Wednesday, 2011-03-23

*** chuck_ has joined #openstack00:03
*** jeffjapan has joined #openstack00:05
*** mahadev has quit IRC00:08
*** chuck_ is now known as zul00:08
diegoparrillavinay_: sounds like your image is not booting :-(00:11
diegoparrillavinay_: are you using ami-tty?00:11
vinay_yes ami-tty00:13
*** gregp76 has quit IRC00:14
vinay_if image is not booting then how come euca-describe-instances shows one as running00:14
vinay_RESERVATION     r-d0a6d9lz      admin   default INSTANCE        i-00000001      ami-tty        running mykey (admin, vin-ubuntu)       0               m1.tiny 2011-03-22 14:50:33.062454       nova00:14
uvirtbotNew bug: #740605 in nova "OpenStack API v1.0 sends computeFault, not cloudServersFault" [Undecided,New]
diegoparrillavinay_: afaik the change to state 'running' happens when the virtual machine tries to boot the image. May be the image is corrupt or incorrect...00:16
notmynameOneiropolo: the simple answer is "don't do that". choose a big enough value to start so you won't have to change it later. unfortunately, swift doesn't yet have any ability to expand a cluster. the only current solution (for no interruption) is "stand up a second cluster and migrate the data". and that's hard00:17
vinay_btw I just used the default image which was registered already00:18
vinay_I installed  nova using the script00:18
*** dendro-afk is now known as dendrobates00:20
vinay_thanks for your help @diegoparrilla... I will look into this deeper... if you have any clues let me know :)00:20
*** gregp76 has joined #openstack00:21
*** lionel_ has quit IRC00:22
*** lionel_ has joined #openstack00:22
Oneiropolonotmyname: thanks for answer00:27
*** diegoparrilla has quit IRC00:27
Oneiropolonotmyname: but at deployment, if i don't have many servers or disks00:28
Oneiropoloand it looks like going to be big enough to extends 10PB or 100PB00:28
notmynamewhen you create your cluster, choose the ring power (which ultimately decides the max size of the cluster) based on what you could have, not what you do have00:29
notmynamefor example, you may be constrained by budget, DC space, power, network ports, etc00:29
*** slyphon has joined #openstack00:29
Oneiropolothat ring power doesn't affect the performance of object storage?00:29
notmynamejust the performance of the ring operations (creation, rebalancing, etc) which all happen out-of-band to your client requests00:30
Oneiropoloas like, at this time i just have 100 disks of 1TB00:30
Oneiropolobut i decided my ring power as 100PB or 200PB00:31
Oneiropolois it possible?00:31
notmynamesure. other things get complicated (like your networking or your DC provisioning), but the ring and swift will be fine00:31
Oneiropoloand the performance isn't going to change.00:32
notmynamekeep in mind though that the bigger the ring power, the more partitions, which do eat up a small, but noticeable, percentage of your disk that is unbilled. it takes space for the filesystem to keep track of all those inodes00:32
Oneiropolookay. i get it.00:33
Oneiropoloand i have one more thing .00:33
notmynamethe ring power, IIRC, is only used in swift as how many bit to use for string slicing, so doing 20 or 30 shouldn't cause much difference in any performance00:34
Oneiropolowhen i need to add more storage nodes to the cluster .00:34
Oneiropolothanks at all .00:35
Oneiropolowhen i add more storage nodes to the swift cluster00:35
*** jfluhmann_ has joined #openstack00:35
Oneiropoloi need to build new ring from the old ring i used00:36
Oneiropoloduring the ring building, is the service going to stop?00:36
Oneiropoloor swift can serve the request during the rebalance and ring building00:37
Oneiropoloand are those behaviour affect swift performance?00:38
notmynameyou rebuild the ring separate from the cluster then deploy it to the cluster. when it's deployed it will begin to rebalance the data, but it will continue to server all requests00:44
*** gregp76 has quit IRC00:44
Oneiropolowhat it mean that deploy the ring to the cluster?00:45
*** sebastianstadil has joined #openstack00:47
*** adjohn has joined #openstack00:50
*** bcwaldon has quit IRC00:56
*** dendrobates is now known as dendro-afk00:57
*** Ryan_Lane has quit IRC01:01
notmynameOneiropolo: actually copy it to the storage nodes01:08
notmynamegenerally, one should generate the ring(s) and use whatever deployment tools you like to copy it to the storage nodes01:09
Oneiropolonotmyname : what happen? if i didn't copy that ring file to the storage nodes?01:17
*** zenmatt has joined #openstack01:18
*** localhost has quit IRC01:19
notmynameOneiropolo: the same thing that happens if you don't copy an html file to your web server. it doesn't get used01:20
*** localhost has joined #openstack01:20
*** shentonfreude has quit IRC01:23
*** User470 has quit IRC01:25
*** jtran has quit IRC01:26
*** dendro-afk is now known as dendrobates01:30
*** joearnold has quit IRC01:32
*** joearnold has joined #openstack01:35
*** vernhart has joined #openstack01:37
*** odyi has quit IRC01:38
*** odyi has joined #openstack01:38
*** odyi has joined #openstack01:38
*** MotoMilind has quit IRC01:38
*** joearnold has quit IRC01:44
*** zenmatt has quit IRC01:44
*** bcwaldon has joined #openstack01:46
*** MarkAtwood has joined #openstack02:02
*** zenmatt has joined #openstack02:04
*** winston-d has quit IRC02:04
*** MarkAtwood has left #openstack02:04
uvirtbotNew bug: #740647 in swift "content-encoding can't be set or deleted via post" [Undecided,Confirmed]
*** rds__ has quit IRC02:13
*** zenmatt has quit IRC02:15
*** azneita has joined #openstack02:18
*** bcwaldon has quit IRC02:21
dendrobatesugh, reviewing and testing  ntt-pf-lab/nova/flatmanager-ipv6  is taking much longer than I expected02:28
*** bcwaldon has joined #openstack02:29
*** zenmatt has joined #openstack02:37
*** santhosh has joined #openstack02:53
*** bcwaldon has quit IRC03:01
uvirtbotNew bug: #740668 in nova "No XML schema for OpenStack API v1.1?" [Undecided,New]
*** mray has quit IRC03:03
jarrodIs there an api call that allows me to initialize a vdi with a remote image (img) ?03:19
*** gondoi has quit IRC03:35
*** gregp76 has joined #openstack03:36
*** gregp76 has quit IRC03:37
*** joearnold has joined #openstack03:37
*** gondoi has joined #openstack03:39
*** hggdh has quit IRC03:47
*** hggdh has joined #openstack03:48
*** hggdh has joined #openstack03:48
*** dendrobates is now known as dendro-afk04:12
*** guynaor has joined #openstack04:12
openstackjenkinsProject nova build #682: SUCCESS in 2 min 4 sec:
openstackjenkinsTarmac: Fix the describe_vpns admin api call.04:13
openstackjenkinsFirstly, use the correct mechanism for mapping ec2 ids from instance ids.04:13
openstackjenkinsSecondly, if a vpn doesn't have an ip/port assignment from the project, don't try utils.vpn_ping, as it will raise an exception.04:13
*** guynaor has left #openstack04:14
*** gregp76 has joined #openstack04:19
*** santhosh has quit IRC04:20
*** dendro-afk is now known as dendrobates04:29
*** kashyap has joined #openstack04:37
*** koji-iida has joined #openstack04:41
dendrobatestime to review and test 3 merge requests == 4 hours, Number of new merge requests during that same period == 604:44
dendrobatesslowly moving backwards04:44
dendrobatesbut hey I'm in a hotel with no distractions, I'll just keep going04:44
*** gregp76 has quit IRC04:51
*** joearnold has quit IRC04:58
koji-iidadendrobates: Hi, I just fixed our ipv6 branch. Could you check again?
*** ovidwu has quit IRC04:58
koji-iidaI always tested ./ -N. But now I hope this branch passes with -V.04:59
*** vernhart has quit IRC04:59
*** ovidwu has joined #openstack04:59
dendrobateskoji-iida: I am testing it now05:02
openstackjenkinsProject nova build #683: SUCCESS in 2 min 6 sec:
openstackjenkinsTarmac: Test the login behavior of the OpenStack API.  Uncovered bug73286605:03
*** hadrian has quit IRC05:10
koji-iidadendrobates: Thank you for your approval.05:10
dendrobateskoji-iida: you're welcome.  It was an important feature.  Thanks for doing the work.05:14
*** vernhart has joined #openstack05:23
openstackjenkinsProject nova build #684: SUCCESS in 2 min 9 sec:
openstackjenkins* Tarmac: Enable flat manager support for ipv6.05:32
openstackjenkins* Tarmac: Adding a talk bubble to the site that points readers to the 2011.1 site and the site - similar to the site. I believe it helps people see more sites are available, plus they can get to the Bexar site if they want to. Going forward it'll be nice to use this talk bubble to point people to the trunk site from released sites.05:32
openstackjenkins* Tarmac: Correctly imports greenthread in  It is used by live_migrate().05:32
*** zigo has joined #openstack05:32
*** zigo is now known as zigo-_-05:33
*** jfluhmann_ has quit IRC05:40
*** koji-iida has quit IRC05:41
*** MarkAtwood has joined #openstack05:48
*** Ted____ has joined #openstack05:49
*** kashyap has quit IRC05:49
*** jfluhmann_ has joined #openstack05:56
uvirtbotNew bug: #740742 in nova "EC2 Admin api returns misformatted instance types" [Undecided,New]
*** kashyap has joined #openstack06:07
*** maplebed has quit IRC06:16
*** ivan has quit IRC06:16
dendrobatesthat's all I got for tonight06:23
*** HugoKuo_ has quit IRC06:24
dendrobatessomeone else's review day now.  Tag!06:24
westmaas1hopefully they start right away06:24
dendrobatesRick Harris is next.  He is a late sleeper, so probably not right away.06:27
westmaas1haha, I was just kidding06:27
westmaas1thanks for working so late, this was your first cisco day right?06:27
*** ivan has joined #openstack06:27
openstackjenkinsProject nova build #685: SUCCESS in 2 min 7 sec:
openstackjenkinsTarmac: Fix for #740742 - format describe_instance_output correctly to prevent errors in dashboard.06:28
openstackjenkinsWithout this change, it is not possible to properly select instance types when launching instances with the OS dashboard.06:28
*** Ted____ has quit IRC06:29
*** koolhead17|afk is now known as koolhead1706:31
*** HugoKuo has joined #openstack06:33
koolhead17hi all06:33
*** Cybo has joined #openstack06:34
*** Cybo has left #openstack06:34
*** slyphon has quit IRC06:34
dendrobateswestmaas: my 2nd day06:35
westmaas1cool, hope its going well!06:35
dendrobatesso far, so good06:37
*** zenmatt has quit IRC06:43
*** guigui has joined #openstack06:44
alekibangoyou should do those days in pairs to make it easier06:44
*** miclorb_ has quit IRC06:58
nerenshi guys, i have a problem where nova-compute and nova-network respawn every couple seconds generating loads of logs.. I'm seeing a "programming error".. you can see details here:
nerensany help/pointers would be greatly appreciated07:06
nerenslooks like "nova-manage db sync" sorted it.. no more errors?07:12
*** allsystemsarego has joined #openstack07:15
*** allsystemsarego has joined #openstack07:15
*** localhost has quit IRC07:17
*** aliguori has quit IRC07:19
*** nidO has quit IRC07:35
*** fayce_ has joined #openstack07:39
*** fayce__ has joined #openstack07:39
*** fayce_ has quit IRC07:39
*** mgoldmann has joined #openstack07:47
*** ramkrsna has joined #openstack07:50
*** ramkrsna has joined #openstack07:50
*** romain_lenglet has joined #openstack07:53
*** lionel_ has quit IRC07:53
*** romain_lenglet has left #openstack07:53
*** lionel_ has joined #openstack07:54
*** miclorb has joined #openstack07:56
*** fayce__ has quit IRC08:01
*** adjohn has quit IRC08:09
*** reldan has joined #openstack08:17
Oneiropolonow i'm reading the swift document.08:22
Oneiropoloi wonder the meaning of array(i) in ring description08:23
Oneiropoloarray(i) of devices ids08:23
Oneiropolowhat is the meaning of array(i)?08:23
*** skiold has joined #openstack08:23
*** aliguori has joined #openstack08:23
*** nidO has joined #openstack08:28
*** taihen has quit IRC08:31
*** fayce has joined #openstack08:31
*** jeffjapan has quit IRC08:33
*** fayce has quit IRC08:34
*** fabiand_ has joined #openstack08:41
*** Nacx has joined #openstack08:42
*** MarcMorata has joined #openstack08:49
*** ptremblett has quit IRC08:56
uvirtbotNew bug: #740810 in nova "nova-manage network creation doesn't respect CIDR" [Undecided,New]
*** viirya has quit IRC08:57
*** viirya has joined #openstack08:57
*** fayce has joined #openstack08:57
*** Oneiropolo has quit IRC09:08
*** fayce has quit IRC09:09
*** fayce has joined #openstack09:09
*** doude has joined #openstack09:25
*** herki has joined #openstack09:31
*** ramkrsna has quit IRC09:33
*** kal24 has joined #openstack09:33
*** bkkrw has joined #openstack09:35
*** ramkrsna has joined #openstack09:45
*** BigValen has joined #openstack09:50
BigValenHey, if anyone is around, I'm doing a little research on people working with cloud-computing type systems.09:50
BigValenIf you could have a look at my two simple questions on and mail (or msg) me an answer, that would be awesome. I won't be attributing anything to people/organisations, just looking for common trends & points.09:51
*** adiantum has quit IRC09:54
*** adjohn has joined #openstack09:54
*** ptremblett has joined #openstack10:01
*** sebastianstadil has quit IRC10:12
*** metoikos has quit IRC10:13
*** reldan has quit IRC10:14
*** metoikos has joined #openstack10:15
*** doude has quit IRC10:15
*** fayce has quit IRC10:16
* kal24 is extraordinary happy, after several days, and many unsuccessful installs, finally found the problem, and i have got my instances launching :)10:16
*** fabiand_ has quit IRC10:19
kal24i have not found anything written about what caused me problem, where should i write it, to share with people new to nova, not to spend time with such a stupid mistake10:21
*** doude has joined #openstack10:24
adjohnkal24: was it a bug, or just from having poor install instructions?10:27
*** miclorb has quit IRC10:28
*** mrchrisadams has joined #openstack10:28
*** fayce has joined #openstack10:30
*** kashyap has quit IRC10:35
*** drico has quit IRC10:36
*** drico has joined #openstack10:36
*** ptremblett has quit IRC10:41
*** ptremblett has joined #openstack10:43
*** drico has quit IRC10:44
*** drico has joined #openstack10:44
*** BigValen has quit IRC10:52
*** BigValen has joined #openstack10:54
*** azneita has quit IRC10:56
*** drico has quit IRC10:58
*** drico has joined #openstack10:59
*** reldan has joined #openstack11:03
*** bkkrw has quit IRC11:05
*** miclorb has joined #openstack11:08
*** adjohn has quit IRC11:11
*** romain_lenglet has joined #openstack11:12
*** nerens has quit IRC11:13
*** romain_lenglet has quit IRC11:22
HugoKuoin this doc    it talks about Cloud Controller.....11:23
HugoKuowhat's Cloud Controller's definition?11:24
HugoKuoCloud Contoller= Nova-API + MySQL ?11:25
BigValenCloud Controller is the master state. What jobs should be running where.11:25
*** fayce has quit IRC11:25
*** ctennis has quit IRC11:28
openstackjenkinsProject nova build #686: SUCCESS in 2 min 7 sec:
openstackjenkinsTarmac: We update update_ra method to synchronize, in order to prevent crash when we request multiple instance at once.11:28
*** reldan has quit IRC11:36
*** miclorb has quit IRC11:39
*** adiantum has joined #openstack11:40
*** joloughlin has joined #openstack11:41
joloughlinhaving an issue with routing from my instances11:44
joloughlinrunning flatdhcp instances get a private ip11:45
joloughlini assign them a public one11:45
joloughlineuca-authorise used to allow ssh11:45
joloughlinthis works fine11:45
joloughlinbut when i ssh into the instance and then want to connect out11:45
joloughlini am having problems11:45
joloughlinfrom iptables -L -n -t nat the snat and dnat rules are there..11:46
*** ctennis has joined #openstack11:48
HugoKuoBigValen : thanks , So where is Cloud controller live , when I seperate nova-network from all-in-one box ?11:51
BigValenI don't know, sorry. I'm very very new to openstack.11:53
joloughlinmy nova-network appears to be dropping new outbound connection from the instances11:53
joloughlininbound connections are fine11:53
joloughlinany ideas ?11:53
*** santhosh has joined #openstack11:54
HugoKuoI think Cloud controller is a kind of abstract to descript entire NOVA system11:55
*** matclayton has joined #openstack11:55
HugoKuojoloughlin : what's your physical architecture ?11:56
HugoKuoall-in-one box ?11:56
*** sateesh has joined #openstack11:58
*** omidhdl has joined #openstack11:58
joloughlinhugokou: one node with api, network and compute12:02
joloughlinone with just compute12:03
joloughlinso inbound traffic to instances is all ok12:03
joloughlinits juts connections originating from the instances cant get out12:03
HugoKuoI have a same archi with you12:04
HugoKuoas your approach12:04
joloughlinim using flatdhcp12:04
HugoKuosame with me12:04
HugoKuolet me think12:04
joloughlinthe instances are given a private ip12:05
joloughlinok, thanks12:05
HugoKuodid you try on two instances those are in different compute nodes?12:05
HugoKuofor now , your instance must runrobin on two compute nodes,12:06
joloughlinyeah, i started 2 instances12:06
joloughlinone is running on each node12:06
HugoKuoI remember I ever faced same problem before12:06
joloughlinstill the same12:06
joloughlinfrom looking at the iptables rules12:07
joloughlinthe snat and dnat rules are there12:07
HugoKuono idea..... I have to see your nova.conf12:08
HugoKuobut I got leave office now12:08
joloughlinthanks for your help12:09
*** dovetaildan has quit IRC12:15
uvirtbotNew bug: #740929 in nova "Spawn error if cidr_v6 is null for network" [Undecided,New]
*** omidhdl1 has joined #openstack12:17
*** dovetaildan has joined #openstack12:17
*** reldan has joined #openstack12:18
*** hggdh has quit IRC12:18
*** adjohn has joined #openstack12:18
*** hggdh has joined #openstack12:19
*** omidhdl has quit IRC12:19
*** jfluhmann_ has quit IRC12:20
*** gaveen has joined #openstack12:21
*** hadrian has joined #openstack12:23
kal24joloughlin, did u try killall dnsmasq and service nova-network restart?12:26
*** santhosh has quit IRC12:27
joloughlinkal24: i needed to add routing_source_ip to nova.conf12:27
joloughlineverything works fine now :)12:28
kal24i see :)12:28
sandywalshphew ... all reviews for me now until they pull me away from the keyboard!12:31
*** Nick_ has joined #openstack12:32
*** Nick_ is now known as Guest951612:32
uvirtbotNew bug: #740935 in nova "wrong local variable name in VMOPS._shutdown causes termination to fail" [Undecided,In progress]
*** westmaas1 has quit IRC12:33
*** sateesh has quit IRC12:35
*** kal24 has quit IRC12:40
*** adjohn has quit IRC12:40
uvirtbotNew bug: #740947 in nova "_write_partition: parted should be executed with sudo" [Undecided,In progress]
*** adjohn has joined #openstack12:53
*** fabiand_ has joined #openstack12:54
*** nerens has joined #openstack12:55
*** nerens has quit IRC12:57
*** nerens has joined #openstack12:57
*** bcwaldon has joined #openstack12:58
*** doude has quit IRC13:01
*** adjohn has quit IRC13:03
*** mdomsch has joined #openstack13:03
*** guigui3 has joined #openstack13:04
*** doude has joined #openstack13:05
*** guigui3 has quit IRC13:06
*** guigui has quit IRC13:07
*** kashyap has joined #openstack13:10
*** omidhdl1 has quit IRC13:14
*** rds__ has joined #openstack13:14
*** MontgoDB has joined #openstack13:16
*** alekibango has quit IRC13:16
*** mray has joined #openstack13:17
*** dprince has joined #openstack13:23
*** jero_ has joined #openstack13:24
*** drico has quit IRC13:26
*** zenmatt has joined #openstack13:33
*** Shentonfreude has joined #openstack13:37
*** slyphon has joined #openstack13:40
*** RobertLaptop has quit IRC13:40
*** at141am has joined #openstack13:43
*** santhosh has joined #openstack13:47
*** pygi has quit IRC13:51
*** Guest9516 has quit IRC13:52
*** m_3 has quit IRC13:58
jaypipessirp_: phew.
*** omidhdl has joined #openstack13:59
*** romain_lenglet has joined #openstack14:02
*** bkkrw has joined #openstack14:03
*** romain_lenglet has left #openstack14:03
*** omidhdl has quit IRC14:03
*** m_3 has joined #openstack14:03
*** omidhdl has joined #openstack14:03
*** BigValen has quit IRC14:04
*** BigValen has joined #openstack14:05
*** Guest9516 has joined #openstack14:08
jaypipesmtaylor: ping me when you're around.14:09
*** dendrobates is now known as dendro-afk14:09
*** doude has quit IRC14:14
*** alekibango has joined #openstack14:15
*** stewart has quit IRC14:16
* soren heads out to pick up his daughter14:21
*** stewart has joined #openstack14:25
*** dendro-afk is now known as dendrobates14:27
*** crescendo has joined #openstack14:35
sandywalshtr3buchet, nice to see people are echoing your multi-nic work to libvirt :)14:35
sandywalshthe power of open source :)14:35
*** RobertLaptop has joined #openstack14:35
*** adjohn has joined #openstack14:39
*** adjohn has quit IRC14:41
*** pygi has joined #openstack14:41
westmaasttx: nice :)14:43
jaypipesttx: very nice :)14:43
ttxshould get refreshed every 30min14:44
westmaassandywalsh: ^^14:44
ttxvery alpha quality, but I thought it makes a lot of sense to push right now14:44
ttxjaypipes: I still need to make it releasecycle-aware... like switch to prioritize bugfixes, targeted bugs etc.14:45
westmaasttx: what does italics mean?14:45
dendrobatesttx: awesome! I will keep reviewing as time allows today.  It is a big mountain to climb14:45
ttxprereq ranch not merged yet14:45
dendrobatesttx: what is the score?14:45
ttxdendrobates: a ttx IA scoring algorithm.14:46
ttxdendrobates: takes into account various things (like whether something is targeted or not, length of patch, age...) to produce a priorityscore14:46
*** mdomsch has quit IRC14:47
dendrobatesttx: I'm impressed14:47
ttxthat's what I need to make a bit more releasecycle-aware14:47
dendrobatesthanks fro doing this14:47
ttxI'm trying to replace myself with a bot, ultimately.14:47
sandywalshttx, cool!14:47
ttxGlance and swift lovers can go directly to
*** doude has joined #openstack14:48
dendrobatesttx have we documented anywhere what is expected for a code review?14:49
westmaashey ttx, one other thing that might be good (if you can get it) is to consider if a branch has prereqs14:49
westmaasyou probably already have that planned and I should just stay quiet :D14:50
ttxwestmaas: in scoring ? yes :)14:50
westmaashaha yeah, figured14:50
ttxCode lives here, btw:
ttxthere is a cli tool and pluggable sorting algorithm :)14:51
ttxcode is probably looking a bit ugly. Mostly written during my train ride yesterday.14:54
*** BigValen has quit IRC14:55
*** BigValen has joined #openstack14:56
*** sparkycollier1 has joined #openstack14:59
*** dragondm has joined #openstack14:59
*** imsplitbit has joined #openstack14:59
*** BigValen has quit IRC15:00
*** nelson has quit IRC15:00
*** nelson has joined #openstack15:00
*** nelson has left #openstack15:00
*** rnirmal has joined #openstack15:01
*** fabiand_ has quit IRC15:04
*** BigValen has joined #openstack15:05
tr3buchetsandywalsh: yep, now i just need to get it mirrored in hyperV :)15:07
justinsbWhich is the "preferred" network driver I should be using for a small (<10 nodes) test nova install?15:17
*** pvo has joined #openstack15:18
*** jbryce has quit IRC15:20
*** dlafferty has joined #openstack15:30
*** enigma has joined #openstack15:30
*** BigValen has quit IRC15:31
dlaffertyjaypipes:  I think you'd mentioned yesterday that you'd success in writing a migrate script that took into account problems with sa-migrate support for mysql.  Where would that be posted?15:32
*** troytoman-away is now known as troytoman15:32
*** dendrobates is now known as dendro-afk15:33
rykeris swift the same as the open stack object storage component?15:33
rykerie. on the page, there are links for 'OS compute' and 'OS object storage'.  Is 'OS object storage' = swift?15:34
*** mdomsch has joined #openstack15:35
rykerI guess it is the same.  The github page shows (swift) in parenthesis.
*** fabiand_ has joined #openstack15:39
*** Rajaram has joined #openstack15:42
*** fabiand_ has quit IRC15:43
annegentlettx: Just curious, and didn't see it on the mailing list anywhere. How is the "Score" calcuated on the Nova branch reviews at
jarrodwhy did you guys write your own image service (glance) rather than use something like s3?15:45
justinsbannegentle: I think its (awesome factor / pep8 violations) ;-)15:46
justinsbannegentle: "awesome factor" is better defined here:
annegentlejustinsb: heh. well it changes daily/hourly? :)15:46
justinsbannegentle: Looks like age is a factor... older patches get a boost15:47
annegentlejarrod: jaypipes probably has a better response, but S3 is supported storage for Glance images if I understand it correctly.15:47
annegentlejustinsb: ah, nice find!15:47
*** rnirmal has quit IRC15:47
*** rnirmal has joined #openstack15:49
ttxannegentle: the definition will evolve based on where we are in the cycle.15:49
*** MotoMilind has joined #openstack15:49
justinsbI'm hitting  virConnect instance has no attribute 'nwfilterDefineXML' when running the default libvirt in Lucid.  Do I take it that Lucid libvirt is not supported?15:51
kim0Hi folks, just letting you know "Ubuntu Cloud Days" starting in 10mins in #ubuntu-classroom .. Thanks15:51
*** blallau has joined #openstack15:52
*** bcwaldon has quit IRC15:57
uvirtbotNew bug: #741077 in nova "IP groups improperly unimplemented [sic]" [Undecided,New]
*** BigValen has joined #openstack16:03
*** gaveen has quit IRC16:05
*** h0cin has joined #openstack16:07
*** gregp76 has joined #openstack16:10
jaypipesdlafferty: it's in review...16:11
*** BigValen has quit IRC16:11
jaypipesannegentle: S3 engine only supports reads, not delete or posts, and it's not tested.16:11
jaypipesjarrod: the reason is because we don't control S3. Amazon does.16:12
jaypipesjarrod: plus, S3 is not just an image service. it's an object storage service, and we already had the excellent Swift project...16:12
jarrodwell, when I say S3, I was really referring to Swift16:13
jarrodmy fault16:13
*** joearnold has joined #openstack16:13
joloughlinhave a problem with glance + nova16:14
joloughlinas per the nova update i am trying to use glance16:14
joloughlinas i have  a number of compute nodes16:14
joloughlinglance is installed and i have added images to it16:14
joloughlinglance details shows those images16:15
*** ramkrsna has quit IRC16:15
joloughlinI've updated nova to use glance with image_service16:15
*** yes456 has joined #openstack16:15
joloughlinand glance_host options16:16
joloughlinbut euca-describe-images returns an unknown error16:16
joloughlinhave i missed something ?16:16
yes456Rackspace made governance changes without talking to the development community or the sitting governance board.16:17
yes456is it write??16:17
yes456if yes why ???16:17
*** blallau has quit IRC16:18
*** bcwaldon has joined #openstack16:19
pvoyes456: this is the developer channel. I'm not sure you're going to get those questions answered here.16:19
*** guigui has joined #openstack16:20
*** maplebed has joined #openstack16:20
*** bcwaldon_ has joined #openstack16:21
*** yes456 has quit IRC16:22
joloughlineuca-describe-images problem after change over to glance16:24
joloughlinanybody have a similar problem ?16:24
*** ScaleUpScott has joined #openstack16:25
*** pharkmillups has joined #openstack16:25
*** skiold has quit IRC16:25
*** adiantum has quit IRC16:26
*** matclayton has joined #openstack16:27
mtaylorjaypipes: morning hombre!16:29
*** enigma has quit IRC16:29
jaypipesmtaylor: hi!16:29
mtaylorjaypipes: how're tricks?16:29
joloughlini have an image service issue after moving to glance, glance itself appears to be working16:30
joloughlinglance details16:30
joloughlinreturns image info16:30
jarrodjoloughlin: does the novatools api work?16:30
joloughlinbut a euca-describe-images breaks16:31
jaypipesmtaylor: need some assistance on Jenkins :) Assistance == need someone to bzr branch lp:~jaypipes/glance/bug731304; cd bug713304; ./ -N16:31
joloughlinjarrod: not sure what that is16:31
*** BigValen has joined #openstack16:31
mtaylorjaypipes: I did - it worked16:32
jaypipesjoloughlin: note that Glance has nothing to do with the EC2 API (and therefore, eucatools doesn't talk to Glance). eucatools talks to Nova, which may use Glance as its image service.16:32
jaypipesmtaylor: hmm... but python test fails. :(16:32
mtaylornope. that works too :)16:32
jaypipesmtaylor: huh? but the Tarmac job is hanging there?16:32
jaypipesmtaylor: over and over again...
mtaylorit works when I do it as jenkins16:33
mtaylorI know16:33
joloughlinjaypipes: so my nova config is probably wrong ?16:33
jaypipesjoloughlin: pastebin it for me.16:33
joloughlini just added an image_service and --glance_host16:33
*** joearnold has quit IRC16:34
*** kashyap has quit IRC16:34
jaypipesmtaylor: so... whatcha think is going on with it?16:34
mtaylorjaypipes: ok - lemme look at something else16:34
mtaylorjaypipes: NO CLUE16:34
jaypipesmtaylor: since that job has now been running for >2 days ;)16:34
mtaylorjaypipes:  oh - weird. ok. so, I'm runing tarmac by hand and now I'm seeing that it's doing something odd...16:35
*** Ryan_Lane has joined #openstack16:42
openstackjenkinsProject nova build #687: SUCCESS in 2 min 8 sec:
openstackjenkinsTarmac: This branch contains the fix for bug #74092916:43
openstackjenkinsIt makes sure cidr_v6 is not null before building the 'ip6s' key in the network info dictionary.16:43
uvirtbotLaunchpad bug 740929 in nova "Spawn error if cidr_v6 is null for network" [Undecided,In progress]
openstackjenkinsThis way utils.to_global_ipv6 does not fail because of cidr==None.16:43
jaypipesjoloughlin: so sorry, mate, I have to run out of the house right now... dealing with an IRS audit thing (sucks..). Perhaps sirp_ could help you out, ok? mtaylor, I'll catch up with you laterz :(16:43
mtaylorjaypipes: tarmac bug. think it's fixed now16:44
*** adiantum has joined #openstack16:44
joloughlinno worries16:44
*** dspano has joined #openstack16:46
joloughlindoes anybody have glance working with nova?16:46
joloughlini have glance working ok16:46
joloughlinas shown by glance details16:46
uvirtbotNew bug: #741106 in openstack-dashboard "Register button out of place" [Undecided,New]
uvirtbotNew bug: #741109 in openstack-dashboard "Users and Roles table out of place" [Undecided,New]
joloughlinbut have problem now getting nova to use it16:47
joloughlineuca-describe-images complains of unknown problem16:47
*** mdomsch has quit IRC16:50
*** kashyap has joined #openstack16:51
joloughlindoes anybody have glance and nova working together ?16:54
*** MontgoDB has left #openstack16:54
*** donkdonk has joined #openstack17:00
joloughlindoes anybody have glance working with nova ?17:11
justinsbjoloughlin: I'm close, but haven't launched an instance yet.  I thought it was supposed to be our preferred config, so I would think that people do.17:12
justinsbjoloughlin: Are you hitting a problem?17:12
justinsbjoloughlin: Sorry, just read back in the chain17:12
joloughlinjustinsb: i have glance working ok17:12
justinsbjoloughlin: I think you may be hitting this issue:17:12
*** mahadev has joined #openstack17:13
joloughlinjustinsb: thanks, I'll take a look at that17:13
*** zigo-_- has quit IRC17:13
justinsbjoloughlin: If you pastie your /images/details output, I can tell you for sure...17:14
*** zigo has joined #openstack17:14
MotoMilindDoes anyone know how to change the project name used by python-novaclient?  It is using "openstack" as the default project name.  [pvo, are you there?]17:14
justinsbjoloughlin: There should also be an error message from nova-api, I believe17:14
joloughlinjustinsb: will do17:15
*** BigValen has quit IRC17:15
*** nelson has joined #openstack17:15
joloughlinthats output of glance details17:16
joloughlinthats my nova.conf17:16
justinsbjoloughlin: Anything in the nova-api log?17:17
justinsbjoloughlin: And can you pastie the raw output from glance, from this:17:17
justinsbjoloughlin: curl
joloughlinok, will do17:18
justinsbjoloughlin: Sorry, wrong URL17:18
justinsbjoloughlin: curl
*** Glace has joined #openstack17:19
*** neogenix has joined #openstack17:19
joloughlinfor curl output17:20
justinsbjoloughlin: Yup, looks like the same bug17:20
justinsbjaypipes: You around?17:20
*** dendro-afk is now known as dendrobates17:20
uvirtbotNew bug: #741123 in nova "EC2 instance ID handling deals with integers, not strings" [Undecided,New]
joloughlinok thanks form help17:22
justinsbjoloughlin: So that's the problem, but I need jaypipes to let me know what the actual fix should be17:22
justinsbjoloughlin: According to the merge proposal I sent to you, if you upload using nova-manage, you bypass the bug17:23
justinsbjoloughlin: You'll probably have to delete the existing images from glance to stop the EC2 API crashing17:23
joloughlini'll give that a go17:24
*** bkkrw has quit IRC17:25
justinsbjaypipes, vishy: Could use some help figuring out what the right thing to do is on
dlaffertyjaypipes: Cool! (Sorry for late response)17:28
*** MarcMorata has quit IRC17:29
*** johnpur has joined #openstack17:30
*** ChanServ sets mode: +v johnpur17:30
joloughlinjustinsb: i deleted images with glance delete, adding them using nova-manage17:35
*** Nacx has quit IRC17:35
joloughlineuca-describe-images now shows them17:35
joloughlinand i have images running on both compute nodes17:35
joloughlinthanks for help17:35
*** dragondm has quit IRC17:37
*** paltman has quit IRC17:38
*** hggdh has quit IRC17:41
*** paltman has joined #openstack17:41
*** matclayton has left #openstack17:41
*** hggdh has joined #openstack17:41
*** enigma has joined #openstack17:45
*** Guest9516 has quit IRC17:46
justinsbjoloughlin: Great - thanks for letting me know!17:48
*** dragondm has joined #openstack17:50
*** dragondm has quit IRC17:51
*** dragondm has joined #openstack17:52
*** guigui has quit IRC17:53
*** joloughlin has quit IRC17:55
*** pvo has quit IRC18:00
*** arun_ has quit IRC18:01
*** pvo has joined #openstack18:03
jaypipesjustinsb, dlafferty, mtaylor: ok, back now. whatup?18:08
jaypipesmtaylor: saw you fixed tarmac. ty so much!18:08
justinsbjaypipes: I'm trying to figure out "who is to blame" and what the fix should be on
jaypipesjustinsb: k, lemme look..18:10
jaypipessoren: looks like swift needs to be installed on the PPA builder:
mtaylorjaypipes: woohoo18:11
MotoMilindHaving trouble with nova-client, images.create('name', id), I get HTTP 500 with empty message.  Any suggestions on how I should debug?  Thanks.18:12
jaypipesjustinsb: so, AFAICT, Glance is behaving exactly as documented in that branch. The properties collection will be blank there, because you aren't add'ing any fields that aren't in the main image attributes collection. If you, for instance did this: glance add kernel_id=12345, then Glance would return the image metadata with a properties collection of {'kernel_id': '123435'} since "kernel_id" is not one of the base attribut18:14
jaypipeses of image. Make sense? You can store any number of arbitrary properties with an image. Arbitrary properties are any field not in ('id', 'name', is_public','status','disk_format','container_format','created_at','updated_at','deleted','deleted_at')18:14
*** lionel_ has quit IRC18:14
*** lionel_ has joined #openstack18:16
*** jaypipes is now known as jaypipes-dentist18:16
MotoMilindAll the REST calls the nova-client is making, should the server log them in nova-api.log?18:16
justinsbjaypipes-dentist: Makes sense.  I guess that my fix is in the right place then - if the type isn't one of the expected values, or is blank, we have to tolerate that.18:17
*** mdomsch has joined #openstack18:18
justinsbjaypipes-dentist: Since type isn't one of the 'non-arbitrary' properties18:18
*** johnpur has quit IRC18:18
justinsbjaypipes-dentist: i.e. the metadata properties :-)18:18
jaypipes-dentistsirp_: *excellent* review on the vsphere patch. excellent.18:18
jaypipes-dentistjustinsb: sure. you can do a type=blah, or you could use the disk_format=aki container_format=aki (or ari, ami, etc...)18:18
sirp_jaypipes-dentist: thanks a bunch! spent 2 hours on that one18:18
jaypipes-dentistsirp_: I know the feeling! but the reviews are worth it. thank you!18:19
jaypipes-dentistok, now i *really* am off to the dentist...18:19
justinsbjaypipes-dentist: Ah.  So perhaps the bug is that the EC2 api should be picking up disk_format or container_format.18:19
jaypipes-dentistjustinsb: bug sirp_ about that one until I return from getting my teefs cleaned. ;P18:19
justinsbjaypipes-dentist: Will do!18:20
justinsbsirp_: Any thoughts on this issue?
sirp_justinsb: looking now (and catching up on IRC above) :)18:22
pvoMotoMilind: should, yes18:26
*** clauden_ has joined #openstack18:28
*** adiantum has quit IRC18:30
MotoMilindpvo: Odd.  The log lists nothing.18:30
MotoMilindpvo: do i need to turn on a debugger?18:30
pvoMotoMilind: you're running the osapi and not the ec2 api, right?18:30
MotoMilindpvo: i am afraid I don't know18:30
sirp_justinsb: sorry for the delay, yeah, ultimately we need to have the BaseImageService abstract out the notion of image_types. I think going foward we want to use `disk_format` and `container_format`; so this would mean changing making changes to the S3ImageService to map image_types to the appropriate disk_format+container_format pair18:31
MotoMilindHow do I find out?18:31
*** Glace has quit IRC18:31
MotoMilindThis is interesting: 2011-03-23 11:35:21,303 INFO nova.api [-] Running ec2 API18:31
MotoMilind2011-03-23 11:35:21,470 DEBUG nova.api [-] App Config: osapi18:31
pvoMotoMilind: one sec. I'm in a business meeting. :/18:32
sirp_justinsb: the related_images patch adds some code to make this translation easier, once that hits, i think we can work on unifiying the concepts of image_types amongst all of the ImageServices18:32
pvoI'm getting called to teh whiteboard.18:32
MotoMilindoh sorry, sure18:32
justinsbsirp_: Thanks. And until that patch hits, I guess we just need to create the images with nova-manage.18:33
*** johnpur has joined #openstack18:36
*** ChanServ sets mode: +v johnpur18:36
*** Glace has joined #openstack18:36
*** ptremblett has quit IRC18:39
sirp_justinsb: yeah (added comment on mergeprop); and if we can't get it fixed TheRightWay soon, i'm okay with a stop-gap with a FIXME going in to work-around18:39
pvoMotoMilind: when you do a netstat -anop, what ports are open? 8774 or 8773?18:40
MotoMilindoh hey, yeah, so seems like both ports are open18:40
edaysandywalsh: can you expand on your last email re: instance id's? why can't a instance ID change? (think of it as a copy+delete rather than a move)18:40
MotoMilindi played with changing the url back and forth18:40
edayif it's within the same zone, it could of course keep the same id18:40
MotoMilindand anything at port 8773 prints out in the nova-api.log file18:40
MotoMilindbut fails to actually do the real work18:41
pvoMotoMilind: hmm. 8773 is the ec2 port18:41
MotoMilindwhereas anything to port 8774 actually works, but fails to log18:41
MotoMilindlet me pastie the startup log, it looks odd18:41
uvirtbotNew bug: #741169 in nova "nova-manage argument parsing problems" [Undecided,New]
westmaaseday: I think the idea is that even if an instance is moving zones (especially internal zones the customer has no knowledge of) we wouldn't want the ID to change18:42
westmaasbut maybe we should wait for sandywalsh to answer you :)18:43
pvowould we actually *move* an instance between zones?18:43
MotoMilindpvo: I have pasted the nova-api.log18:43
pvoinstead of snapping and restoring?18:43
edaywestmaas: if an instance changes zones, internal or not, I guess I don't see an issue with a rename18:43
*** ScaleUpScott has quit IRC18:43
*** vinay_ has quit IRC18:43
edaymost likely moving zones means an IP change too, which is a bigger deal18:44
westmaasyeah, true18:44
westmaas(to both of you)18:44
edaybut within a zone, we don't need to change ID or IP18:44
ironcamelxtoddx: are you around?18:44
*** mrchrisadams has quit IRC18:44
pvoeday: agree with that18:45
pvoMotoMilind: looking18:45
MotoMilindpvo: thanks18:45
ironcamelsandywalsh: can you flip the switch on it looks like it has 2 approvals, but the status is not updated because it was resubmitted18:45
*** adiantum has joined #openstack18:46
uvirtbotNew bug: #741171 in glance "mark properties as deleted if they don't exist on an update" [Undecided,In progress]
pvoMotoMilind: it'll be a bit. I'll need to launch my api nodes to look at the logs. Will do soon as I'm back at my desk.18:47
pvounless someone else verifies it first. : )18:47
*** hggdh has quit IRC18:48
sandywalshwestmaas, eday pvo My thinking was an instance id may be stored by the customer for automation, etc. Therefore we can't change instance id in the background... however, perhaps some operations respond with a new instance_id (such as migration)18:49
*** joearnold has joined #openstack18:49
pvosandywalsh: they shouldn't change in teh same zone18:49
sandywalshno, they shouldn't change in the same zone.18:50
sandywalshI don't think18:50
pvoif they have to go to another zone, it would be on the request of the customer or something else bad18:50
westmaasnever a need to balance zones?18:51
pvoyou'd have to re-ip18:51
westmaasor if there is, we'd have to initiate customer contact18:51
pvowhich is a nono, unless its super critical18:51
sandywalshdepends on the resolution of a zone18:51
comstudre-iping depends on what a 'zone' is18:51
edaywestmass: I think we just need to avoid rebalancing at a zone level with proper capacity planning (can rebalance hosts within a zone)18:51
sandywalshwhat if zone = huddle?18:51
pvosandywalsh: thats likely the resolution18:52
pvoI don't think RS will change its network configuration18:52
pvobut for others it will be different18:52
sandywalshso you don't foresee migrations between huddles?18:52
pvoyou'd *have* to reip18:52
*** h0cin has quit IRC18:53
*** kbringard has joined #openstack18:53
*** joearnold has quit IRC18:53
pvoits more about networking than how big a zone can scale with code18:53
pvowe could go super wide if it was just a code issue18:53
kbringardhey guys, I looked around the wiki and launchpad a bit but didn't see anything obvious... is there a targeted release for Glance to be supported by the EC2 api?18:53
sandywalshyeah, true18:53
donkdonksorry, i just stepped into this. can we not have something that does the re-ip as part of a extra-huddle move?18:54
comstudyou'd still probably want to break up 'super wide' into sub zones18:54
comstudbut networking restrictions kinda force that18:54
edayI think not allowing proper migrations between zones is an ok restriction. you can always backup+restore in another zone18:54
*** fabiand_ has joined #openstack18:54
donkdonkeday: good point18:54
pvoeday: yea, thats what I said above too.18:55
xtoddxironcamel: whats up?18:55
edayand if you have automated systems that do that, just don't hardcode your instance IDs, just autodiscover/cache your working set18:55
pvomigrating myself to another zone. brb18:55
*** pvo has quit IRC18:56
*** AimanA has joined #openstack18:56
*** mgoldmann has quit IRC18:56
sandywalsheday, I think that's a change in the assumptions of instance_ids18:58
sandywalshthere was a quote in the thread about how amazon treats them18:58
*** old_devel has joined #openstack18:59
edaysandywalsh: ahh, I must have missed that19:01
edaywe can't really compare to amazon though for any of this, since they control 100% of their deployments and do whatever they like19:01
*** hggdh has joined #openstack19:01
*** hggdh has joined #openstack19:01
*** adiantum has quit IRC19:01
edayfederation is a whole new game19:01
*** h0cin has joined #openstack19:04
sandywalsheday, true ... I need to think more about the implications of saying "instance id's can change".19:06
sandywalsheday, are you in agreement that it can only happen via a user-initiated action?19:06
edaysandywalsh: they can't, depending on how you define an instance19:06
sandywalsheday, ah, so you're saying migration implies killing an old instance and creating a new one?19:07
edaysandywalsh: if I make a copy of an instance an deploy it in a new zone, I get a new ID for that instance. I now have two instances. nothing changed. I can now delete my first instance19:07
sandywalsheday, right, I think we're saying the same thing in this case19:08
edaywithin a zone, I can migrate between compute hosts, and nothing changes (ids, ips, ...)19:08
sandywalsheday, my concern was that instance id might be ephemeral in some fashion19:08
*** _vinay has joined #openstack19:08
sandywalsheday, true19:08
edaysandywalsh: I was wondering that too, but I don't think so if we define things this way19:09
*** AimanA has quit IRC19:09
*** scottsanchez has joined #openstack19:09
* scottsanchez waves19:10
creihtjustinsb: are you going to take another stab at the openstack volume api?19:10
*** patcoll has joined #openstack19:10
*** AimanA has joined #openstack19:11
justinsbcreiht: I essentially re-proposed my volumes API as a clean patch, as a starting point...19:12
*** pvo has joined #openstack19:13
justinsbcreiht: Sorry, was waiting on LP :-)
creihtahh cool, just found it19:13
justinsbcreiht: No attach / detach yet...19:13
creihtYeah that was going to be my next question :)19:14
justinsbcreiht: I'm actually trying to bring up a nova install so I can test some of this stuff in a more sustainable way19:14
creihtok, just wanted to make sure that something was happening on that side of things19:14
justinsbcreiht: If you have a suggestion on attach / detach, feel free to suggest it!19:14
justinsbcreiht: Or any objections to the volume API as it stands19:15
creihtWe'll take a look at it19:15
justinsbcreiht: As it's on a tight schedule if we're going to try to get it into Cactus19:15
creihtas for attach/detach, I'm not at a point where I can even begin to comment :)19:15
creihtyeah I know19:15
dprincejustinsb: If we were to make volumes an extension we could add action extensions to /servers.19:15
*** adiantum has joined #openstack19:15
*** fabiand_ has quit IRC19:15
dprincejustinsb: That would allow you to implement attach/detach actions for volumes cleanly.19:16
*** drico has joined #openstack19:16
dprincejustinsb: Additionally the existing extensions mechanism supports top level controllers as well (/volumes).19:16
dprinceJustinsb: making what you've got an extension would be fairly trivial.19:16
creihtjustinsb: is your hesitation on attach/detach mostly due to the different vms supporting it differently?19:16
dprincejustinsb: and the actions would be a drop in.19:16
*** BigValen has joined #openstack19:17
justinsbcreiht, dprince: No hesitation on attach / detach, just one step at a time and then got side-tracked onto a few snafus with setting up nova (hence the stream of bugs I keep filing)19:17
*** fabiand_ has joined #openstack19:18
justinsbdprince: I looked at making it an extension, but if it's going to live in nova, I sort of didn't see the point.19:18
justinsbdprince: I'd have to add an extensions directory that pointed inside the nova code tree19:18
creihtyeah, it is kinda core to everything19:18
justinsbdprince: But I think we all agreed that it should be an extension19:18
creihtjustinsb: needs tests ;)19:18
dprinceAgreed. It is core. But until it is part of the SPEC making it an extensions seems somewhat clean.19:19
justinsbdprince: So could use some help!19:19
*** fabiand_ has left #openstack19:19
justinsbcreiht: I know... I do have tests in a separate branch... will get them merged in.19:19
justinsbcreiht: Just feel like I'm so close to getting nova instance to boot, and then keep hitting one snag after another19:19
termiehrm, trunk seems broken to me19:19
termie(unittests failing)19:19
termieAttributeError: 'IptablesFirewallTestCase' object has no attribute 'out_rules'19:20
creihtjustinsb: yeah sorry... not trying to rush you19:20
dprinceAnd by clean I mainly mean a /v1.0 OS API request wouldn't handle /volumes. Just v1.1 extensions or later.19:20
*** johnpur has quit IRC19:21
justinsbcreiht: No worries!  I'm not territorial either if you want to add attach/detach :-)  I should get there today though19:22
justinsbdprince: Do you think I should create a nova/extensions/volume directory and patch your extensions patch to look at nova/extensions ?19:23
justinsbdprince: And then maybe allow users to enable / disable "built-in" extensions through flags?19:23
creihtjustinsb: I'm not anywhere near being able to do nova dev yet19:24
creihtJust trying to help this part any way that I can at the moment19:24
dprinceThat sound reasonable to me. Volumes would show up as a v1.1 extension then which seems reasonable.19:24
creihtand it is a bit late in the game for cactus for me to jump in right now :)19:25
*** lionel_ has quit IRC19:25
dprinceBesides foxinsox you'll be the first to use the new extensions stuff so I'm happy to work with you if you need it.19:25
*** Ryan_Lane is now known as Ryan_Lane|food19:26
dprincejustinsb: Also the extensions path (directory) is configurable already.19:29
*** lionel_ has joined #openstack19:29
dprincejustinsb: and simply moving the extension file ( or whatever out of that directory would disable it.19:29
justinsbdprince: I'm thinking that maybe there should be 2 extensions directory - one for 'user extensions' and one for 'system extensions' aka 'incubating extensions'19:30
justinsbdprince: But I'll just pass the flag for now :-)19:31
*** reldan has quit IRC19:33
kim0soren: hi o/ Just a reminder .. your session in 30mins19:33
*** brd_from_italy has joined #openstack19:33
*** adiantum has quit IRC19:33
*** dprince has quit IRC19:36
*** viirya has quit IRC19:37
*** viirya has joined #openstack19:37
*** dprince has joined #openstack19:39
*** donkdonk has left #openstack19:40
vishyjustinsb, sirp_, jaypipes: I added container_format and disk_format to the s3 image service, but I continued to key the prefix off of properties['type'].  I was just trying to minimize changes.  There is no reason why it couldn't key off of disk format and container format.  The only complaint that I have is that I'm not sure ami is not really a disk format, and glance forces disk format and container format to be == if they are ami.19:42
*** bkkrw has joined #openstack19:42
*** ptremblett has joined #openstack19:42
termievishy: how are you online? shouldn't you be on a plane?19:44
termiedoes test_retry_on_failure fail for anybody else?19:45
kim0soren: hi .. there ? your session in 15mins19:45
sorenkim0: I'm here now.19:45
kim0soren: great thanks :)19:45
termiei get a resource unavailable error after latest checkout19:45
sorenJust wandered in.19:45
kim0soren: #ubuntu-classroom19:45
sorentermie: Weird. Haven't seen that.19:46
termiei am on mac but we do expect unittests to pass there19:47
termiei suspect it is something that works on linux but not mac if it got past jenkins19:48
termieoh, that test is doing all kinds of fancy things19:49
vishytermie: i left way to early19:50
sorenfsvo "fancy"19:50
termiesoren: ;) it writes a shell script out to a temporary file desscriptor and then attempts to execute it19:50
sorentermie: Oh, *that* unit test.19:51
sorentermie: Sorry, was thinking of one of the virt ones.19:51
sorentermie: I wrote test_retry_on_failure, I think. :)19:51
sorentermie: Can I see the traceback?19:51
termiesoren: looked like your handywork ;) i'll see if i can tweak it to work on mac19:52
sorenWeird. Why would stderr be unavailable?19:53
* soren blames eventlet19:53
termiesoren: well if the test is passing for you on linux i would probably start by blaming mac19:54
termiesoren: although i don't yet see anything that looks too out of place19:54
sorentermie: Hm... Python 2.6.5..19:54
sorenYeah, 2.6.5 should be fine.19:55
termiesoren: why did we add "attempts" code instead of just putting the retries in a loop where applicable?19:56
sorentermie: Less duplication?19:57
termiesoren: seems to make the execute function much smarter than it should be19:57
uvirtbotNew bug: #741214 in nova "Automatically unrescue rescued instances after N hours" [Undecided,In progress]
termiesoren: seems like a wrapper method might be better if it is being used often, i'm not familiar with where it is in use yet19:58
*** joearnold has joined #openstack19:58
*** zigo has quit IRC19:59
sorentermie: Feel free to make it so.20:00
*** pvo has quit IRC20:01
*** mahadev_ has joined #openstack20:01
*** hggdh has quit IRC20:02
termiesoren: are we even using this retries code?20:03
termie(nova)termie@preciousroy:p/nova/nova % grep -r execute * | grep attempts                              [master] 13:02:3720:03
termieresults in no output20:03
*** mahadev has quit IRC20:03
*** Ryan_Lane|food is now known as Ryan_Lane20:04
*** jaypipes-dentist is now known as jaypipes20:05
termiesoren: ah found a couple in linux_net20:05
*** hggdh has joined #openstack20:05
*** hggdh has joined #openstack20:05
*** BigValen has quit IRC20:08
*** Rajaram_ has joined #openstack20:11
*** Rajaram has quit IRC20:11
*** Rajaram_ is now known as Rajaram20:11
*** reldan has joined #openstack20:13
*** imsplitbit has quit IRC20:15
*** Rajaram has quit IRC20:16
*** BigValen has joined #openstack20:19
*** arun_ has joined #openstack20:19
*** Rajaram has joined #openstack20:20
*** hvaldivia has joined #openstack20:22
*** jrisch has joined #openstack20:23
*** MarcMorata has joined #openstack20:23
*** sparkycollier1 has quit IRC20:25
*** mahadev_ has quit IRC20:25
btorchok I have created a maverick kvm image and I can boot that image fine using libvirt but when I try with nova it fails due to some /dev/vda issue. has anyone seen this ?
*** mahadev has joined #openstack20:26
hvaldiviaHi everyone20:27
*** ctennis has quit IRC20:27
hvaldiviaI tried to create a volume using euca-create-volume but it doesnt work20:27
hvaldiviaI followed the instructions from
*** mdomsch has quit IRC20:30
*** brd_from_italy has quit IRC20:30
*** sparkycollier1 has joined #openstack20:30
hvaldiviahere the log of nova-volume service
old_develhvaldivia: aptitude install iscsitarget open-iscsi and then restart nova-volume ?20:33
*** dprince has quit IRC20:33
*** joearnold has quit IRC20:33
termiejustinsb: btw, fixed the rename vs remove-add thing for git-bzr-ng20:37
hvaldiviaold_devel: thanks20:37
termiejustinsb: somebody submitted a patch for it20:37
hvaldiviaold_devel: I got this error
justinsbtermie: Thanks... giving bazaar a try for now.  Might revisit once we're past feature-freeze :-)20:38
*** ctennis has joined #openstack20:39
*** dinnerjacket has joined #openstack20:39
*** ctennis has quit IRC20:39
*** ctennis has joined #openstack20:39
dinnerjacketHi guys, any of you had a problem with rabbitmq hanging?20:39
dinnerjacketwe're getting a lot of queues with unconsumed messages, and no new connections can be made to the rabbitmq server, etc.20:40
dinnerjacketbtw running maverick rabbitmq-server package 1.8.0-120:40
uvirtbotNew bug: #741246 in nova "file_inject method missing from agent plugin" [Undecided,New]
*** adiantum has joined #openstack20:42
*** sparkycollier1 has quit IRC20:43
jaypipessirp_: pls see my comment in review on
old_develhadrian: looks like a bug, probably fixed in trunk - what version are you running?20:45
sirp_jaypipes: will do, if you get a sec, pls checkout related_images, i think this might clarify some of the issues around metadata in the ImageServices20:45
hadrianold_devel: i didn't ask anything now20:45
jaypipesvishy: ya, re: the ami disk_format must equal container_format... think we should get rid of that rule? was trying to keep some semblance of data cohesion in the Glance data store (for instance, having disk_format of aki and container_format of something other than aki doesn't make any sense, no?20:46
jaypipessirp_: it's next on my list :)20:46
* jaypipes notes that if someone doesn't understand software, some of the questions on Launchpad would seem, well, odd. "Weird behaviour of my new node" is a perfect example. :)20:49
*** sparkycollier1 has joined #openstack20:52
hvaldiviadoes anyone knows how to create volumes ?20:54
hvaldiviaI lost 3 hours trying to create and attach a volume to my instances20:54
*** bcwaldon has quit IRC20:56
jaypipeshvaldivia: justinsb can help you I believe.20:57
*** jaypipes is now known as jaypipes-afk20:57
Ryan_Lanehvaldivia: which driver are you using?20:57
Ryan_Lanehvaldivia: and is nova-volume running properly?20:57
justinsbhvaldivia: Yes, what Ryan_Lane said.  And are you still getting the volume_ref is not defined error?20:58
hvaldiviaRyan:  I get this error from nova-volume
justinsbhvaldivia: Can't believe we haven't fixed that yet... are you running from source or Bexar or something else?20:59
justinsbhvaldivia: Looks like we have fixed that problem in trunk, so this is dependent on what version of code you're running21:00
justinsbhvaldivia: Which is probably rapidly going to get out of my area of "expertise"21:00
hvaldiviajustingsb: Im running this
justinsbhvaldivia: The issue is that if you have some volumes where creation didn't work properly (e.g. software not present), then you can't restart the volume controller due to a typo-level bug21:03
*** h0cin has quit IRC21:04
justinsbhvaldivia: Well, if you deleted the volumes from the DB then you would bypass the bug21:04
hvaldiviajustinsb: yes I have some volumes listed as error and creating21:05
justinsbhvaldivia: This is only safe if you haven't used the volumes, otherwise you risk losing any data you put on them21:05
justinsbhvaldivia: but it sounds like you don't yet have any data on there to lose21:05
hvaldiviayeah… :)21:05
justinsbhvaldivia: You running MySQL?  SQLite?21:05
justinsbhvaldivia: Let me find the table name... probably volumes, but I'll check!21:06
hvaldiviajustinsb: thanks justin21:06
*** DigitalFlux has joined #openstack21:07
justinsbIck... just spotted that our tables are MyISAM :-(21:07
*** Rajaram has quit IRC21:08
justinsbhvaldivia: Anyway, if you log into the MySQL database, you can either do this:21:08
*** Rajaram has joined #openstack21:08
justinsbhvaldivia: delete from volumes21:08
justinsbhvaldivia: update volumes set deleted=121:08
justinsbhvaldivia: I'd probably do the former, but on a production system I'd probably do the latter21:08
justinsbhvaldivia: I'm assuming you don't have any volumes yet!!21:09
justinsbhvaldivia: (At least any working ones :-) )21:09
justinsbhvaldivia: Once you've done that, you should then be able to start the volume controller21:09
hvaldiviajustinsb: That's right, I can delete those records21:10
justinsbhvaldivia: The volumes code is generally going to be much more stable in Cactus - there are some good fixes going in21:10
hvaldiviajustinsb: Let me try...21:10
justinsbhvaldivia: Though that's not much good for you at the moment ;-)21:10
_vinayI am trying to debug an instance which shows running in euca-describe-instance21:11
_vinaybut I cant ssh to it21:11
_vinayI found this error in /var/log/libvirt/qemu/i-0001.log21:12
_vinayqemu: -device rtl8139,netdev=hostnet0,id=net0,mac=02:16:3e:7e:80:8a,bus=pci.0,addr=0x2: pci_add_option_rom: failed to find romfile "pxe-rtl8139.bin"21:12
_vinayany ideas what it might mean???21:12
justinsb_vinay: I believe that error is not serious (unless you're trying to PXE boot, which you're not)21:13
justinsb_vinay: Any other errors in there?21:13
_vinayno.. thats the only error21:14
_vinaythats the last line in the log21:14
_vinaytherefore the suspect21:14
_vinayI use euca-reboot-instances <id>21:15
justinsb_vinay: Well, it's not always the last line that is the problem :-)21:15
_vinaybut nothing happes21:15
_vinay:) agree21:15
kbringardI'd modify the libvirt.xml on the compute node it's running on to add a vnc head, and virsh destroy/create it then VNC in to see what's happening21:15
justinsbkbringard: I was the only person that did that :-)21:15
_vinayI am running this on a single node for now21:15
justinsbkbringard: I mean I thought I was ...21:15
kbringard <graphics type='vnc' port='-1' autoport='yes' listen='' keymap='en-us'/>21:15
kbringardhaha, newp, I do it all the time :-)21:15
kbringard_vinay: ssh to the compute node, and look for /var/lib/nova/instances/$instance-id21:16
justinsbkbringard: I'm hoping sleepsonthefloor  's VNC patch lands soon!21:16
kbringardjustinsb: true story21:16
kbringard_vinay: and add the line I pasted to you (replacing the x's with the IP you want it to listen on) before the last </devices> tag in the libvirt.xml21:17
kbringardthen do virsh list21:17
kbringardthen virsh destroy $instance21:17
kbringardthen virsh create libvirt.xml21:18
kbringardthen you can do netstat -an | grep LIST | grep 590 to verify that the VNC listener came up21:18
vishyjaypipes: not really sure about aki and ari.  A ramdisk is in generally a tar.gz format I think, but i think matching aki and ari for both is probably fine21:18
kbringardthen you can use any VNC client to connect to it and see the console21:18
*** n1md4 has joined #openstack21:19
_vinayok.. kbringard  thanks...21:19
hvaldiviajustingsb: I deleted the volumes records, I got this error, But the cmd euca-create-volume works :)21:20
kbringardoff that topic... anyone know what the status of cactus support in the ec2 api is? justinsb; based on the launchpad repo I found it looks like you're working on that?21:20
kbringard_vinay: no worries, hopefully it helps21:20
_vinaythe ip should be what? the ip of the host where I am running the compute node?21:20
hvaldiviajustinsb: thanks again.21:20
kbringard_vinay: yea21:20
kbringardthe IP that you ssh to the compute node21:20
kbringardif you don't supply one, it starts on localhost21:20
kbringardwhich isn't a problem, but unless you like to setup x forwarding or ssh tunnels it doesn't help so much21:21
justinsbkbringard: Not sure what you mean by cactus support in the ec2 api ?21:21
kbringardI meant21:21
kbringardglance support21:21
kbringardin the ec2 api21:21
uvirtbotNew bug: #741274 in nova "MySQL database tables are using the MyISAM engine" [Undecided,New]
justinsbhvaldivia: Glad to hear it helped.  I'm not sure what that error is.  Looks like you tried to use a volume 4 first?  Maybe one you'd previously deleted?21:22
justinsbhvaldivia: But does look like volume 6 worked.  Let me know if the attach works!!21:22
*** bkkrw has quit IRC21:22
justinsbkbringard: Hmmm.. I'm trying to use glance with EC2.  It seems to work if you create the volume "right"21:23
justinsbkbringard: "right" = with nova-manage21:23
justinsbkbringard: The important thing being that you set the type="machine" property in glance21:23
kbringardoh, interesting21:23
justinsbkbringard: But there seem to be some bugs with nova-manage command line argument processing21:23
*** sebastianstadil has joined #openstack21:23
justinsbkbringard: And also, currently, if you have _any_ images that aren't "right", the EC2 API barfs21:23
termiejustinsb: any idea why changes to the objectstore might affect integrated.test_login tests?21:24
*** dinnerjacket has quit IRC21:24
n1md4evening. After running the script, I'm supposed to run "source /root/creds/novarc" but there's no directory called root/creds21:24
justinsbtermie: Err... no.  Is this on trunk?21:24
n1md4Should I create it?21:25
_vinaykbringard: I didnt supply a ip .. I installed nova using (single node)21:25
_vinayso in this case the listen ip will be ??21:25
termiejustinsb: the branch is based off trunk, yes, but it is a branch that removes twisted from the objectstore code21:25
termiejustinsb: everything is passing except these random login tests21:25
kbringard_vinay: well, you have to manually add that VNC line to the libvirt.xml for the specific instance21:25
termiejustinsb: and your name is stamped a couple places on the tests.integrated.* code21:25
kbringard_vinay: so I'd made the IP be whatever the IP of that machine is21:26
_vinayok I see21:26
kbringard_vinay: that way it'll listen on that address21:26
_vinayand the port remains -1 ??21:26
kbringard_vinay: yea, that should make it assign them incrementally, starting from 590021:26
n1md4Also, when installing the node using the sudo ./ script, I get the following "./ 53: [[: not found" and "Unsupported OS, sorry!" even though I'm running ubuntu, any idea?21:26
justinsbtermie: Yes, tests.integrated is entirely my fault :-)21:27
kbringardjustinsb: cool, do you happent to have any notes on how you've been doing it that you'd be willing to share?21:27
justinsbtermie: It's bringing up the API "for real" and then hitting the TCP socket "for real"21:27
*** dspano has quit IRC21:28
*** bcwaldon has joined #openstack21:28
justinsbtermie: Ah ... here's a thought... are you running the nova-api while also running the tests?21:28
justinsbtermie: And do you have an error message?21:28
justinsbtermie: As it doesn't pick an unused port at the moment21:29
termiejustinsb: nope, but i think there may be something up with a xenapi plugin21:29
termiejustinsb: so i'm looking in ther enow21:29
hvaldiviajustinsb: It works :)21:30
*** nerens has quit IRC21:30
termiejustinsb: looks like it is trying to use the objectstore directly instead of glance21:30
justinsbkbringard: Starting a pastie here... nothing particular exciting yet though...
justinsbhvaldivia: Great news! :-)21:31
justinsbtermie: It = my tests?21:31
kbringardcool, thanks, I'll check it out and perhaps add to it (or start my own with additional notes)21:31
*** enigma has quit IRC21:31
termiejustinsb: xenapi, which i assume is what your tests are testing against21:32
justinsbtermie: They shouldn't be... let me check21:32
*** hvaldivia has quit IRC21:32
kbringardjustinsb: so there are no official blueprints/plans for supporting glance in ec2 yet?21:33
justinsbtermie: LoginTest.test_login is failing with a xen error?  It's just looking for a list of flavors as a "canary" test that login is working21:33
justinsbkbringard: Well, I think it's supposed to "just work", but I've hit a few bugs because I'm not very good at following instructions :-)21:34
kbringardah, i see... I thought vishy, or someone else, said that glance doesn't work with ec2 yet21:34
sirp_re: , could we have it always sort pre-reqs above dependents?21:34
kbringardonly the RS/OS api21:34
sirp_the reviewlist is awesome btw, whoever is responsible for that!21:34
termiejustinsb: i'm not running anything extra locally, i assume if this is testing running services it is starting them itself21:34
*** ecarlin has joined #openstack21:35
justinsbtermie: Yes, it is starting services directly.  (In integrated_helpers)21:35
*** bkkrw has joined #openstack21:35
*** ptremblett has quit IRC21:35
*** ptremblett has joined #openstack21:36
justinsbtermie: That's really odd, if you look at the source of the tests.  All the tests are doing the same thing, it's just that we expect the first to fail and the others to throw exceptions21:36
termiejustinsb: maybe it fails after the first request21:36
termiejustinsb: and the service dies21:36
termiejustinsb: aye, i've looked through the code, nothing seems amiss21:37
justinsbtermie: Yes... I wonder if it's a socket reuse issue...21:37
justinsbtermie: Although I don't think I'm able to restart the service correctly at the moment21:37
vishykbringard: it merged recently21:37
justinsbtermie: So I just leave it running21:37
kbringardvishy: ahh, ok, awesome, thanks21:37
kbringardI'm still running a pretty stock bexar, so that's probably why it doesn't work at all for me21:38
justinsbtermie: Oh ... I see the problem21:38
justinsbtermie: Bugger... false alarm21:38
kbringardso is cactus the target for officially supporting it?21:38
*** allsystemsarego has quit IRC21:39
justinsbtermie: Can I pull the branch from somewhere?21:39
vishykbringard: yes, old objectstore is coming out21:39
termiejustinsb: sure, one sec21:39
kbringardvishy: awesome, perfect, thank you sir21:39
justinsbkbringard: Updated the pastie with the line I'm using to upload images using glance, but no real miracles in there:
termiejustinsb: or lp:~termie/nova/eventlet_objectstore21:42
justinsbtermie: Thanks... I'll branch and see if I can figure it out21:42
*** bcwaldon has quit IRC21:44
justinsbtermie: Good news for you is I can reproduce the problem :-)21:44
justinsbtermie: Bad news for me is I can't just blame it on you :-)21:44
*** BigValen has quit IRC21:45
termiejustinsb: am reading the relevant bug, not aware of a wsgi service shutdown prob21:45
justinsbtermie: It's just I don't know how to shutdown our wrapper around wsgi...21:46
termiejustinsb: i wrote some code for such things, lemme see if i can find it21:46
termiejustinsb: oh, it is actually in that branch21:48
termiejustinsb: check out test.py21:48
termiejustinsb: also is probably why stuff fails21:48
termiejustinsb: because i clean up services after each test :p21:48
openstackjenkinsProject nova build #688: SUCCESS in 2 min 10 sec:
openstackjenkins* Tarmac: Updates the previously merged xs_migration functionality to allow upsizing of the RAM and disk quotas for a XenServer instance.21:48
openstackjenkins* Tarmac: Fix lp735636 by standardizing the format of image timestamp properties as datetime objects.21:48
openstackjenkinsSo far, only glance supports returning image create, update, and delete timestamps. And only the openstack api reports those dates. Thus, the fixes were confined to those two modules.21:48
termiejustinsb: so i guess try removing your recycling21:49
termiejustinsb: and just start a new thing every time21:49
*** romain_lenglet has joined #openstack21:49
*** ecarlin has quit IRC21:49
termiejustinsb: and if it works, tell me how to do it and i'll apply it to my branch :p21:49
*** romain_lenglet has quit IRC21:49
justinsbtermie: Thanks for the link on how to shut down wsgi.  I'm still going to try to track down what's really going wrong here.21:49
termiejustinsb: that is what is really going wrong21:49
termiejustinsb: my stuff kills the service21:50
justinsbtermie: Oh!21:50
termiejustinsb: so after the first test it gets wiped out21:50
termiejustinsb: which is what we really wanted anyway21:50
termiejustinsb: so this branch can close the bug, too21:50
justinsbtermie: Why would you go and do a mean thing like that? :-)21:50
termiejustinsb: in, _monkey_patch_wsgi... works automagically21:50
btorchdoes nova for some reason start booting an instance without the full disk being downloaded ?21:51
*** ianw has joined #openstack21:51
termiebtorch: not that i can think of21:51
termiebtorch: is this related to your injection issue?21:51
termiebtorch: is it possibly you are creating an instance without assigning it any disk space or something odd like that?21:51
btorchtermie: hmm you mean the vda disk issue ?21:51
termiebtorch: yeah, the issue comes up when it tries to inject data, i thought21:52
termiebtorch: from your pastie21:52
termiebtorch: i'm not super familiar with that step though21:52
*** DigitalFlux has quit IRC21:52
*** BigValen has joined #openstack21:52
btorchtermie: I found this bug report 736745 where soren stated that since using a disk image instead of a filesystem image I should not need to specify a kernel/ramdisk while uploading/registering  the  image21:53
*** jrisch has quit IRC21:54
btorchtermie: so I'm trying that out, the disk file stopped growing (just 80MB) and the nova-compute.log just says that instance instance-00000006: booted21:54
btorchI don't see any errors this time
termiejustinsb: fixed on my end21:57
termiejustinsb: made it not be a singleton anymore21:57
termiejustinsb: checking rest of the tests now, but i think i've got it21:57
kbringard_vinay: did that VNC stuff work for you? and did it help resolve your problem?21:58
_vinaykbringard I just setup the vnc21:58
kbringardah, cool21:58
*** ecarlin has joined #openstack21:58
_vinayso this was just to obtain the console on compute node?21:58
kbringardwell yea, on the instance that isn't booting21:58
kbringardso you can see why it's not booting21:59
_vinayhmm well it gave me the console on my machine21:59
_vinaywhich I already have :)21:59
_vinayyeah.. and I tried to give it ip of -- which is supposed to be the ip of the instance22:00
kbringardif you do netstat -an | grep 590 is there a 5901, 5902... etc?22:00
kbringardwell, you'll need to give it the IP of the host, which is the machine you're running it on22:00
termiejustinsb: yup, all good, yay :)22:00
_vinayYes I have all these ports22:00
kbringardhmmmmm, OK22:00
justinsbtermie: I've also fixed it here22:00
justinsbtermie: But great, no need for a singleton...22:01
kbringarddo any of them have in there?22:01
kbringardlike say or whatever22:01
_vinayI put my IP in libvirt22:02
*** ecarlin has quit IRC22:02
_vinaywhich is the IP of machine I installed nova on22:02
kbringardthat's the IP of the machine you're running nova on22:02
kbringardok, cool22:02
kbringardthen if you run22:02
kbringardnetstat -an | grep
kbringarddoes it return anything?22:03
_vinayso the netstat cmd shows the same *.92 address22:03
_vinayyes that port is 590122:03
kbringardah, ok22:03
_vinaytcp        0      0*               LISTEN22:03
kbringardthen try connecting to with your vnc client22:03
kbringardanother thing you can try is running22:04
justinsbtermie: Shall I pastie the code without all the complaing about wsgi?22:04
kbringardlsof -i | grep 590122:04
*** ecarlin has joined #openstack22:04
kbringardand make sure that a kvm process is actually what has that port open22:04
kbringardor xen22:04
kbringardor whatever hypervisor you're using22:04
*** miclorb_ has joined #openstack22:05
kbringardbut, hopefully connecting to will give you a console on the VM that you're having problems with22:05
*** sparkycollier1 has quit IRC22:06
*** omidhdl has quit IRC22:07
kbringardyou can also mess with the port= and autoport settings to force it to a specific port, so remove any ambiguity22:07
termiejustinsb: i already pushed my branch up22:07
termiejustinsb: but i can add you as a reviewer :)22:07
*** omidhdl has joined #openstack22:07
kbringardanyway, I need to go before my wife gets too pissed, best of luck _vinay!22:07
*** kbringard has quit IRC22:07
termiejustinsb: donez22:08
*** omidhdl has quit IRC22:11
*** patcoll has quit IRC22:12
*** RobertLaptop has quit IRC22:13
*** AimanA has quit IRC22:15
*** ecarlin has quit IRC22:15
*** AimanA has joined #openstack22:15
*** johnpur has joined #openstack22:16
*** ChanServ sets mode: +v johnpur22:16
*** gondoi has quit IRC22:17
*** troytoman is now known as troytoman-away22:17
*** Shentonfreude has quit IRC22:17
btorchis it possible to use kvm disk images with the nova-objectstore ?22:18
btorchI changed my setup to use the nova-objectstore to see that my disk image size issue was perhaps related to glance/swift being on a remote location22:19
sirp_jaypipes: what's the policy on mergeprop rejection, does -2 qualify (heh not regarding any of your patches :P )22:20
justinsbbtorch: What's the issue at the moment with nova-objectstore.  Disk image is only 80MB or so?22:21
MotoMilindpvo: Odd.  On this thread about osapi v/s EC2 Api, when I connect to port 8774, I don't see any debug printed in nova-api.log for list() family of methods, such as flavors.list(), servers.list() and images.list() on the OpenStack instance.  However, I do see a stack trace when I run images.create('test', 12)22:22
MotoMilindpvo: I am pasting the stack trace on the nova-api side for you, but it just tells me that the method is not implemented.  Wonder if you have checked in some code already and that I have an older revision22:23
*** BigValen has quit IRC22:24
*** littleidea has joined #openstack22:24
justinsbbtorch: Ah ... the disk is only 80MB because it's a differencing disk22:24
justinsbbtorch: My is only 8MB, but I think that's why.  So that's not necessarily a problem.22:25
btorchjustinsb: no that was the problem when I was using nova with glance... I tried to use the objectstore now but it keeps asking for a kernel image for the machine image.22:29
btorchjustinsb: going back to my nova/glance setup, so even though my qcow2 disk image is 900MB the disk image file will only be 80MB ?22:29
*** RobertLaptop has joined #openstack22:29
justinsbbtorch: That's what I'm seeing (8MB though)22:30
justinsbbtorch: It's a differencing disk in your pastie:22:30
justinsb qemu-img create -f qcow2 -o cluster_size=2M,backing_file=/var/lib/nova/instances/_base/0000000d_sm /var/lib/nova/instances/instance-00000006/disk22:30
btorchjustinsb: also I never got anything in the console.log file22:31
*** BigValen has joined #openstack22:31
uvirtbotNew bug: #741328 in nova "utils.execute fails retries" [Undecided,New]
*** joearnold has joined #openstack22:32
justinsbbtorch: I'm also struggling (I got it to launch and could log in on VNC), but injection didn't work for me22:32
justinsbbtorch: But from your logs you got injection to work22:32
justinsbbtorch: Hopefully we can both figure it out together :-)22:32
btorchjustinsb: yeah with the injection issue I at least had something in my console.log :)22:33
*** joearnold has quit IRC22:33
justinsbbtorch: What was the injection issue again?  From your pastie, it looks like it worked...22:33
*** topper1 has joined #openstack22:34
btorchjustinsb: but the injection issue apperantly was because I uploaded the disk image providing kernel and ramdisk IDs. Once I uploaded it again without the ramdisk, kernel ids I had no injection issue22:35
justinsbbtorch: Ah... I know this one!22:36
justinsbbtorch: The issue is that there are two disk formats... you can either upload a raw disk image, or just the image of a partition22:36
*** topper1 has quit IRC22:36
justinsbbtorch: If you're uploading just a partition, you need to boot from a kernel you provide to nova22:36
btorchjustinsb: true22:37
justinsbbtorch: If you're uploading a raw disk image, you can boot from the boot sector22:37
justinsbbtorch: So, if you're using a kernel, it assumes you're using a partition22:37
btorchjustinsb: ok, my kvm disk image though is of qcow format22:37
justinsbbtorch: And if you're using no kernel, it assumes you're using a whole disk image22:37
justinsbbtorch: Yes, so if you built your disk image using vmbuilder, you have a whole disk image22:37
justinsbbtorch: And if you specify a kernel it won't work22:37
*** RobertLaptop has quit IRC22:37
justinsbbtorch: Well, injection won't work22:37
btorchnot vmbuilder but yes I have a full disk image22:38
justinsbbtorch: The image may still boot, depending on how clever your image is :-)22:38
*** RobertLaptop has joined #openstack22:38
btorchjustinsb: it boots fine if I use libvirt or qemu directly22:38
justinsbbtorch: So, with glance/nova, without a kernel, is it still broken?22:39
btorchjustinsb: I'll play around more with this later. the other thing I was going to try is tweek the libvirt.xml template22:39
justinsbbtorch: To include VNC?22:40
*** pharkmillups has quit IRC22:40
*** rnirmal has quit IRC22:40
justinsbbtorch: What's the problem you're hitting though?22:40
btorchjustinsb: it never booted up ... nova list shows it active but console.log shows nothing in there and I couldn't ping the vm22:40
justinsbbtorch: Did you check /var/log/libvirt/<instance> ?22:41
justinsbbtorch: I had an issue with permissions on /dev/kvm22:41
btorchjustinsb: no I was checking /var/lib/nova/instance ... indeed I should have checked that ... let me give a quick try before heading out22:41
MotoMilindpvo: I checked my /etc/nova/nova.conf and found out that I don't have any value set for image_service.  However, the code path followed (based on the stack trace) nova-api is using as opposed to local.py22:42
MotoMilindpvo: I am going to try to change the config to use local, s3, glance, etc to see if the results are different22:42
justinsbbtorch: Kudos to _vinay for reminding me about this logfile!22:42
*** mray has quit IRC22:43
btorchjustinsb: that log just shows that is starting up ... no errors22:48
justinsbbtorch: :-(22:48
justinsbbtorch: Can you VNC in?22:48
btorchhold on I can ping it now wth22:48
*** sebastianstadil has quit IRC22:49
*** Tom____ has joined #openstack22:49
btorchjustinsb: ha got in :)22:52
*** RobertLaptop has quit IRC22:52
justinsbbtorch: No way!  I literally just got into mine as well :-)22:52
justinsbbtorch: Pair problem solving at its best :-)22:52
*** RobertLaptop has joined #openstack22:53
*** westmaas1 has joined #openstack22:54
btorchjustinsb: so I guess the adminPass that nova displays is not related to the root account ?22:55
btorchjustinsb: this is custom ubuntu10.10 image so I've setup some accounts and installed apps I wanted already22:55
uvirtbotNew bug: #741339 in nova "libvirt does not support snapshotting" [Undecided,New]
uvirtbotNew bug: #741342 in nova "The word "None" is injected into authorized_keys" [Undecided,New]
justinsbbtorch: I think adminPass is only implemented for xen at the moment22:58
btorchah cool22:58
*** mahadev has left #openstack22:58
justinsbbtorch: I think SSH key injection works though, which is much more secure anyway :-P22:58
btorchjustinsb: true haven't tried that.. I guess with the latest nova I got there were new iptable chains added :(22:59
justinsbbtorch: Your adminPass starts with "ubun"... that can't just be luck (?)23:01
justinsbbtorch: Don't know about the iptable chains (yet).  Is it blocking something?23:01
btorchjustinsb: no, it's just that with my network layout, I have to insert some snatting rules and make a gateway change since my vms are on my localnet23:02
btorchin order to gain access to the tubes :)23:03
MotoMilindpvo: Ok, so setting the —image_service=nova.images.local.LocalImageService apparently allowed me to create an image just fine.  However, that setting prevented me from listing images for any of my projects in the dashboard or in python nova-client!  Quite a conundrum :)23:03
justinsbbtorch: Ah, OK23:03
justinsbAnyone know why the adminPass is being prefixed with the first chars of the server name?23:03
*** mahadev has joined #openstack23:04
*** jero_ has quit IRC23:07
*** MarcMorata has quit IRC23:08
*** Rajaram has quit IRC23:12
*** ianw has quit IRC23:13
*** ianw has joined #openstack23:14
*** rnirmal has joined #openstack23:16
uvirtbotNew bug: #741354 in nova "Generated adminPass is prefixed with the first 4 characters of the server name" [Undecided,New]
*** RobertLaptop has quit IRC23:21
*** RobertLaptop has joined #openstack23:21
*** motohashikn has joined #openstack23:22
*** pvo has joined #openstack23:22
*** sebastianstadil has joined #openstack23:23
*** at141am has quit IRC23:29
*** Oneiropolo has joined #openstack23:31
*** vernhart has quit IRC23:33
*** motohashikn_ has joined #openstack23:35
*** BigValen has quit IRC23:36
*** motohashikn_ has quit IRC23:36
*** joearnold has joined #openstack23:42
*** johnpur has quit IRC23:42
*** RobertLaptop has quit IRC23:44
*** joearnold has quit IRC23:46
*** jeffjapan has joined #openstack23:48
*** dlafferty has quit IRC23:49
*** _vinay has quit IRC23:51
*** Tom____ has quit IRC23:55

Generated by 2.14.0 by Marius Gedminas - find it at!