yum install python-passlib[sudo] password for micadeyeye:14:43
micadeyeyeLoaded plugins: langpacks, presto, refresh-packagekit14:43
micadeyeyeSetting up Install Process14:43
micadeyeyeNo package python-passlib available.14:43
micadeyeyeError: Nothing to do14:43
eolo999in particular for its dispersion14:43
anplnew error on loggin in with dashboard14:43
anplUnable to get usage info: This error may be caused by a misconfigured nova url in keystone's service catalog, or by missing openstackx extensions in nova. See the dashboard README.14:43
mjforkanpl: install openstackx on compute14:43
*** PeteDaGuru has joined #openstack14:44
*** krish has joined #openstack14:44
anplmjfork: I used EPEL testing repo14:44
anplit doesnt got openstackx14:44
*** tryggvil_ has joined #openstack14:45
*** lts has joined #openstack14:45
*** jokar has joined #openstack14:46
anplhow can I install Openstackx on Redhat14:47
*** chuck_ has joined #openstack14:48
*** zul has quit IRC14:48
*** chuck_ is now known as zul14:49
*** zul has joined #openstack14:49
*** bobi111 has joined #openstack14:49
anplmjfor: do you have any installation steps for openstackx14:49
anpland it says that Openstackx is being deleted in favor of python-novaclient. We are working on moving all the functionality into nova core and novaclient.14:50
anplis it so14:51
*** dspano has joined #openstack14:51
mjforkanpl: i got it from the griddynamics repo14:52
*** rnirmal has joined #openstack14:53
anplI had tried the setup with Griddynamic and wit Diablo everything went fine but I was not able to start keystone service14:53
mjforkkeystone packages are broken for diablo14:53
*** troytoman-away is now known as troytoman14:53
*** rnorwood has joined #openstack14:54
mjforki did nova + glance from package14:54
mjforkkeystone + dashbaord from source14:54
anplthanks mjfork, let me try that way14:54
*** zul has quit IRC14:54
anplmjfork: I will do an installation from scratch using Griddynamics repo14:56
anpland install Dashboard and Keystone from source as you suggested14:56
*** robbiew has joined #openstack14:57
*** hggdh has joined #openstack14:58
anplmjfork: thank you so much14:59
*** zul has joined #openstack14:59
mjforkanpl: good luck14:59
heyhoI'm stuck to a problem with glance (I think) unable to launch instances from the dashboard.. the main problem is that nothing gets created in /var/lib/nova/instances folder.. Also, glance-api and glance-registry can't start properly but glance-control all start succed14:59
Madkisscould somebody do a wild guess on what exactly is going wrong here?
* uksysadmin is about to resign to the fact the only way to instal Keystone is by source, but is also lazy...15:04
uksysadminwho do I need to pester in the Ubuntu ranks to get the working source install of Keystone made into deb packages?! ;-)15:04
uksysadminMadkiss, what do the logs show of the nova host you're connecting to?15:07
kodapaAnyone got dashboard working with diablo on ubuntu 11.10?15:11
kodapaI remember getting it to work using a PPA, but I can't remember which one15:11
*** nati2 has quit IRC15:11
*** nati2_ has joined #openstack15:11
uksysadminwhen the going gets tough, the tough goes for coffee... or java... Eucalyptus is written in Java... hmmm!... ;-)15:12
*** objectiveous has quit IRC15:13
Spirilisspeaking of coffee, I've had none this morning15:14
*** shaon has joined #openstack15:15
*** mnour has quit IRC15:15
*** ldlework has joined #openstack15:16
*** mrjazzcat has joined #openstack15:17
zuluksysadmin: whats not working?15:18
Madkissuksysadmin: nothing suspicious, really15:19
uksysadminI can't seem to find a suitable api-paste.ini file - docs reference the latest (so I've gone to the latest/trunk) and still no joy... anywhere else docs make vague references to setting up for use by nova15:19
zuluksysadmin: you might want to bug adam_g15:19
*** shaon has quit IRC15:20
uksysadmincheers zul15:20
uksysadminI'm only frustrated because it's bleeding edge stuff and that development is bound to be ahead of packaging... BUT... having a repeatable, documented procedure that follows where nova is up to now helps15:21
uksysadminAND (while I'm ranting) given that keystone is a pre-req to horizon AND we all like nice interfaces to our cloud environments (its a great sell to the execs where their investments run) means we have to get over this hurdle of built packages available that can be used but don't work15:22
* uksysadmin stops ranting now. promise. and I won't mention Euca again unless referring to the euca2ools ;-)15:23
uvirtbotNew bug: #911805 in glance "Scriptify the generation of man pages" [Undecided,New]
*** blamar has joined #openstack15:23
Birk_mjfork: i don't think that my problem is the host field in the database, because this field is filled with the host of the nova controller15:24
Birk_in my case cpcaserver315:24
Madkissuksysadmin: I am in the very same boat you are.15:24
mjforkuksysadmin: which paste file?15:25
mjforkBirk_: should it be the host of the volume?15:25
heyhocan anyone guess what could be the problem here?
Madkissmjfork: /etc/nova/api-paste.ini15:25
Birk_mjfork: yes15:26
mjforkBirk_: did yuo update it?15:26
mjforkMadkiss: i have it for diablo nova, here is what mine is15:26
uksysadminto be fair - lots of people will be - we're used to things working - that's a testament to lots of things, particularly around the established distros... and even though openstack isn't ready for production - we have to prepare environments where they're close to prod... and that doesn't involve running development tools and having to craft init scripts when in a few months we'll be replacing these with official packages.15:26
uksysadmin(sorry - carried on my rant)15:26
Birk_no, because the host field was already with this information15:26
uksysadminmjfork, api-paste.ini15:26
Birk_and I can attach a volume to a VM that is running in the machine with all nova components, but when i try to attach a volume to the VM that is running in the machine with just nova compute i got that error.15:26
Madkissmjfork: diablo nova, to my knowledge, won't work with diablo keystone properly, because the current keystone packages for 11.10 have a changed API (and 11.10 didn't have keystone out of the box)15:27
uksysadminthe problem is the ec2md paste section15:27
uksysadminpaste.app_factory = nova.api.ec2.metadatarequesthandler:MetadataRequestHandler.factory15:27
mjforkMadkiss: i installed dashboard/keystone it from source15:27
Madkisswhy don't you just delete the ec2-section, uksysadmin?15:27
uksysadmincauses ImportError: No module named metadatarequesthandler in nova-api15:27
mjforkuksysadmin: diablo or essex15:28
uksysadmindiablo and essex15:28
mjforkBirk_: paste output of select * from volumes;15:28
*** dubsquared has joined #openstack15:28
*** drewlander has joined #openstack15:28
mjforkuksysadmin: i can speak to diablo, which i have working with nova/galnce from packages and dashboard/keystoen from source15:28
uksysadmindiablo is fine - something will be great.15:29
Madkissmjfork: which version of keystone did you actually compile?15:29
uksysadminwhich is why, to be fair, I was about to grab a coffee and do the steps from source as you suggested15:29
mjforkMadkiss: diablo/stable15:29
uksysadmin(finally - I wanted to exhaust all avenues in packages before giving up)15:29
mjforkuksysadmin: have you tried the ubuntu packages from RCB suggested on th forums?15:30
uksysadminyeah - but I'm using proposed packages to fix a couple of issues - floating_ip assignment and I found out today it fixes the mysql start up /race condition causing services to not start on boot properly - rcb reverts this)15:30
* uksysadmin wants the earth ;-)15:31
Madkissyou are using the milestone-proposed packages?15:31
uksysadminno - oneiric-proposed15:32
mjforkuksysadmin: did the diablo packages from RCB have the keystone issue?15:32
uksysadminmjfork, not tried it - I could do a fresh install and try them - but as I had a working nova, it broke some stuff as ^^ above15:32
uksysadminmjfork, that paste file is the one I'm using15:33
mjforkuksysadmin: if you don't mind keystone/dashboard from source15:34
mjforkthen you can probably just go that route15:34
*** mchenetz has quit IRC15:34
*** mchenetz has joined #openstack15:34
uksysadmingoing to have to - and assume that down the line the installation through packages will simply catch up15:34
* uksysadmin gets coffee15:36
*** laurent\ has quit IRC15:36
Madkisskeystone      1.0~d4~20110909.1108-0ubuntu3.1 015:37
MadkissI wonder whether that's supposed to work with Diablo.15:37
*** jaypipes has joined #openstack15:40
*** n81 has joined #openstack15:40
*** n81 is now known as n8815:40
n88has anyone seen this error when trying to launch an instance: TRACE: RemoteError: AMQPConnectionException (503, u"COMMAND_INVALID - second '' seen", (20, 10), '')15:40
kodapaAnyone got dashboard working with diablo on ubuntu 11.10?15:41
mjforkkodapa: i do (from diablo/stable source)15:41
Birk_mjfork: here is the result of the select: Sorry for the bad formatation, i don't know how to make a output file with headers in mysql. There are 2 volumes, the first is atached to an instance running in the cpcaserver3.15:42
kodapamjfork: I have nova-api etc installed from oneric repos15:42
kodapamjfork: and keystone and horizon from ppa diablo/stable branch15:43
*** dolphm has joined #openstack15:43
n88nevermind, seems to already be reported:
mjforkBirk_: nova-volume is running on cpcaserve3?15:43
kodapamjfork: I can login to dashboard, but it says Error: Unable to get service info: This error may be caused by a misconfigured Nova url in keystone's service catalog, or by missing openstackx extensions in Nova. See the Horizon READ15:43
*** ttx has quit IRC15:43
*** ttx has joined #openstack15:43
*** ttx has quit IRC15:43
*** ttx has joined #openstack15:43
mjforkkodapa install openstackx on nova15:43
kodapamjfork: I fixed this error before by installing nova-api and openstackx from a PPA, but I can't remember which one15:44
Birk_mjfork: yes, nova-volume is running on cpcaserver315:44
kodapamjfork: I have openstackx from oneric repos15:44
mjforkwhat is your extensions dir in nova.conf15:44
*** Shentonfreude has joined #openstack15:44
mjforkBirk_: pastebin your nova-volume.log15:44
kodapamjfork: there's the problem; I can't find the extensions dir15:44
*** ttx has quit IRC15:45
*** ttx has joined #openstack15:45
*** ttx has quit IRC15:45
*** ttx has joined #openstack15:45
*** blamar has quit IRC15:46
*** sidarta has joined #openstack15:47
*** blamar has joined #openstack15:47
mjforkkodapa in conf file or?15:48
kodapamjfork: system15:48
kodapamjfork: even that I installed openstackx from apt, I cannot find the extensions dir15:49
mjforkkodapa ok, what is extesnsions setting in nova.conf15:49
mjforkkodapa locate a file named admin.py15:49
kodapamjfork: installed openstackx from git, branch diablo15:51
kodapamjfork: and put --osapi_extensions_path=/usr/local/lib/python2.7/dist-packages/openstackx-0.2-py2.7.egg/extensions/ in nova.conf15:51
kodapamjfork: still same error :S15:51
mjforkpastebin ls -l /usr/local/lib/python2.7/dist-packages/openstackx-0.2-py2.7.egg/extensions/ in nova.conf15:52
Madkisskeystone.middleware.ec2_token:EC2Token.factory -- was that one added in a version of keystone later than 1.0-d4?15:54
mjforkkodapa did you restart nova-compute and nova-api after changing the extensisno15:54
Birk_mjfork: follow the nova-volume.log:
kodapamjfork: of course15:55
Birk_i found just some informations about the creation and exclusion of volumes15:55
*** cloudgeek has quit IRC15:56
mjforkkodapa pastebin nova-api and nova-compute15:56
mjforkBirk_: did it go ok that time?15:57
*** dragondm has joined #openstack15:57
heyhoI've tried to delete my pending instances via nova delete.. I specify username and apikey but the response is "Invalid Openstack Nova Credentials" is Keystone messing up?15:57
Birk_mjfork: nop. The error is in the nova-compute.log on the cpcaserver2 (just nova-compute)15:58
mjforkBirk_: send that15:58
mjforkkodapa pastebin nova-compute.log16:01
mjforkmakes ure debug is on in conf file16:01
kodapawhy nova-compute16:01
kodapait just says Updating host status16:02
*** zzed has joined #openstack16:02
mjforkkodapa on startup it should log all kidns of info related to extesnsions16:03
Birk_mjfork: follow
mjforkBirk_: does cpcaserver3 resolve? and is that where your nova-volume is running16:06
*** GheRivero_ has joined #openstack16:06
Birk_yes, I can ping cpcaserver3 ( outside openstack network. Nova-Volume is running in cpcaserver3.16:10
*** freeflyi1g has joined #openstack16:10
mjforkcan you do it from this host16:12
Birk_I can attach volumes to VM that are running in cpcaserver3, if i try to attach volumes to VM that are running in cpcaserver2 (nova-compute only) i got that error on the nova-compute.log.16:14
*** freeflying has quit IRC16:14
mjforkfrom cpcaserver2 , can you "ping cpcaserver3"16:14
mjforkand on cpcaserver2 , run "sudo iscsiadm -m discovery -t sendtargets -p cpcaserver3"16:15
kodapamjfork: got it working, was my keystone endpoints that was fucked :P16:16
kodapamjfork: thanks anyway16:16
Birk_mjfork: yes, i can ping cpcaserver3 from cpcaserver2. The output of this command is iscsiadm: Login I/O error, failed to receive a PDU iscsiadm: retrying discovery login to iscsiadm: Connection to Discovery Address closed.16:16
Birk_the same as in .log16:16
mjforkis = cpcaserver316:17
*** NONONO has quit IRC16:17
Birk_mjfor: yes16:17
*** NONONO has joined #openstack16:17
mjforkwhat if you run "sudo iscsiadm -m discovery -t sendtargets -p cpcaserver3" on cpcaserver316:18
Birk_i got no output whe running this on cpcaserver316:19
*** bradm has quit IRC16:20
*** rgidron has quit IRC16:20
*** bobi111 has quit IRC16:20
*** jj0hns0n has joined #openstack16:21
*** nyeates has joined #openstack16:22
mjforkBirk_: try this thread?
*** aliguori has quit IRC16:23
Birk_mjfork: I will read it, thanks !16:23
*** nati2 has joined #openstack16:23
mjforkBirk_: as a workaround, manually set the provider_location in the DB16:24
*** rnorwood has quit IRC16:24
mjforkBirk_: format is ‘<ip>:<port>,<portal> <target IQN>’16:25
*** nati2 has quit IRC16:25
Birk_mjfork: ok16:25
*** nati2_ has quit IRC16:25
*** nati2 has joined #openstack16:25
*** rnorwood has joined #openstack16:27
*** reidrac has quit IRC16:27
heyhoDoes anyone know what's wrong in this command? sudo nova --user="admin" --apikey="admin" delete 0 Invalid OpenStack Nova credentials.  I set keystone to use admin user with admin password..16:28
dspanoI've searched high and I've searched low, but I have not been able to find if keystone will be ported to Ubuntu 10.04. Does anyone know if it will?16:29
dspanoIt looks like it was at one time, but nothing since August.16:30
*** mindpixel has quit IRC16:31
mjforkheyho: need a token, not a password16:32
mjforkheyho: what is nova-manage token list16:32
mjforkerrrrrr keystone-manage token list16:32
mjforkdspano: production or "test" use?16:32
tightworkIs it possible to auto assign floating ip on creation?16:32
*** heut2008 has quit IRC16:33
tightworkauto_assign_floating_ip ?16:33
uksysadmintightwork, yes16:33
uksysadminand I didn't even have to pay him16:34
kbringarduksysadmin: gotta love it when people reference your blog, eh?16:34
* uksysadmin 's ego just grew a bit16:34
mjforkdspano: i can give you directions for source for diablop16:34
dspanomjfork: I know I can install from the bazaar repos, but I wasn't sure if they're going to build .debs for it eventually.16:34
mjforkdspano: not sure16:34
uksysadminI need to update it tbh - there's an easier route16:34
uksysadminthe oneiric-proposed deb repo has the fix in now16:34
kbringarduksysadmin: what about the diablo-stable?16:35
uksysadminnot sure - I tested the ubuntu repo packages that are lined up for release for oneiric users at some point16:35
kbringardah, cool no worries16:35
dspanomjfork: I'm cool with using the deprecated authentication, I'm just afraid, I'll paint myself into a corner once I get my cloud into production.16:35
kbringardwas just curious :-)16:35
*** galthaus has quit IRC16:36
*** aliguori has joined #openstack16:36
uksysadminttx, /might/ know though on updated packages to diablo-stable ? or #ubuntu-cloud will16:36
ttxhmm ?16:37
uksysadminmaybe - maybe not the ubuntu channel16:37
kbringardeh, I'll just check the git repo and see if it's been backported16:37
* kbringard is lazy16:37
uksysadminsorry ttx - are there any proposed updates to diablo-stable repo planned?16:37
uksysadminI presume?16:37
ttxwhich ppa ? The openstack-release/2011.3 one ?16:38
*** bepernoot has quit IRC16:38
uksysadminkbringard ? (I'm half in this conversation and nearly a break through with keystone...)16:39
*** nati2 has quit IRC16:39
kbringardoh, sorry, I don't even mean a repo, I just mean backported in the github16:40
kbringardbut like I said, I can look myself, I didn't mean to start a whole thing :-)16:40
ttxkbringard: you mean the stable/diablo branch ? Yes, backports happen there all the time16:40
uksysadminlol - my fault - I see the release manager and pounce ;-)16:40
ttx\ignore uksysadmin16:40
* uksysadmin just shipped some Black Rat over to ttx16:41
kbringardttx: right, yea, I was just asking if uksysadmin knew if the patch for the auto floating IP had been pulled in16:41
ttxuksysadmin: that's nasty.16:41
kbringardbut I can check myself, I just thought someone may know off the top of their head16:41
kbringardsorry to distract y'all16:41
uksysadminI'm sorry - was hitting below the belt then16:41
*** jokar has left #openstack16:43
*** GheRivero_ has quit IRC16:44
*** yshh has joined #openstack16:45
mjforkany one know a rough # of cactus downloads? I  know at the conference in October, Diablo had 50,000+16:45
heyhomjfork: I've tried all the tokens in keystone-manage list but I always get invalid credentials16:47
mjforkheyho: pastebin output of token list and teh exact command you are running16:48
uksysadminbreakthrough with keystone trunk packages was short-lived... at least I got a different error. :-S16:49
uksysadmin(yeah yeah - I'll go for the source install)16:49
*** bradm has joined #openstack16:49
heyhomjfork: here you re
mjforkheyho all installed on one baox\16:50
heyhomjfork: yes16:50
*** rgidron has joined #openstack16:50
*** MarkAtwood has joined #openstack16:51
heyhomjfork: and I can view  image list via Glance.. Glance is up with the command "sudo glance-controll all start" but "sudo service start glance-api" and "sudo service start glance-registry" can't startup16:52
mjforkheyho: speciffy version of 1.116:52
heyhomjfork: where?16:53
heyhomjfork: in nova.conf, for nova or for keystone?16:53
rgidronmjfork: thanks so much for helping with the glance server configuration issue before16:54
*** tryggvil_ has quit IRC16:55
mjforkheyho:  in your command16:55
eolo999In the github's keystone readme: "Use swift to check everything works (note: you currently have to create a container or upload something as your first action to have the account created; there's a Swift bug to be fixed soon):"16:57
rgidroncan anyone help with a failed uec-publish-tarball (
eolo999So how do I currently create a container???16:57
eolo999I try with: % curl -X PUT -D - -H "X-Auth-Token: 999888777666" http://localhost:5000/v1/new_container16:58
eolo999but i get a 300 Multiple Choices response16:58
mjforkrgidron: i would use galnce direct like this
*** donaldngo_hp has joined #openstack16:59
*** donaldngo_hp has quit IRC16:59
*** donaldngo_hp has joined #openstack16:59
*** nati2 has joined #openstack16:59
*** laurent\ has joined #openstack16:59
*** rnorwood has quit IRC17:00
*** oubiwann has joined #openstack17:00
*** mattray has joined #openstack17:00
rgidronmjfork: OK, thanks17:01
*** uksysadmin has quit IRC17:01
heyhomjfork: sudo nova --username admin --apikey 57a0004a-3b82-4169-9c8b-3026fcf10af6 --version 1.1 --projectid 1 --url  delete 0 the response is: Expecting auth (HTTP 400)17:01
mjforkheyho: i am trying this now, but i think URL needs to tbe the nova API, not keystone17:01
*** maplebed has joined #openstack17:02
*** GheRivero_ has joined #openstack17:02
heyhomjfork: now I got invalid credentials again.. :(17:03
heyhowell, I'll try again tomorrow with a fresh mind.. bye and thansk17:04
*** heyho has quit IRC17:04
eolo999help, plz17:05
*** donaldngo_hp has quit IRC17:05
*** rnorwood has joined #openstack17:06
*** donaldngo_hp has joined #openstack17:06
*** vdo has quit IRC17:07
*** marcuz has quit IRC17:07
*** Gordonz has joined #openstack17:09
*** whenry has joined #openstack17:09
tightworkI set the floating ip fix, but still had pending build problems. Turns out all the addresses were allocated. No error message from horizon though, got it from nova-compute.log.17:12
*** jog0 has joined #openstack17:14
*** mchenetz has quit IRC17:14
*** aculich has quit IRC17:14
*** joesavak has joined #openstack17:16
*** warik has joined #openstack17:17
*** shaon has joined #openstack17:18
uvirtbotNew bug: #911880 in nova "nova-rootwrap causes Tracebook" [Undecided,New]
*** hugokuo has joined #openstack17:22
uvirtbotNew bug: #911876 in quantum "Quantum Admin Guide installation section needs updating" [Undecided,New]
uvirtbotNew bug: #911879 in nova "Update OSAPI and EC2's to properly handle InstanceInvalidState exceptions" [Medium,New]
*** oneiroi has quit IRC17:24
*** Remco__ has quit IRC17:24
*** eolo999 has quit IRC17:25
*** spiffxp has joined #openstack17:29
*** troytoman is now known as troytoman-away17:29
*** tryggvil has joined #openstack17:31
*** eglynn has quit IRC17:32
*** Razique has quit IRC17:36
*** bcwaldon has quit IRC17:36
*** derekh has quit IRC17:37
*** zigo has joined #openstack17:37
*** Guest54318 is now known as tobin17:39
*** dotdevops has joined #openstack17:39
*** eqx311 has quit IRC17:43
*** shaon has quit IRC17:44
*** dolphm has quit IRC17:46
*** aculich has joined #openstack17:47
*** praefect has quit IRC17:47
*** jj0hns0n has quit IRC17:50
*** dragondm has quit IRC17:51
*** ejat has quit IRC17:52
*** nati2 has quit IRC17:52
*** jdurgin has joined #openstack17:52
*** nati2 has joined #openstack17:53
*** oneiroi has joined #openstack17:53
*** oneiroi has quit IRC17:53
*** Ramonster has quit IRC17:54
*** ejat has joined #openstack17:55
uvirtbotNew bug: #911892 in openstack-manuals "identity concept image needs clarification" [Medium,New]
*** jog0 has left #openstack17:57
*** jj0hns0n has joined #openstack18:00
*** troytoman-away is now known as troytoman18:00
*** andrewbogott has joined #openstack18:00
*** redconnection has quit IRC18:01
*** ejat has quit IRC18:01
*** arBmind has quit IRC18:05
*** Remco_ has joined #openstack18:06
*** darraghb has quit IRC18:06
*** rustam has quit IRC18:07
*** rustam has joined #openstack18:07
*** rustam has quit IRC18:12
*** Remco_ has quit IRC18:12
*** mchenetz has joined #openstack18:20
*** lborda has joined #openstack18:25
*** krish has quit IRC18:25
*** bepernoot has joined #openstack18:29
*** martine has joined #openstack18:30
*** zigo has quit IRC18:31
*** bepernoot has quit IRC18:31
*** yshh has quit IRC18:38
*** yshh has joined #openstack18:38
*** yshh has quit IRC18:43
*** sidarta has quit IRC18:45
*** symb0l_ has joined #openstack18:46
*** dpippenger has joined #openstack18:48
*** lesagjea has joined #openstack18:49
*** hugokuo has quit IRC18:51
*** foexle has quit IRC18:56
*** deshantm_ has joined #openstack18:57
*** tmichael has joined #openstack18:59
*** deshantm has quit IRC18:59
*** mattray1 has joined #openstack18:59
*** mattray has quit IRC19:00
*** jj0hns0n has quit IRC19:03
*** troytoman is now known as troytoman-away19:04
*** dotdevops has quit IRC19:06
*** aculich has quit IRC19:06
sniperdI have had a change in review for a few weeks now, should I be +1'ing it myself (I would assume this is frowned upon), or does this thing wait for a bit since fewer people review chef changes?,257119:08
*** donaldngo_hp has quit IRC19:08
*** rocambole has quit IRC19:13
Kiallsniperd: you might want to raise it in the #openstack-dev channel, many of the devs are not in this channel or don't monitor it...19:13
uvirtbotNew bug: #911925 in swift "Bad handling of large "files" in swift-client and greenio" [Undecided,New]
Kiallfrom the topic.. "Openstack Support Channel, Development in #openstack-dev" :)19:14
sniperdKiall: ty19:14
*** amatk has quit IRC19:15
*** aculich has joined #openstack19:16
Kiallsniperd: see? That was faster ;)19:16
sniperdmuch faster :)19:17
*** jj0hns0n has joined #openstack19:20
*** Gordonz has quit IRC19:22
*** yossarian has joined #openstack19:24
yossarianI'm confused about which services need to be installed on varion nodes in a multi node configuration19:25
yossarianis there any service that must be installed everywhere, or any service that should NOT be installed on more than one host?19:25
*** dotdevops has joined #openstack19:25
*** donaldngo_hp has joined #openstack19:26
Kiallyossarian: for nova? or? .. all nova services can be installed on many nodes..19:28
yossarianlike nova-compute, nova-scheduler, nova-network19:28
*** sidarta has joined #openstack19:28
yossariando you have to run nova-compute on every node?19:29
Kialltypically though, you might have 1 (or n) "controller" nodes with the api, scheduler, objectstore, network and multiple compute nodes with just compute19:29
yossarianare you supposed to run nova-scheduler on more than 1 node?19:29
*** dolphm has joined #openstack19:29
Kiallyou can also use the multi-network mode, which means everywhere you run nova-compute, you also run nova-network.. this means network traffic doesn't get NAT'ed via the controller node19:29
yossarianit seems that running "apt-get install nova-compute", and setting up /etc/nova/nova.conf correctly is not enough to get a node to join the cluster19:30
Kiall1 copy of the scheduler will do, but multiple provides redundancy...19:30
KiallI guess your missing something in the config file so :)19:30
KiallHere's an example nova.conf that I use on all nodes, controller or not,
Kiallwatch out for that "--multi_host=True", it multiple nova-network hosts, not hosts in general...19:32
yossarianhow is the compute server supposed to obtain images? does it just download them from the controller node?19:33
KiallThe compute node downloads the images from glance, which could be on the controller node for smaller installs, or its on set of nodes for larger installs19:34
yossarianand the glance client automatically installs along with nova-compute?19:34
KiallYea, "apt-get install nova-compute" should be everything needed on the additional compute nodes (unless you use the multi-node networking, which means you will also need nova-network on them)19:35
Kiall(BTW - If you are planning on using keystone or dashboard, the ubuntu oneiric packages are currently broken ..)19:36
*** bcwaldon has joined #openstack19:36
yossarianwhich network is the multi-node one? I know about flat, flatdhcp, and vlan19:36
KiallAh, it applies to all the network managers...19:36
KiallThe standard way of handing networking is you have 1 nova-network node which all instances proxy/NAT their network traffic via19:37
*** andyandy has quit IRC19:37
*** andyandy_ has quit IRC19:37
Kiallmulti-node removes that SPOF by using nova-network on each nova-compute node19:37
*** primeministerp has quit IRC19:37
yossarianhow do I turn off multi-node then for simplicity?19:37
*** donaldngo_hp has joined #openstack19:38
KiallStart with installing nova-network everywhere you install nova-compute, then enabled that nova.conf flag, and delete+re add all the network settings via nova-manage (or just update the DB manually)19:38
*** dwcramer has joined #openstack19:38
yossarianto turn OFF?19:39
Kiallits off by default19:39
yossarianso the --multi_host=True flag enables it?19:39
yossarianleaving that out turns it off?19:39
Kiallthe --multi_host=True half enables it, it changes the default behaviour of "nova-manage network create" to enable it..19:40
yossarianokay I wonder if that was my problem19:41
uvirtbotNew bug: #911957 in swift "static web can make an extra response to the proxy" [Undecided,Fix committed]
*** tmichael has quit IRC19:41
Kiallyossarian: what was the problem? the new nodes never shows in `nova-manage service list` ?19:41
Kiallnever show in*19:41
*** tmichael has joined #openstack19:41
yossarianno, the node shows up, but attemps to create a new instance get stuck in pending19:41
yossarianand I can't find any log entry indicating an error19:42
SpirilisKiall: does the stable diablo version have multi_host fully supported or is there any patching we need to do to the code?19:42
KiallHave you checked the nova-compute logs on the 2nd node, and the nova-network logs on the first? one of those should tell you why it failed19:42
KiallSpirilis: It works out of the box19:42
Spiriliscool, thx19:42
yossarianit didn't fail, it just gets stuck in limbo19:42
Spirilisgoing to switch to that :)19:42
Spirilismjfork explained what it was19:42
Kiallyossarian: yea, ignore the instance status, thats a bug in older versions of diablo19:43
Kiall(The ubuntu packages are fairly behind on bug fixes last I checked)19:43
KiallThey (the ubuntu packages) are sorting it, but I don't belive they have released updated packages yet.. Hence why i built my own...19:44
Kiallthe ubuntu packagers*19:44
yossarianyeah I had been hoping to use something more stable than the development trunk19:44
Spirilishey anyone see a scenario where libvirtd seems to go out to lunch?19:45
Spirilisnoticed today I couldn't terminate a vm, restarted nova-compute and it was hanging connecting to qemu:/// ... had to restart libvirt-bin and then it worked19:46
*** sidarta has quit IRC19:46
KiallSpirilis: yea, I haven't seen it happen since upgrading to libvirt 0.9.719:46
Spirilisthen a short bit later it caught up with the rabbitmq I guess and the terminate completed19:46
Spirilisahh k19:46
mjforkyossarian: i am late to the party, what are you trying to do you need trunk for?19:46
yossarianKiall was just suggesting that the ubuntu-oneiric packages are bug ridden19:47
Kiallyossarian: the nova packages work (but are missing lots of the diablo updates)..19:48
KiallIts just the KS+Dash packages that are .. non-functional...19:48
*** sidarta has joined #openstack19:48
KiallAnd, Just asked one of the ubuntu packages.. I think updated packages are landing soon19:48
yossariansoon as in tomorrow or soon as in next week?19:49
KiallI've no clue!19:49
mjforkyossarian: supposedly these packages are more stable
KiallThey have been working on it for a while, I would hope soon... ;)19:50
Kiallmjfork: they look pre-diablo, and diablo-final to me?19:50
KiallLots of bugfixes since them19:50
mjforkKiall: there are diablo-final there, which is actually from ~dec19:51
yossarianso this is stable diablo stuff without any essex newness?19:51
Kiallthe diablo-final stuff is all dated Nov..19:51
Kiallyossarian: yes, there is a bugfix only branch for diablo .. neither that repo or ubuntu have updated to it yet though19:52
mjforkyossarian: yes, all stable diablo19:52
KiallI have my packages here.. I keep them fairly up to date with the bugfixes...
mjforkKiall: my undesrstanding, as that post says, is that it is the repo Rackspace Cloud Builders uses19:52
mjforkand that it contains bug fixes, etc. needed to be stable and suitable for deployment19:53
Kiallmjfork: that may be so, but regardless, there have been maybe, 100 or so bugfixes since the date those packages were published19:54
mjforkKiall: oh yea, don't get me wrong, but if you are looking for relatively stable packages for deployment, then that is the place to go19:54
KiallI'm sure it has many fixes over diablo-final (the ubuntu packages do too), but there is likely a lot they dont have :)19:55
Kiall(based on the dates)19:55
KiallAnyway - It doesn't bother me, I've build my own packages that I keep updated with the stable branch fixes19:55
mjforknot sure what yossarian's goal is, but that would be the starting point i would sue19:55
yossarianokay I'll look into using those packages if my current tests fail again19:56
mjforkyou can use kiall's as well19:56
*** deshantm_ has quit IRC19:56
tightworkIs there anyway to move a vm from one tenant to another?19:56
KiallYea, A good few people have had success with mine.. But I'd still prefer Ubuntu got it sorted ;)19:56
yossarianyou know, it appears that installing nova-network on my second node solved the problem19:56
yossarianor else it may have been random luck that solved it19:57
Kialltightwork: not directly, but you can just edit the DB directly19:57
Kialltightwork: not officially, but you can just edit the DB directly*19:57
KiallWatch out for things like SSH keys inside the instance though, they would belong the a user in other tenant (assuming its a different user)19:58
Spirilissimilar kind of question, if a nova-compute node spontaneously crashes, and you got the VM instance store on NFS, do you have to direct-edit the DB to "move" it to a different (surviving) node?19:58
KiallSpirilis: I can't remember who, but someone did script that scenario.. From their description, it was basically update the DB to "move" the node, then issue a reboot19:59
tightworkKiall:  I could import the key, I'm working with a chef knife-openstack and it seems to only want to build a vm in admin:admin and not admin:demo.19:59
SpirilisKiall: yep that's what I figured... sounds easy enough to script I guess19:59
KiallSpirilis: watch out though for the instance running on two nodes at the same time..19:59
Spirilisyeah that would be bad19:59
KiallThat would be disastrous ;)20:00
*** sidarta has quit IRC20:00
*** rmk has quit IRC20:00
KiallEg - if you automate this, then you should as part of that automation physically power down the old compute node.. (Look into IPMI)20:01
Spiriliswe're talking turkey here at work about getting a cloud started for our web developers... getting exciting20:01
*** Kyle__ has joined #openstack20:01
Spirilisah yes, indeed... I think we'll be handling this in a manual scenario so that shouldn't be a problem20:01
Spirilisbut it's one of those questions that come up20:02
uvirtbotNew bug: #911959 in swift "swift-init doesn't support object-expirer" [Undecided,Fix committed]
uvirtbotNew bug: #911960 in swift "auditors" [Undecided,Fix committed]
uvirtbotNew bug: #911962 in swift "saio instructions reference ~/swift/trunk" [Undecided,Fix committed]
Kyle__I'm about to start setting up an openstack cluster, and was wondering first: which OS does most of the community run/develop on?  I was under the impression it was ubuntu, but wanted to be sure before going forward.20:03
*** lesagjea has quit IRC20:03
mjforkKyle__: ubuntu is your best bet20:05
Kyle__mjfork: OK.  Had a rather dissapointing eucalyptus cluster based on ubuntu.  Didn't want to make the same mistake of using a non-standard platform twice!20:06
*** datajerk has quit IRC20:06
*** dolphm has quit IRC20:07
Kyle__mjfork: And the latest-greatest (11.10 server) is the reccomended version?20:10
*** jkyle has joined #openstack20:10
mjforkKyle__: i have seen varying ones, but the newest (11.10) seems to have the most packages20:10
jkyleWhat section of the docs covers how to add users to projects? Im on the nova-manage page, I see how to create a project with an admin user but not how to add a user to a project20:10
mjforkjkyle: keystone?20:11
jkyleum, no. this deployment is older20:12
jkyleI htink I found it, nova-manage project add <projectname> <username>. 2011.320:13
*** Gordonz has joined #openstack20:13
jkyleversion is 2011.320:13
*** Gordonz has quit IRC20:13
mjforkok, diablgo20:13
*** Gordonz has joined #openstack20:14
jkyleso when I add a role within the context of a project the user is added to the project?20:15
*** Ryan_Lane has joined #openstack20:15
uvirtbotNew bug: #911972 in horizon "Subprocess call to curl must DIE!" [Undecided,Confirmed]
*** dolphm has joined #openstack20:16
*** dolphm has quit IRC20:17
mjforkjkyle:  i would guess you need to add them to project first20:17
jkylehow to add to the project is not on the linked page provided, continueing to search other doc content for that20:21
*** n88 has quit IRC20:22
*** mchenetz has quit IRC20:23
mjforkjkyle: based on what is missing (add to project) must just assign a role in the project20:26
*** Gordonz has quit IRC20:26
jkylethere mustn't be an actual add to project. just roles20:27
jkyleI've assigned the role 'sysadmin' to a user within a project, nova-manage role add jkyle sysadmin someproj20:30
jkylehowever, this user cannot run instances. if I: euca-run-instances ami-0000123 I get a 401 error. the nova-api.log says => POST /services/Cloud/ CloudController:RunInstances 40120:32
Spirilisso, to do multi_host, you just run nova-network on all compute nodes?  anything else special you need to consider?20:32
*** datajerk has joined #openstack20:32
Spirilisjust tried firing up a VM on a (multi_host=T) private net with flatdhcp... bombed20:32
Spirilissome weird error in the AMQP code20:33
mjforkSpirilis: is your network in the nova db set to multihsot?20:33
Spirilisboth nova-compute and nova-network registered some kind of AMQP error20:34
*** tdi has quit IRC20:35
KiallSpamapS: has this node ever worked? I'm betting you're missing some nova.conf flags20:35
Kialldo you have the "--rabbit_host=" flag? Most obvious contender20:35
Kiallsorry - not SpamapS .. Spirilis :)20:36
Spirilisah, yes it has worked20:36
Spirilisbeen able to light up VMs on all 3 nova-compute nodes20:36
KiallWell, Its the rabbitmq connection thats bombing.. Thats a place to start looking :)20:36
SpirilisI did just do an apt-get update, upgrade, installed new kernel, rebooted...20:36
KiallIs rabbitmq running?20:37
Spirilisthe beam.smp is running20:37
KiallMaybe it wasnt up when nova-network etc started?20:37
Spirilishmm, that could be possible.  this is on the head/controller node too...20:37
*** katkee has joined #openstack20:38
KiallYea, thats it..20:38
KiallIt happens with release day diablo ;) Pretty sure it was fixed in diablo/stable20:38
Spirilisk gotcha... yeah started another vm, it's running on the 2nd node now and active20:39
Spiriliskvm is running20:39
*** praefect has joined #openstack20:39
Spirilisbtw is it typical for "nova delete <vmname>" not to work if the vm's status is ERROR?20:40
*** dolphm has joined #openstack20:40
*** bepernoot has joined #openstack20:40
Spiriliscool my VM's GW is, not .1 ... multi_host is definitely working :D20:41
Spirilisthanks for your help!20:41
Spirilisthat is some supreme coolness right there20:41
*** n81 has joined #openstack20:42
praefectSpirilis: watchout once I had multi_host enable I ran into troubles with auto assignment of floating IPs, I free my unused floating ips quickly so they are re-used, on a different NIC (multi_hsot) causing ARP poisoning20:43
praefectthere is a fix in place but you need to make arping advertise the new owner of the IP20:43
Kiallpraefect: interesting, I haven't noticed that?20:44
praefect(install arping, set a flag in nova.conf and add arping to sudoers file) I think that was all20:44
Spirilispraefect: ahh I recall reading that somewhere, a link mjfork sent me I think.  they were discussing whether some kind of arping would be necessary20:44
Kiallpraefect: any idea what flag? :)20:44
Spirilisoh so the code to do it is already there...20:44
praefectI'll check the flag, gimme a min20:44
Kiallpraefect: thanks :)20:44
praefect--send_arp_for_ha=true and don't forget to add arping to the sudoers20:45
*** mattray1 is now known as mattray20:45
*** mattray has joined #openstack20:45
Kiallpraefect: yea, I'm updating my packages with "arping" as a dep and arping in the sudoers now :)20:45
praefectI did the fix manually (without realizing it was in the code).. I was all proud and then I see there is a flag for it... hahaha.. openstack is moving fast!20:45
Spirilislooks like my nova_sudoers had arping already, but the /usr/bin/arping ... not /usr/sbin/arping20:45
*** GMI has joined #openstack20:46
praefectSpirilis: I guess it all depends on what version you have, I use RPM packages so it wasn't in mine.. I think I'm somewhere between diablo and essex-1 regarding version..20:46
Kiallspiffxp: /usr/bin/arping is in the iputils-arping package20:47
Spirilisoh weird20:47
Spirilisthey have an 'arping' and 'iputils-arping' deb?20:47
Kiallyea .. there is  "iputils-arping" and "arping" packages20:47
praefectthe thing I'm missing the most is a good integration of keystone, dashboard and nova.. but I'm too lazy to do it myself.. quite sad really..20:47
Spirilissilly ubuntu20:47
Kiallpraefect: I've repackaged working/updated versions everything for ubuntu ;)20:48
praefectKiall: I'm on CentOS (non-negotiable).. it's a business requirement20:48
praefectbut it will come...20:48
*** pguth66 has joined #openstack20:48
Kiallmarkmc and his guys are working on it I believe :)20:49
*** jedi4ever has quit IRC20:49
praefectour goal here is the virtualization of GPUs so I'll probably end up using Xen.. I do already but KVM runs better currently...20:49
praefectif only KVM had GPU virtualization in the roadmap somewhere20:50
*** bepernoot has quit IRC20:50
*** tdi has joined #openstack20:52
*** exprexxo has joined #openstack20:55
*** bepernoot has joined #openstack20:58
*** GMI has quit IRC20:59
*** bcwaldon has quit IRC20:59
*** bepernoot has quit IRC21:00
*** dprince has quit IRC21:00
*** troytoman-away is now known as troytoman21:01
Spirilislol word to the wise, if you delete a flavor and try to re-create it with the same name, end up nuking the DB entry for the old one so it works..... check to make sure the 'instances' table doesn't have any old deleted entries with instance_type_id = <id of the flavor you nuked>21:04
Spiriliselse the horizon GUI bombs trying to show usage stats21:05
*** exprexxo has quit IRC21:07
Spirilisweird, now my VM isn't getting its dhcp request replied to21:07
tightworkKiall: would tenant be known as project_id in the db?21:07
*** dendro-afk is now known as dendrobates21:09
*** jj0hns0n_ has joined #openstack21:10
*** jj0hns0n_ has joined #openstack21:11
*** amatk has joined #openstack21:11
Kialltightwork: yea, some of the terms conflict between keystone and nova :)21:11
KiallSpirilis: more than that..21:12
KiallTry terminating that VM via the CLI ;)21:12
KiallI completely forgot to file a bug against that21:12
Kialland the fact that the flavor deletion via the dash is actually deleting the record, rather than marking it deleted as it should be21:12
*** jj0hns0n has quit IRC21:13
*** jj0hns0n_ is now known as jj0hns0n21:13
Spirilishuh, had to nuke dnsmasq and restart nova-network, dnsmasq was still using --listen-address= (the old default GW pre-multi_host) rather than .421:15
Spirilisnow it's listening on the .421:15
*** mgoldmann has quit IRC21:25
*** drewlander has quit IRC21:30
*** Tobbe has joined #openstack21:32
*** nati2 has quit IRC21:37
*** mattray has quit IRC21:38
*** mchenetz has joined #openstack21:39
*** Arminder has quit IRC21:39
*** Arminder has joined #openstack21:39
adam_ghmm. if requests for an image are hitting glance without tenant/project id tacked on to url (ie, v1/images/ref as opposed to v1/admin/ref) glance correctly returns a 401 and nova-api returns a 500. does this suggest an error in nova-api or keystone? glance's endpoint appears to be configured with the proper /%tenant_id% suffix21:42
Kiallthe glance endpoint doesnt have %tenant_id%..21:44
KiallIt should be "http://XXX.XXX.XXX.XXX:9292/v1"21:45
KiallLots of misinformation out there that it should be "http://XXX.XXX.XXX.XXX:9292/v1.1/%tenant_id%"21:45
*** vendemiat has joined #openstack21:46
adam_gKiall: hmm21:47
Kiall should be enough proof ;)21:48
adam_gKiall: perhaps the %tenant_id% was a red herring on my part, i originally hit the same issue without21:48
KiallAlso, Watch out for "v1" vs "v1.1"21:48
KiallAnyway, paste the error logs and maybe someone can help :)21:49
adam_gerror logs dont provide much, in this case. the best i can tell is that the tenant id is missing from requests to the glance server in setups where this si failing, but present on devstack (working). trying to figure out whats misconfigured, nova-api, nova-compute21:51
Madkiss2012-01-04 21:51:00,049 CRITICAL nova [-] No module named metadatarequesthandler21:51
Madkisswhen starting nova-api21:51
mjforkadam_g: what is output of keystone-manage endpointTemplates list21:51
adam_gmjfork: one sec, ive been trying different formats for the glance endpoint.21:52
mjforkdiablo or trunK?21:52
*** tmichael has quit IRC21:52
*** joesavak has quit IRC21:53
adam_gmjfork: currently failing with this:  previosly failing the same way w/o %tenant_id% tacked on21:53
*** tmichael has joined #openstack21:53
adam_gmjfork: e2 milestone21:53
*** tmichael has quit IRC21:53
*** bengrue has quit IRC21:54
Madkisswhen trying to start nova-api, I see this:
Madkissthat's my api-paste.ini:
MadkissThis is happening on Ubuntu Precise development branch with e2 packages21:55
Madkisse2 milestone, i mean21:55
Madkisswhen trying to get nova to work with keystone21:56
Madkisscould anyone lend me a helping hand please? :)21:56
mjforkadam_g: looks ok, i haven't do anything wtih essex so can't be of much help.21:56
*** Gordonz has joined #openstack22:00
*** lts has quit IRC22:02
adam_gMadkiss: where id you base the api-paste.ini off?22:03
Madkisswell, it's a mixture of and
*** Hakon|mbp has quit IRC22:05
*** andrewbogott_ has joined #openstack22:05
Madkissalthough while using the first mainly, I now at least made nova-api start again22:05
MadkissI have no idea whether my file is correct now, though22:06
*** martine has quit IRC22:07
adam_gMadkiss: this is what im using on precise, its working quite well with the exception of the glance stuff above. im pretty sure that is unrelated, though22:07
*** clauden__ has quit IRC22:07
*** clauden has joined #openstack22:07
*** miclorb has joined #openstack22:08
adam_gMadkiss:   update the bottom keystone section to your env22:08
*** andrewbogott has quit IRC22:08
*** andrewbogott_ is now known as andrewbogott22:08
Madkisswoohoo, thanks a lot22:08
*** miclorb has joined #openstack22:08
adam_gMadkiss: also22:08
adam_gMadkiss: apt-get install openstackx and add the following to your nova.conf22:08
*** llang629 has joined #openstack22:09
adam_gMadkiss: apt-get install python-keystone22:09
*** llang629 has left #openstack22:09
adam_gpackaging will hopefully be updated in the future to take care of that, or at least some of it22:10
*** MarkAtwood has quit IRC22:11
*** MarkAtwood has joined #openstack22:12
Madkisslet's see22:12
*** dspano has quit IRC22:14
*** dubsquared has quit IRC22:15
Madkissnow i'm down again to this:
Madkisskeystone-manage endpointTemplates add RegionOne keystone http://%HOST_IP%:5000/v2.0 http://%HOST_IP%:35357/v2.0 http://%HOST_IP%:5000/v2.0 1 122:19
*** Shentonfreude has quit IRC22:19
Madkissis that one correct?22:19
adam_gMadkiss: looks good22:20
Madkissthen i'm out of ideas22:20
adam_gMadkiss: Madkiss what does keystone_ec2_url point to in nova.conf?22:20
uvirtbotNew bug: #912015 in horizon "Dashboard uses curl to check that endpoints are up" [Undecided,Invalid]
Madkissadam_g: just like you said: --keystone_ec2_url=
*** paulmillar has joined #openstack22:21
*** llang629 has joined #openstack22:21
*** llang629 has left #openstack22:21
*** clopez has quit IRC22:22
adam_gMadkiss: strange, wondering why that trace shows 'POST /v2.0/tokens HTTP/1.1\r\nHost:'22:22
*** dotdevops1 has joined #openstack22:22
*** dotdevops has quit IRC22:22
*** MarkAtwood has quit IRC22:22
Madkissadam_g: I have set NOVA_URL=""22:22
*** MarkAtwood has joined #openstack22:22
Madkissand declare -x OS_AUTH_URL=""22:23
adam_gMadkiss: oh, hmm. i haven't really played much with the nova client or user facing OSAPI stuff.22:25
*** troytoman is now known as troytoman-away22:26
*** MarkAtwood has quit IRC22:26
Madkissii  python-novaclient               2.6.7~bzr128-0ubuntu1                         client library for OpenStack Compute API22:29
Madkissis that the most recent version?22:30
*** imsplitbit has quit IRC22:31
*** katkee has quit IRC22:31
Madkissinterestingly enough, btw. "curl -d '{"tenantName": "admin", "passwordCredentials":{"username": "admin", "password": "hastexo"}}' -H "Content-type: application/json"" works like a charm22:31
*** GheRivero_ has quit IRC22:35
*** mchenetz has quit IRC22:35
judd7anyone know how to turn OFF swift container sync for a single container?  Should I just remove the headers?  How??22:35
*** mchenetz has joined #openstack22:36
*** yossarian has quit IRC22:37
*** willaerk has joined #openstack22:37
*** mchenetz has quit IRC22:38
*** spackest has joined #openstack22:38
spackestso, I'm trying to get openstack working with ubuntu 11.10 and think (hope) I am getting close, but I get this error swift-container-sync does not exist.  It looks like it is a known bug that has been fixed22:40
spackestguess I'm just not sure how to get swift, version 1.4.4~20111108.1612-0ubuntu122:40
*** n81 has quit IRC22:43
spackestfrom here?
*** n81 has joined #openstack22:47
*** dolphm has quit IRC22:47
*** markvoelker has quit IRC22:47
*** spiffxp has quit IRC22:49
ppradhandouble checking.. do the iimages at should work with openstack?22:50
*** n81 has quit IRC22:50
*** galthaus has joined #openstack22:52
*** bsza has quit IRC22:56
*** jj0hns0n has quit IRC22:57
Kiallppradhan: yes, kinda ;)22:57
*** jj0hns0n has joined #openstack22:57
Kiallppradhan: for maverick on, use the .img rather than .tar.gz versions22:57
*** jj0hns0n has quit IRC22:57
*** rnorwood has quit IRC22:58
Kiallfor lucid, use the .img inside the .tar.gz and the kernel and ramdisk from
Kiall(for an example how to do that .. have a look at the glance-upload-*.sh scripts I have here: )22:59
ppradhani've download the oneric.uploaded using cloud publish.. while running I can only see the bios23:02
KiallThats likely unrelated to the download, they will all appear to work fine, until you try and update the kernel inside an instance...23:02
Kiallwhat does the console show? (rather than vnc)23:03
Kyle__A layout question: for two interface systems, one interface would be hooked up to a public switch, and all VMs would appear as part of that network, and the other interace would be hooked up to a private switch.  Is that correct?23:03
*** Gordonz has quit IRC23:04
KiallKyle__: yes, bar all instances would be on the private switch... and floating IPs+NAT on the public switch23:04
Madkissis there something i need to do with nova for the "initial auth"?23:04
*** rnorwood has joined #openstack23:04
*** RicardoSSP has joined #openstack23:04
*** micadeyeye_ has quit IRC23:04
*** micadeyeye has quit IRC23:04
Kyle__Kiall: OK, thank you.23:04
*** cdub has quit IRC23:05
*** kbringard has quit IRC23:05
*** RicardoSSP has quit IRC23:05
ppradhanKiall:  euca-describe-instances says its running23:05
Kialleuca-describe-instances says that nova ran the KVM command, and that command did not fail.23:06
KiallNothing to do with if the instance booted or not :)23:06
ppradhanso where should i check?23:07
Kiallwhat does euca-get-console-output show?23:07
*** cdub has joined #openstack23:07
*** jfluhmann has joined #openstack23:10
notmynamejudd7: remove headers by setting them to an empty string23:11
ppradhanKiall: says "[    2.547418] VFS: Cannot open root device "vda" or unknown-block(0,0)23:11
KiallSo VM cant find its hard drive23:12
KiallIt could be related to the uploaded image, or a few other things...23:13
*** nati2 has joined #openstack23:13
ppradhanKiall: Yes but I can see kernel and the main image from euca-describe-images23:13
KiallSure, but was it corrupted in transit? while saving? which transferring from glance to nova? while being resized for the instance? etc etc..23:14
KiallIf you try start another, does it succeed? (ie was it a once off muck up?)23:14
ppradhanis there anyway to check if the downloaded tar.gz is not corrputed?23:16
Kiallwell the md5, but i doubt it would have extracted if it was corrupted during the download from uec-images...23:17
KiallYou're best bet is to try another, if it fails.. dont do anything! Just go look at the logs23:17
ppradhanok.. i will get another one and try23:18
KiallRight now, the logs will be filled with all sorts of extra stuff below the real error, making it harder to find unless you know what to look for23:18
*** mdomsch has quit IRC23:18
*** PotHix has quit IRC23:19
*** jj0hns0n has joined #openstack23:19
*** nati2_ has joined #openstack23:19
*** ppradhan has quit IRC23:21
uvirtbotNew bug: #912033 in nova "GET /servers/detail can return 404" [Undecided,New]
*** nati2 has quit IRC23:22
*** ppradhan has joined #openstack23:23
ppradhanKiall: irc client crashed .. can you paste the 2 lines that contains the url to kernal and ramdisk?23:24
Kiall<Kiall> ppradhan: for maverick on, use the .img rather than .tar.gz versions <Kiall> for lucid, use the .img inside the .tar.gz and the kernel and ramdisk from
Kiall<Kiall> (for an example how to do that .. have a look at the glance-upload-*.sh scripts I have here: )23:24
*** jfluhmann has quit IRC23:24
ppradhanKiall:  Thanks !23:26
*** ironcamel2 has quit IRC23:27
*** dragondm has joined #openstack23:29
*** gyee has joined #openstack23:29
*** zzed has quit IRC23:31
*** ironcamel is now known as ironcamel223:31
*** ironcamel2 is now known as ironcamel23:32
*** spiffxp has joined #openstack23:32
*** dwcramer has quit IRC23:32
*** ironcamel2 has joined #openstack23:35
judd7notmyname: curl -i -XPOST -H 'x-auth-token: AUTH_tk39cf30deaeb04ff4a83d02544bc8edb5' -H 'X-Container-Sync-To:' -H 'X-Container-Sync-Key:'
judd7doesn't actually make the change to ''23:41
*** rnirmal has quit IRC23:41
notmynamethat should delete the header23:41
judd7That's what I tought. :(23:41
judd7root@proxy02-c01:/etc/swift# curl -ik -H 'x-auth-token: AUTH_tk39cf30deaeb04ff4a83d02544bc8edb5'  -XHEAD<hash>/aaa23:42
notmynamehmm...I'd have to look at the docs (I'm getting confused on the different ways metadata works on container vs objects)23:42
judd7X-Container-Sync-Key: abcde23:42
*** clauden has quit IRC23:42
*** clauden__ has joined #openstack23:42
*** jkyle has quit IRC23:43
*** jj0hns0n has quit IRC23:43
*** gyee has quit IRC23:45
judd7notmyname: if I try to set it to blank with the swift client, validate_sync_to errs.23:45
judd7I'm out.  please respond on list if you can think of anything.23:48
*** judd7 has quit IRC23:48
ppradhanKiall: the new downloaded images is working !23:50
ppradhanKiall: whats the default password they put it?23:50
KiallCool, I guess "corrupted-in-transit " was the issue..23:50
Kiallthere is none, can you imaging how many EC2 instances would be owned if they had a default? ;)23:51
KiallWhen you start the instance, you specify which SSH key to use with the -k option23:51
*** vendemiat has quit IRC23:52
*** rnorwood1 has joined #openstack23:52
*** rnorwood has quit IRC23:53
notmynamejudd7: (although he has left) curl won't send blank headers :-(. But that's the way to delete them23:54
Kiallnotmyname: i remember someone posted some PHP on the mailing list that used the PHP curl ext to delete headers..23:55
KiallI assume the same thing could be done from the CLI curl too23:56
notmynameKiall: unfortunately, no23:58
notmynameat least no documented way, and there are patches changing this behavior on the curl mailing list23:59
notmynameso I think the only way to do it is by crafting the http request yourself23:59

Generated by 2.14.0 by Marius Gedminas - find it at!