Monday, 2011-03-21

*** arun_ has quit IRC00:01
*** sophiap has quit IRC00:06
*** pvo has joined #openstack00:06
*** omidhdl has left #openstack00:10
*** arun_ has joined #openstack00:14
*** arun_ has joined #openstack00:14
*** hggdh has quit IRC00:15
*** hggdh has joined #openstack00:16
*** bkkrw has quit IRC00:17
*** sophiap has joined #openstack00:19
*** arun_ has quit IRC00:19
*** Ryan_Lane has joined #openstack00:21
*** arun_ has joined #openstack00:24
annegentlen1md4: For OpenStack Object Storage (Swift), refer to There is not a similar guide for OpenStack Compute (Nova) that I know of, though. I'm the doc coordinator and I haven't seen one yet.00:27
*** Ryan_Lane has quit IRC00:27
*** Ryan_Lane has joined #openstack00:32
n1md4annegentle: thanks.  I've seen that page, but I'm not sure why it's titled for Debian.  The guide uses an ubuntu base system, and ubuntu specific commands, nor does the package 'swift' exist.  I might just install an ubuntu system.00:39
*** kashyap has joined #openstack00:51
*** vapor has joined #openstack00:55
*** Ryan_Lane has quit IRC00:57
*** cascone has joined #openstack01:02
*** sophiap has quit IRC01:04
*** sophiap has joined #openstack01:24
*** sophiap_ has joined #openstack01:25
*** sophiap has quit IRC01:25
*** sophiap_ is now known as sophiap01:25
*** zenmatt has joined #openstack01:29
*** sophiap has quit IRC01:32
*** miclorb_ has quit IRC01:44
*** vish1 is now known as vishy01:51
*** sophiap has joined #openstack01:53
*** freeflyi1g is now known as freeflying02:01
*** cascone has quit IRC02:04
*** vapor has quit IRC02:12
winston-dHi all, I've some questions about CloudFiles PHP bindings.  What if the object I create with '$obj->create_object($name)' already exists on Swift?  I didn't see any special treatment for this condition.  And what if I write content to the object using 'write()' or 'load_from_filename()'?  It will overwrite the existing object or not?02:13
*** miclorb has joined #openstack02:14
*** sophiap has quit IRC02:20
*** Lantizia has joined #openstack02:27
LantiziaHey I'm kinda new to OpenStack... if I was to Compute it doesn't actually do any virtualization does it and requires a hypervisor?  What hypervisor would be recommended for KVM VM's?02:28
HugoKuo_I use nova-manage service list & euca-describe-avalibility-zone vebose ..........a compute-node shows XXX by nova-manage , but it's normal by euca command...02:29
HugoKuo_is this a bug ?02:29
HugoKuo_openstack support various hypervisor02:29
HugoKuo_default is kvm02:29
LantiziaHugoKuo_, right but does that run on the same system as Compute runs on?02:30
HugoKuo_I do think so....02:30
Lantiziayou _do_ think so?02:30
HugoKuo_sorry , I thought you means which hypervisor is recommand.....02:31
HugoKuo_I don't know the real question XD02:31
winston-dLantizia: I think OpenStack itself doesn't do virtualization, it requires underline hypervisor to do that.02:31
winston-dLantizia: what do you mean by 'KVM VM's'?02:32
*** sophiap has joined #openstack02:35
Lantiziawinston-d, what I mean is I have a bunch of VM's that run using KVM on a hypervisor called Proxmox VE... now I'd like to try OpenStack and I'm willing to swap my HyperVisor for whatever is recommended/supported for OpenStack02:38
LantiziaProxmox VE does _not_ use libvirt02:39
*** rchavik has joined #openstack02:50
*** pvo has quit IRC03:08
*** sophiap has quit IRC03:09
*** miclorb has quit IRC03:15
winston-dLantizia: still there?  I haven't tried Proxmox VE yet.  But I guess the VM it uses can be easily put to other KVM implementation, such CentOS, RHEL, Ubuntu or even Xen.03:23
winston-dLantizia: these hypervisors _use_ libvirt. :D03:24
*** pvo has joined #openstack03:24
*** fayce has joined #openstack03:29
Lantiziawinston-d, yeah I'm here - was kinda after a distro that sets up the OS and KVM with a WebUI as well... similar to proxmox - in case you can't manage it via openstack03:37
*** miclorb has joined #openstack03:39
winston-dLantizia: I haven't used any WebUI with KVM.  But as far as I know, with RHEL you can use virt-manager to mange VMs.03:39
Lantiziathat's not exactly web based though03:40
westmaas1there is a dashboard for openstack available - but you mean you want a lower level webui on each host in case the management piece is broken?03:40
Lantiziakind of like how you can connect to ESX/ESXi boxes directly but also manage them all from connecting to the VCenter03:41
westmaas1yeah, I'm not sure about that.  I haven't been working at that level at all yet.03:41
Lantiziagranted that's not web based though03:41
westmaas1maybe there is a web ui for xen server?03:42
Lantiziawould prefer kvm03:43
Lantiziai mean otherwise would I simply be installing debian 6 and kvm and libvirt from repo's and that'd be enough for openstack to manage?03:43
winston-dLantizia: I think that's enough03:44
westmaas1me too, but I'm not sure.03:44
winston-dand maybe less painful to enable OpenStack03:44
Lantiziaopenstack doesn't manage openvz though right? or can it via libvirt?03:45
winston-dopenstack can run on top of any hyerpvisor libvirt supports, i guess03:46
*** pvo has quit IRC03:50
*** fayce has quit IRC03:59
*** kashyap has quit IRC04:23
*** Xenith has quit IRC04:27
*** nijaba has quit IRC04:30
*** Xenith has joined #openstack04:33
*** kashyap has joined #openstack04:58
*** vernhart has joined #openstack04:59
*** mdomsch has quit IRC05:12
*** aimon_ has joined #openstack05:20
*** aimon has quit IRC05:23
*** aimon_ is now known as aimon05:23
*** arun_ has quit IRC05:33
*** koolhead11 has joined #openstack05:35
*** MarkAtwood has joined #openstack05:41
*** MarkAtwood has left #openstack05:42
*** ovidwu has quit IRC05:43
*** f4m8_ is now known as f4m805:50
openstackjenkinsProject nova build #673: SUCCESS in 2 min 9 sec:
openstackjenkinsTarmac: fixes nova-manage instance_type compatibility with postgres db05:53
*** ovidwu has joined #openstack05:57
*** arun_ has joined #openstack05:59
*** fayce has joined #openstack06:00
*** ovidwu has quit IRC06:10
*** ovidwu has joined #openstack06:13
*** zenmatt has quit IRC06:34
*** aimon has quit IRC06:35
*** aimon has joined #openstack06:36
*** ovidwu has quit IRC06:45
*** ovidwu has joined #openstack06:46
*** arun_ has quit IRC06:48
*** ramkrsna has joined #openstack06:48
*** ovidwu has quit IRC06:56
*** ovidwu has joined #openstack06:57
*** ovidwu has quit IRC06:58
*** ovidwu has joined #openstack07:01
*** gasbakid has joined #openstack07:01
*** allsystemsarego has joined #openstack07:07
*** naehring has joined #openstack07:08
*** miclorb has quit IRC07:09
*** gregp76 has joined #openstack07:18
*** fayce has quit IRC07:22
*** zul has quit IRC07:22
*** gregp76 has quit IRC07:25
*** guigui1 has joined #openstack07:28
HugoKuo_is latest trunk the best choice ?07:30
*** nijaba has joined #openstack07:38
*** nijaba has joined #openstack07:38
*** CloudChris has quit IRC07:46
*** befreax has joined #openstack07:47
*** omidhdl has joined #openstack07:47
*** lionel_ has quit IRC07:51
*** Nacx has joined #openstack07:51
*** lionel_ has joined #openstack07:51
*** ovidwu has quit IRC07:57
*** ovidwu has joined #openstack07:59
*** lionel_ has quit IRC08:06
*** lionel_ has joined #openstack08:10
sorenHugoKuo_: Probably.08:11
HugoKuo_soren: thx08:12
*** befreax has left #openstack08:14
*** mgoldmann has joined #openstack08:23
*** mgoldmann has joined #openstack08:23
*** ramkrsna has quit IRC08:25
*** skiold has joined #openstack08:31
*** miclorb_ has joined #openstack08:37
*** ramkrsna has joined #openstack08:37
*** miclorb_ has quit IRC08:48
*** reldan has joined #openstack08:53
*** romain_lenglet has joined #openstack08:56
*** romain_lenglet has left #openstack08:56
*** nijaba has quit IRC09:01
*** bkkrw has joined #openstack09:01
*** miclorb_ has joined #openstack09:02
*** nijaba has joined #openstack09:04
*** nijaba has joined #openstack09:04
*** bkkrw has quit IRC09:07
*** fayce has joined #openstack09:15
*** uksysadmin has joined #openstack09:21
*** kal24 has joined #openstack09:28
*** fayce has quit IRC09:30
*** fayce has joined #openstack09:30
*** guigui1 has quit IRC09:47
*** 15SAACGVK has joined #openstack09:48
*** rchavik has quit IRC09:51
*** CloudChris has joined #openstack09:53
koolhead11how openstack handle OVF better than existing cloud infra in market?10:00
*** z0__ has quit IRC10:02
sorenkoolhead11: We don't.10:06
sorenkoolhead11: Yet, at least.10:06
koolhead11soren, but the openstack compute book says soo.10:07
koolhead11ok. go tit10:07
sorenkoolhead11: Where?10:07
koolhead11soren, am editing wikipedia entry10:09
*** reldan has quit IRC10:11
*** aimon has quit IRC10:15
*** uksysadmin has quit IRC10:28
*** naehring has quit IRC10:39
uvirtbotNew bug: #739283 in nova "Nova API service stuck in endless loop if image_service=nova.image.s3.S3ImageService" [Undecided,New]
koolhead11soren,  is the code names  nova swift and Glance version ]dependant ?10:43
sorenkoolhead11: What do you mean?10:44
koolhead11soren, i wanted to know the code name for compute object storage will always remain same? it has notthing to do with newer releases?10:45
sorenkoolhead11: Oh, no, the code names should be the same.10:46
sorenkoolhead11: We have other code names for the versions.10:46
koolhead11soren, thanks. yeah bexer and cactus coming :)10:47
sorenkoolhead11: 2010.1 = Austin..10:47
sorenRight :)10:47
sorenBexar, not Bexer, by the way.10:47
koolhead11soren, sorry typo10:47
koolhead11all this while i was in sense that OVI is one main reason for the popularity of open-stack besides rackspace and nasa along with many other companies10:48
*** reldan has joined #openstack10:51
*** ryker has joined #openstack10:54
*** uksysadmin has joined #openstack10:58
*** zul has joined #openstack11:01
*** ovidwu has quit IRC11:02
*** Hexadecimal has joined #openstack11:19
*** bkkrw has joined #openstack11:19
koolhead11soren, by default i see 3 roles. how will i know more about the permission these roles have associated with it? Can i create a custom role ?11:20
*** urchit has joined #openstack11:21
koolhead11Is there a role via which i can create some uses in my specific project and assign them certain roles?11:21
urchitdoes anyone know what needs to be done to remove a node from a multi node install other than killing nova-compute?11:22
*** ctennis has quit IRC11:23
sorenurchit: That should suffice.11:23
sorenkoolhead11: I'm having trouble following your questions.11:24
*** bkkrw has quit IRC11:24
urchitsoren: thanks11:24
koolhead11soren, in my open-dashboard manage section i see 3 roles available.  "sysadmin"  "netadmin"  "developer"  my question is 1. can i add custom roles? 2. can i add users to my project  and give  one of these roles.11:28
soren1. No, I don't believe you can. 2. Yes.11:28
kal24i have the following error in nova-network.log11:37
kal24does anyone know the reason?11:37
*** jfluhmann has quit IRC11:41
*** miclorb_ has quit IRC11:43
*** omidhdl has quit IRC11:45
*** nRy has joined #openstack11:45
*** zul has quit IRC11:46
*** 15SAACGVK has quit IRC11:46
*** rchavik has joined #openstack11:49
*** ctennis has joined #openstack11:50
*** ctennis has joined #openstack11:50
BK_mankal24: yep. you need to fix your sudo configuration11:51
BK_mankal24: Defaults:nova !requiretty11:51
*** reldan has quit IRC12:07
*** adiantum has joined #openstack12:08
*** pygi has quit IRC12:08
*** Hexadecimal has quit IRC12:09
*** westmaas1 has quit IRC12:09
*** adiantum has quit IRC12:13
*** aliguori has joined #openstack12:16
*** jeremyb has quit IRC12:16
*** jeremyb has joined #openstack12:17
*** gasbakid has quit IRC12:19
*** ramkrsna has quit IRC12:21
*** ramkrsna has joined #openstack12:21
uvirtbotNew bug: #739327 in glance "glance-combined still referenced in rev92" [Undecided,New]
*** chuck_ has joined #openstack12:31
*** chuck_ is now known as zul12:32
*** zul has quit IRC12:32
*** zul has joined #openstack12:32
*** ryker has quit IRC12:33
*** ryker has joined #openstack12:33
*** nelson has joined #openstack12:38
*** B2_ has joined #openstack12:48
*** jfluhmann has joined #openstack12:48
*** omidhdl has joined #openstack12:50
*** jeremyb has quit IRC12:53
*** jeremyb has joined #openstack12:53
*** tildedave has joined #openstack12:55
*** h0cin has joined #openstack12:56
*** h0cin has joined #openstack12:56
*** zigo has joined #openstack12:57
*** sophiap has joined #openstack12:58
kal24BK_man, i have changed the sudoers file, rebooted, and still get the same message12:58
zigoHi there.12:58
kal24i have put in the first line of  /etc/sudoers.d/nova_sudoers12:59
zigoI'm a Debian developer, I'm willing to work on packaging nova in Debian.12:59
zigoI tried to build from the current Natty package, but got this error:12:59
zigoAre some build-depends missing???12:59
*** tildedave has quit IRC12:59
*** tildedave has joined #openstack13:00
*** B2_ has quit IRC13:03
*** ptremblett has joined #openstack13:03
*** kapil has joined #openstack13:05
*** hadrian has joined #openstack13:07
*** dendro-afk is now known as dendrobates13:07
*** jero has joined #openstack13:11
sorenzigo: We need sphinx > 1.0.13:13
sorenzigo: Maybe the build-depends isn't versioned :(13:13
sorenzigo: /me fixes13:13
zigoOh ok, so that's what I said: build-depends are wrong.13:13
zigoI'd be happy to fix all of these! :)13:13
*** omidhdl has quit IRC13:13
zigoIs the packaging of nova in bzr as well?13:13
zigoIs bzr close to git? :)13:14
zigo(I never used bzr, only git...)13:14
*** joloughlin has joined #openstack13:15
sorenThey're both DVCS's.13:15
sorenBoth let you commit things and stuff.13:15
sorenTheir metadata model is a bit different, though.13:16
soren..but for most intents and purposes, they do the same thing.13:16
joloughlinhi, which project roles does a user need to have in order to launch an instance ?13:16
zigoOk, thanks for the advice.13:16
zigosoren: So you are the person doing the Ubuntu packaging?13:16
sorenMostly, yes.13:16
zigoNice to meet you then.13:16
zigoMy qa page.13:17
zigoI intend to work on packaging nova in Debian ...13:17
sorenNo Python packages?13:17
zigoSo, do you mind doing a bit of team work with me? :)13:17
zigoWell, final goal is having the full of openstack to run in Debian.13:18
sorenNot at all.13:18
zigoWhere do you live?13:18
zigo(I'm a French guy living in China)13:18
sorenShouldn't be hard. I've been meaning to put out a call for help for that.13:18
sorenI'm in Denmark.13:18
zigoI'm asking because of timezones issues ...13:18
*** kashyap has quit IRC13:18
sorenIf this is off-hours stuff for you, we should match up just fine.13:19
*** timrc has quit IRC13:20
joloughlinim looking for some advice on having multiple users in one project13:20
zigoI'll need to dig out how all this work, so it might need a lot of time before I can really help.13:21
joloughlinwhen i create a user and add to an existing project with the developer role they can't start instances13:21
zigoBut I'm quite sure it will happen.13:21
joloughlinshould they be able to ?13:21
zigoPeople here in Shanghai are really motivated for doing cloud stuff.13:21
*** kashyap has joined #openstack13:22
*** zigo has quit IRC13:22
*** zigo has joined #openstack13:23
*** zigo is now known as zigo-_-13:23
zigo-_-[21:21:06] <zigo> I'll need to dig out how all this work, so it might need a lot of time before I can really help. But I'm quite sure it will happen. People here in Shanghai are really motivated for doing cloud stuff.13:23
zigo-_-soren: Seems the package also needs python-rrdtool in the build-depends13:24
*** bkkrw has joined #openstack13:25
joloughlinim looking for some advice on having multiple users in one project13:26
joloughlincan somebody help ?13:26
*** mjmac has left #openstack13:27
*** patcoll has joined #openstack13:27
*** bcwaldon has joined #openstack13:28
*** joloughlin has quit IRC13:29
zigo-_-soren: You also should add a rm -rf doc/doc/build in your clean target !13:30
*** urchit has quit IRC13:31
*** drico_ has quit IRC13:31
*** joloughlin has joined #openstack13:33
*** Ephur has joined #openstack13:33
zigo-_-Ok, it built in my Squeeze! :P13:36
joloughlinhi, can anyone give some advice on roles and users within projects ?13:37
*** reldan has joined #openstack13:45
*** jaypipes has quit IRC13:47
*** pygi has joined #openstack13:49
*** zenmatt has joined #openstack13:49
*** jaypipes has joined #openstack13:51
jaypipessoren: around? time to tell me your packaging secrets?13:51
*** da_ has joined #openstack13:53
*** da_ has left #openstack13:53
*** j05h has joined #openstack13:55
kal24i have this error:13:59
kal24could anyone help pls?13:59
*** urchit has joined #openstack13:59
sorenzigo-_-: Why would rrdtool be needed in build-depends?13:59
zigo-_-I had some errors about the python module not being present... doens't mater, ignore that one.14:00
zigo-_-But don't ignore the one about the rm -rf doc/doc/build ! :)14:00
*** evil_e has joined #openstack14:02
zigo-_-soren: There's some dependencies for upstart that wont fit with Debian.14:03
zigo-_-The following packages have unmet dependencies:14:04
zigo-_- upstart : Conflicts: sysvinit14:04
sorenI don't see how that's something we can fix.14:06
sorenWe don't require upstart.14:06
soren...and if we did, if upstart doesn't work on Debian, that's not an OpenSTack problem.14:07
sorenThat said, I'm sure our init (i.e. non-upstart) scripts need to be updated.14:07
zigo-_-soren: What is upstart-job for?14:07
zigo-_-Well, Debian uses insserv ...14:08
BK_manzigo-_-: does Debian have an upstart?14:08
BK_manzigo-_-: welcome, a packager for another distro :)14:09
zigo-_-BK_man: We thought about using it, but finally, we use insserv !14:09
jaypipesBK_man: damn, beat me to it! :) Thanks!14:09
BK_manzigo-_-: ok. you need to write initscripts from scratch14:09
BK_manzigo-_-: I'm packager for RHEL6. I'm based in Russia14:09
zigo-_-And frankly, Ubuntu has taken the job initiated in Debian in 2002 to do dependency boots, so it's really a shame that we don't use the same stuff...14:10
zigo-_-Cool! :)14:10
BK_manzigo-_-: you can leech useful sources for packaging here and adapt them for Debian:
jaypipeskal24: what is your sql_connection setting?14:11
sorenzigo-_-: upstart-job is a wrapper for stuff that moved from a sysvinit style boot script to an upstart job.14:11
sorenjaypipes: Ok, so the problem with the glance builds is that they still refer to something called glance-combined.14:11
zigo-_-Where's the init.d script that needs to be re-written then?14:11
sorenzigo-_-: In the source package.14:12
sorenzigo-_-: Where you'd expect them (usual debian style)14:12
zigo-_-debian/*.init ?14:12
zigo-_-All of them?14:12
zigo-_-k :)14:12
zigo-_-I didn't see anything that would need upstart there ...14:12
zigo-_-insserv is fine! :)14:12
jaypipessoren: hehe, ok. BK_man just submitted a fix for that. heading off to Tarmac.14:13
sorenDid Debian drop going to upstart?14:13
jaypipessoren: any way to automate testing of that?14:13
BK_manjaypipes: soren14:13
sorenjaypipes: Packaging?14:13
jaypipessoren: and fail the tarmac job if the build fails?14:13
zigo-_-soren: Again, we don't use upstart, we use insserv instead.14:13
sorenjaypipes: Hm... Yeah, I suppose.14:13
BK_manjaypipes: soren: can we use Jenkings for packaging automation?14:14
sorenzigo-_-: They're hardly equivalent.14:14
sorenHow do you figure that?14:14
zigo-_-Why hardly equivalent?14:14
BK_manzigo-_-: you need like that:
soreninsserv is still does old school sysvinit style booting.14:15
zigo-_-soren: Nop, it does dependency booting.14:15
sorenupstart is event driven.14:15
zigo-_-We don't use the links in /etc/rcX.d anymore since Squeeze (eg: since last 4th of february).14:16
jaypipesBK_man: we do use it for package automation. problem is that the build automation doesn't prevent a breaking merge from being merged :)14:16
BK_manjaypipes: agreed. do you want to break a merge if my (RHEL6) build failed? I think not.14:16
zigo-_-BK_man: What's the git clone URL then? (I'm not use to github...)14:17
*** dendrobates is now known as dendro-afk14:17
BK_manjaypipes: Ubuntu is not only distibution for OpenStack. It's primary, but OpenStack should be disto-agnostic14:17
BK_manzigo-_-: I'm not a git guru, sorry. Try to ask reldan or adiantum14:18
j05hBK_man:  git clone
BK_manj05h: thanks!14:19
zigo-_-I was about to try that! :)14:19
j05hnp. i can help with git much more than i can bzr :D14:21
jaypipesBK_man: actually, yes, I do want the merge to be rejected.14:21
jaypipesBK_man: that's exactly what I'm saying.. if we can automate this stuff so that we don't break distro builds, I think that's A Good Thing. Agree?14:22
BK_manjaypipes: not sure14:23
sorenNot entirely.14:23
BK_manjaypipes: I have a local patches (RHEL specific). I do not want to reject a merge if one of my patches will fail to apply.14:23
jaypipessoren: well, in this particular case, the Debian/Ubuntu build broke, so baby step would be to just at least have Tarmac fail the merge if that particular build fails..14:23
sorenI'd be happy for us to be notified if something breaks, but I think setting ourselves up so that any bug in any of the distros that have chosen to integrate OpenStack can block us from merging stuff is a bad plan.14:23
BK_mansoren: +114:24
sorenjaypipes: Oh, I agree the Ubuntu builds should be required to work.14:24
sorenjaypipes: Anything else shouldn't block us, IMO.14:24
jaypipessoren, BK_man: just as an example, Drizzle denies any merge that breaks a large number of target platforms: Not *all* those platforms trigger a complete failure, but a number of them do...14:25
sorenjaypipes: One of the reasons for having a reference platform is that we won't let more than one platform slow us down.14:25
jaypipessoren: no, don't get me wrong, I understand the reasoning. I would argue a couple more "target" or "important" platforms would be useful, too.14:26
*** mustfg2 has joined #openstack14:26
BK_manjaypipes: can we vote for second reference platform on upcoming summit?14:27
jaypipesBK_man: sure, that would be cool.14:27
jaypipesBK_man, soren: I'm not arguing for having RH as a target platform at this very moment...14:27
* BK_man hoping to be on Summit. Airfare is HUGE money :(14:27
jaypipesBK_man: yeah, I know :(14:28
sorenI, for one, have no problem at all having just a single target platform. I like to be able to give clear answers when people ask which platform to run our stuff on.14:28
jaypipesBK_man: hopefully, the next one may be in Europe...14:28
jaypipessoren: but we are supposed to be "platform-agnostic", no?14:28
sorenI wish there was a similarly clear answer as to which hypervisor to use, for instance.14:28
jaypipessoren: or is that just for the virt layer?14:28
BK_manjaypipes: it will be another nighmare with obtaining Schengen visas14:28
sorenjaypipes: I don't think we've ever said we were agnostic.14:29
BK_mansoren: it would be nice not to use distro-specific features (like NBD and SELinux)14:30
sorenjaypipes: My personal opinion is that I'd be happy to support more, but I won't myself put resources into supporting them. If people want to provide patches that make us support another distro, I won't reject them "because they don't matter for Ubuntu" or something.14:30
sorenBK_man: How is NBD distro-specific?14:30
BK_mansoren: NBD is not in RHEL anymore. I don't think that it will be easily intergated to upcoming CentOS614:31
mustfg2when i want to install an additional computing node for nova14:31
sorenBK_man: *sigh*14:31
BK_mansoren: there were important reasons for dropping NBD from RHEL kernel - I saw some complicated bugs in bugzilla14:32
mustfg2can i do that with the multible server script?14:32
BK_manmustfg2: CC stands for Cloud Controller. Have no idea what that script does however :)14:32
jaypipesBK_man: well, as luck would have it, your patch fixed one thing, but there's still a (separate) bug that is preventing the Glance build from succeeding...14:32
sorenjaypipes: Build log link?14:33
sorenzigo-_-: Oh, you're looking at the binary packages, aren't you?14:33
zigo-_-Of what?14:33
jaypipessoren: known bug, AFAICT:
uvirtbotLaunchpad bug 731684 in glance "Glance lazy loading issue -- still coupled with Swift" [High,Confirmed]14:33
zigo-_-I'm building from your .dsc file ...14:33
sorenzigo-_-: Whatever. You mentioned upstart-job. You would only ever see that dependency if you're looking at binary packages.14:34
zigo-_-And I don't even get how nova-api got the upstart-job dependency in, as it's not in your debian/control ...14:34
BK_manjaypipes: mine builded ok: Wrote: /home/abr/rpmbuild/RPMS/noarch/openstack-glance-2011.2-0.9.bzr93.noarch.rpm14:34
zigo-_-So, how it ended in the deps is a mistery ... :)14:34
sorenzigo-_-: As I said: binary packages. dh_installinit adds it (through ${misc:Depends})14:34
jaypipesBK_man: cool, that's good at least! :)14:34
zigo-_-I don't think it would in Debian.14:35
sorenjaypipes: There's a bunch of other test suite errors that don't mention Swift. Are they just masking the swift problem?14:35
sorenzigo-_-: Of course not.14:35
sorenzigo-_-: Hence, you must be looking at my binary packages. Don't :)14:35
zigo-_-I'm not using any binary directly from Ubuntu, I've dget your .dsc and build them ...14:35
zigo-_-And it ended-up having this upstart-job depends, I don't even understand how !14:36
sorenzigo-_-: So your dh_installinit must have upstart support, too.14:36
jaypipessoren: ya, same root problem I believe.14:37
zigo-_-Seems so, yes.14:37
*** adiantum has joined #openstack14:37
zigo-_-Oh, I get it now ...14:37
zigo-_-It's because there's some debian/*.upstart files !14:38
*** rnirmal has joined #openstack14:38
zigo-_-If I remove them and just use dh_installinit, it's gona be ok, I guess.14:38
*** hub_cap has joined #openstack14:40
*** urchit has quit IRC14:42
*** urchit has joined #openstack14:43
zigo-_-Tomorrow, I'll rewrite the debian/rules so it doesn't use the minimalist dh_helper style ... :)14:43
*** rchavik has quit IRC14:44
zigo-_-soren: bingo, it does work now.14:46
zigo-_-(if removing the debian/*.upstart files)14:46
zigo-_-but nova-api doesn't start, it seems stuck ...14:47
sorenzigo-_-: Uh... why would you do that?14:47
sorenzigo-_-: Rewrite debian/rules, I mean.14:47
*** romain_lenglet has joined #openstack14:47
zigo-_-Because it has some useless calls that are slowing down the build.14:47
zigo-_-That's always the result of using the things with14:48
zigo-_-        dh $@14:48
zigo-_-Seems nova-api doesn't go in the background ...14:50
sorenIt doesn't.14:50
sorenYou can probably use start-stop-daemon14:50
BK_mansoren: but it can't track if background process was successfully launched or not14:51
zigo-_-BK_man: start-stop-daemon will do that for me.14:51
zigo-_-Anyway, it's late, I'll see that tomorrow.14:51
BK_manzigo-_-: really? Did you tried it? Mine did!14:51
joloughlinhi, when a user is added to a project, what role must they be assigned so they can run instances ?14:51
BK_manzigo-_-: /sbin/start-stop-daemon --start -b -c nova:nobody --make-pidfile --pidfile $pidfile  --exec /usr/bin/nova-$suffix -- --flagfile=$flgfile --logfile=$logfile --pidfile $pidfile14:52
BK_manzigo-_-: this is how I spawn nova-compute14:52
zigo-_-Cool !14:52
BK_manzigo-_-: please note -b switch14:52
sorenjoloughlin: projectmanager or sysadmin.14:52
zigo-_-I didn't know redhat had start-stop-daemon ...14:52
zigo-_-I thought it was a Debian thing.14:53
BK_manzigo-_-: with that start-stop-daemon don't want child process14:53
BK_manzigo-_-: it's not RHEL this14:53
BK_manzigo-_-: I downloaded it and packaged in RPM.14:53
*** romain_lenglet has left #openstack14:54
BK_manzigo-_-: RHEL's initscripts can't put daemons to background14:54
zigo-_-BK_man: Do you work using Fedora?14:54
joloughlinsoren: thanks, was trying developer14:54
zigo-_-What's your distro?14:54
BK_manzigo-_-: nope. My build is RHEL6.0 at the moment14:54
zigo-_-But there's no commit in RHEL ...14:55
zigo-_-I mean, the only way to get into RedHat is to package for Fedora.14:55
zigo-_-Anyway, I'm happy to find nice people in this project!14:55
zigo-_-You'll see me again.14:55
BK_manzigo-_-: :)14:56
zigo-_-But it's late here now, bye !14:56
BK_manzigo-_-: bb14:56
*** sophiap has quit IRC14:56
*** zigo-_- has quit IRC14:56
*** Shentonfreude has joined #openstack14:59
*** omidhdl has joined #openstack15:01
*** dragondm has joined #openstack15:01
*** sophiap has joined #openstack15:02
*** pvo has joined #openstack15:03
*** koolhead11 is now known as koolhead11|15:05
*** koolhead11| is now known as koolhead11|afk15:05
uvirtbotNew bug: #739433 in glance "Glance Image Limit is 2^32 on 32bit Systems" [Undecided,New]
*** troytoman-away is now known as troytoman15:10
*** guigui has joined #openstack15:13
*** mray has joined #openstack15:15
*** rds__ has joined #openstack15:17
*** omidhdl has quit IRC15:24
*** kashyap has quit IRC15:24
*** dprince has joined #openstack15:32
*** spectorclan_ has joined #openstack15:33
*** ianweller_ has joined #openstack15:34
joloughlinwhen I try to launch an instance with a user having a role other than projectmanager i get a 40115:37
joloughlini've tried sysadmin, netadmin and developer15:37
spectorclan_If anyone has issues with the voting system please feel free to contact me. Thanks,15:37
mustfg2when i have a 2 server setup (on one server are all services on the second there is only nova-compute installed)15:40
mustfg2can i specify which nova-compute server is used15:40
mustfg2when im running an instance15:41
mustfg2just for testing :)15:42
*** lvaughn has quit IRC15:44
*** burris has quit IRC15:44
*** lvaughn has joined #openstack15:44
jarrodmustfg2: you mean specify which nova-compute instance to use when creating a vm from the nova-api?15:45
jarrodyou can disable the service on other nova-computes temp15:45
mustfg2yep thats what i thought :D15:45
joloughlincurrently on a user with a projectmanager role can launch an instance15:46
joloughlinis this correct ?15:47
*** burris has joined #openstack15:50
*** urchit has quit IRC15:53
*** urchit has joined #openstack15:54
*** ramkrsna has quit IRC15:54
spectorclan_Election - Some Candidate Bios at  before you vote; Some candidates listed info there.15:54
*** urchit has quit IRC15:56
justinsbThanks spectorclan_ - good idea.  I'm standing for the PPB and wrote a mini-essay to make up for the idea that I might not be as well known as the Rackspace candidates15:57
justinsbHopefully other non-Rackers will do the same!15:57
mustfg2how can i check on which nova-compute my instance is running ?16:02
*** gregp76 has joined #openstack16:02
*** bcwaldon has quit IRC16:05
spectorclan_pvo: are you around?16:07
*** enigma1 has joined #openstack16:10
*** joloughlin has quit IRC16:11
*** clayg_ has joined #openstack16:11
mustfg2how can i check on which nova-compute server my instance is running ?16:13
*** guigui has quit IRC16:15
*** johnpur has joined #openstack16:19
*** ChanServ sets mode: +v johnpur16:19
*** bkkrw has quit IRC16:24
*** arun_ has joined #openstack16:24
*** Ryan_Lane has joined #openstack16:26
*** urchit has joined #openstack16:29
*** bcwaldon has joined #openstack16:32
jaypipesspectorclan: if I don't know the candidates in the list, I voted them an 11. I'm not voting anyone to the PPB that doesn't contribute to OpenStack. Period.16:34
jaypipesjustinsb: and no, I didn't vote you an 11. ;)16:34
justinsbjaypipes: I like that our voting goes up to 11.  Excellent potential for Spinal Tap jokes16:36
justinsbjaypipes: I hope everyone puts up bios.  I don't even know how to choose between some of the higher profile candidates!16:37
spectorclan_jaypipes: I agree, just nice to see profiles of people16:37
justinsbjaypipes: Did you see my comment about the SA-migrate issues?16:38
jaypipesjustinsb: yeah. I'm very frustrated with the whole thing right now. :)(16:39
justinsbjaypipes: I'm not a big fan of SA migate either16:39
justinsbjaypipes: But liked it a lot better once I figured out reflection on tables16:39
justinsbjaypipes: Which made me think my beef was mostly with the docs16:39
jaypipesjustinsb: I wasted an absolute shit-ton of time this weekend with it.16:39
jaypipesjustinsb: reflection on tables? that's just SA, not SA-migrate...16:40
justinsbjaypipes: I know ... it's not like what we're doing is that hard to do in raw SQL (and probably safer as well)16:40
justinsbjaypipes: Well, it may have got it from SA, but I'm thinking big picture here16:40
jaypipesjustinsb: in Nova or Glance? Because Nova's scripts aren't even tested. There's bugs all over them, and data loss bugs at that.16:40
justinsbjaypipes: Well that's great news :-)16:41
jaypipesjustinsb: in Glance migrate scripts, you don't just drop a column and move it to another table. You first have to make sure the data in the column you are dropping is saved to a temp table, and then re-inserted into the other table. That isn't done anywhere in Nova's scripts, AFAICT.16:41
*** arun_ has quit IRC16:41
justinsbjaypipes: Have we moved any columns?16:42
jaypipesjustinsb: and there's zero testing of any of the migrate stuff.16:42
jaypipesjustinsb: look at the migrate scripts from NTT. I know that certain columns were moved from Network to FixedIP at the very least.16:42
justinsbjaypipes: Perhaps we should have a database backup command as the first command of all our migrate scripts as part of the release16:42
*** CloudChris has quit IRC16:43
jaypipesjustinsb: I spent about 2 hours looking through sa-migrate's test cases yesterday. they don't test barely anything.16:43
justinsbjaypipes: I have no objection to doing things in raw SQL, but isn't it a relatively big change at this stage of the game?16:44
justinsbjaypipes: One thing we _could_ do is make creating a migrate script part of the release process16:44
jaypipesjustinsb: sa-migrate is *supposed* to work with just straight SQL scripts...16:44
justinsbjaypipes: And that could be optimized, hand-checked SQL16:44
justinsbjaypipes: If SA-migrate is as bad as you say it is, it's a bit scary to rely on it in the real world16:45
justinsbjaypipes: Where there will be different versions of it etc16:45
justinsbjaypipes: The best approach I've ever seen to DB migration just works by table inspection16:45
jaypipesjustinsb: re: "a little late at this stage of the game". a) we're talking only about Glance here, not Nova, and b) I've got branches and bugs sitting in Work In Progress *just* because of this issue.16:45
jaypipesjustinsb: it's holding up everything in Glance.16:46
justinsbjaypipes: Ick16:46
*** spectorclan_ has quit IRC16:46
jaypipesincluding Critical blueprints because the migrate scripts won't pass the unit test because of bugs in SA-Migrate.16:46
justinsbjaypipes: I think the beauty of the 'unified' migrate script is that we can develop in the "more productive" SQLAlchemy...16:46
justinsbjaypipes: ... or you can write your own SQL script16:47
justinsbjaypipes: But when it comes to production we use a single script16:47
jaypipesyou need progressive, not unified, scripts, otherwise trunk will break existing dev setups every time a DB schema change is made.16:47
jaypipesjustinsb: off to meeting, back laterz.16:48
justinsbjaypipes: ttyl16:48
*** skiold has quit IRC16:48
*** arun_ has joined #openstack16:51
*** hub_cap has quit IRC16:51
*** hub_cap has joined #openstack16:51
*** mustfg2 has quit IRC16:52
*** maplebed has joined #openstack16:56
*** Nacx has quit IRC17:00
*** joearnold has joined #openstack17:01
kal24i have no error message in logs17:04
kal24but my VM stuck in pending state17:04
kal24if i try to reboot, then it remains in shutdown state17:05
kal24i have tried kvm directly, it works, and can images uploaded17:05
kal24i am completeley clueless17:05
kal24could u help?17:05
*** kakoni has joined #openstack17:09
zulspectorclan: ping17:10
*** shawn has joined #openstack17:12
*** spartha has joined #openstack17:13
*** Ryan_Lane has quit IRC17:13
sparthaHas anyone found a fix to this bug?
uvirtbotLaunchpad bug 714260 in nova "creating of new instance fails while using objectstore for images (curl exit code 22)" [Undecided,Incomplete]17:13
*** Ryan_Lane has joined #openstack17:13
*** urchit has quit IRC17:14
sparthaHas anyone found a fix for this bug?
kal24please help, i have no error message, but my vm are pending17:15
kal24what to do?17:15
mtaylorpvo: hey! sorry I missed you the other day - whazzup?17:19
kal24pleasee help me17:22
kal24i cant get my vm's out of pending state17:22
*** aryan_ has quit IRC17:23
j05hkal24: are you vms getting network?17:24
kal24they get ip17:24
vishykal24: do you have any errors in the log at /var/log/libvirt/qemu/instance-XXXXXX.log17:24
kal24but cant ssh or ping them, i used the euca-authorize command17:24
vishypending means libvirt failed to start them17:24
vishythere should also be an error in nova-compute.log17:25
kal24nothing in libvirt log17:25
kal24and nothing in nova log17:26
*** kashyap has joined #openstack17:26
kal24i have tested libvirt seperatly17:28
kal24i could access vms with virsh17:29
BK_mankal24: is there anything in nova logs? cd /var/log/nova; ls -otrh17:29
*** MotoMilind has joined #openstack17:30
*** dspano has joined #openstack17:32
*** sophiap has quit IRC17:33
kal24it can caused maybe missing firewall rule?17:35
kal24i have added a rule like this one: iptables -t nat -A PREROUTING -d -p tcp -m tcp --dport 80 -j DNAT --to-destination
kal24but did not change more on rules17:36
*** enigma1 has quit IRC17:38
BK_mankal24: can you check console output from KVM?17:39
*** sophiap has joined #openstack17:39
BK_mankal24: /var/lib/nova/instances/instance-0000000e/console.log17:39
* BK_man stupid - it does not starting up. sorry17:40
*** enigma1 has joined #openstack17:40
*** johnpur has quit IRC17:40
kal24i cant see the vm at all with virsh17:40
kal24oh, i check the link17:41
kal24there is nothing here  /var/lib/nova/instances/17:41
kal24oh, another useless day, with struggling :( if u have any idea or solution for pending VM problem, dont hesitate to drop me a mail onlany@freemail.hu17:44
kal24good night guys17:44
*** kal24 has quit IRC17:44
*** urchit has joined #openstack17:44
*** nelson has quit IRC17:45
*** jc_smith has joined #openstack17:45
uvirtbotNew bug: #739546 in openstack-manuals "image_service default flag is incorrect" [Undecided,New]
*** dendro-afk is now known as dendrobates17:47
*** uksysadmin has quit IRC17:47
*** urchit has left #openstack17:51
*** urchit has joined #openstack17:52
*** dfg has joined #openstack17:52
*** zul has quit IRC17:54
*** johnpur has joined #openstack17:56
*** ChanServ sets mode: +v johnpur17:56
*** chuck_ has joined #openstack17:57
*** jc_smith has left #openstack17:57
*** chuck_ is now known as zul17:57
openstackjenkinsProject nova build #674: SUCCESS in 2 min 5 sec:
openstackjenkinsTarmac: Changed fixed_range (CIDR) to be required in the nova-manage command; changed default num_networks to 1.17:58
*** clauden_ has joined #openstack18:05
pvomtaylor: oh hey. I had some dpkg questions but I think I got most of them answered. I may ping you about packing some other debs18:08
zulso about the would be nice if there are some bio on the wiki so i can be better informed who im voting for18:09
pvonot everyone wil have a bio, since it is all voluntary18:10
openstackjenkinsProject nova build #675: SUCCESS in 2 min 7 sec:
openstackjenkinsTarmac: Openstack api 1.0 flavors resource now implemented to match the spec18:13
*** zul has quit IRC18:14
*** chuck_ has joined #openstack18:14
*** chuck_ is now known as zul18:14
clayg_vishy: the quickstart tty.tgz images aren't compatible with cactus?
uvirtbotLaunchpad bug 735641 in nova "Unclear error when images directory wasn't upgraded" [Low,In progress]18:27
*** zul has quit IRC18:30
vishyclayg_: that is correct, I haven't had time to upgrade them yet, but you should be able to put them in and then do a nova-manage image convert, perhaps I'll add that to for now18:30
*** chuck_ has joined #openstack18:30
*** chuck_ is now known as zul18:30
justinsbpvo: I don't think that's very helpful, considering we were directed to add bios here:
justinsbzul: Bios are supposed to be here:  Additional bios are available on pvo's link18:31
*** sophiap has quit IRC18:36
*** nelson has joined #openstack18:39
clayg_vishy: ok, nova-manage image convert <path_to_old_images> worked ok, didn't know about that one18:40
vishygood deal.  I'll add it to nova.sh18:40
clayg_but the names change... so tutorials can't just refer to "ami-tty" anymore...18:41
uvirtbotNew bug: #739601 in nova "Instance files should be tombstoned and cleaned up later instead of immediately deleted" [Undecided,New]
vishyclayg_: you can, there is a little hack to allow you to use names instead of ids18:42
vishynote the (ami-tty) in parentheses18:42
clayg_oic, ok nice18:42
vishyclayg_: can you do a quick test for me?18:43
vishyclayg_: what happens if you run image convert again after they are already converted?18:43
jaypipesdubs: got a spare 15 minutes? mind looking at
clayg_vishy: nothing bad - it doesn't work:
notmynamespectorclan: next time there is an election, can we have IRC names on the ballot too (next to the real names)?18:44
justinsbnotmyname: Great idea - I just added mine to the official bio page, though it's probably easy to guess mine, because mine is my name :-)
vishydoes it give an error message?18:46
vishyah i see the paste18:46
vishycool i'll just run it every time then18:46
creihtI wonder how many people on that page don't realize they have been nominated :)18:46
clayg_vishy: maybe worth noting - you can't convert the old images in place... I had to move them to old_images, then mkdir images, then convert old_images18:47
*** jero has quit IRC18:47
justinsbcreiht: Perhaps we should create a really awful position just to test that theory :-)18:47
*** jero has joined #openstack18:48
creihtspectorclan: It also may useful for the 2 general board seats to indicate which 2 seats they will be replacing18:49
*** lionel_ has quit IRC18:50
uvirtbotNew bug: #739609 in nova "Cannot set body to unicode without a charset errors" [Undecided,New]
*** CloudChris has joined #openstack18:51
*** gasbakid has joined #openstack18:52
openstackjenkinsProject nova build #676: SUCCESS in 2 min 6 sec:
openstackjenkinsTarmac: Provide more useful exception messages when unable to load the virtual driver.18:53
*** chuck_ has joined #openstack18:54
*** lionel_ has joined #openstack18:54
*** zul has quit IRC18:55
*** chuck_ is now known as zul18:55
*** zul has joined #openstack18:55
vishyclayg_: hmm, it is supposed to do that if the dirs match18:58
vishyclayg_: maybe the symlink confuses it?18:58
vishyclayg_: how did you get that last error that you pasted?19:00
*** vjsamuel has joined #openstack19:00
*** vjsamuel has left #openstack19:03
*** j05h has quit IRC19:04
*** burris has quit IRC19:05
uvirtbotNew bug: #739629 in glance " needs --unittests-only option" [Low,New]
*** fabiand_ has joined #openstack19:07
*** dprince has quit IRC19:08
*** spectorclan_ has joined #openstack19:13
*** burris has joined #openstack19:16
jaypipessoren: around?19:17
kpepple_clayg: i see your error with images ... i think i fixed that in trunk this weekend ... are you still seeing problems ?19:19
*** j05h has joined #openstack19:19
*** kashyap has quit IRC19:21
*** burris has quit IRC19:21
*** Ryan_Lane has quit IRC19:21
uvirtbotNew bug: #739641 in nova "Smoketests failing for ipv6" [Undecided,In progress]
*** Ryan_Lane has joined #openstack19:21
*** omidhdl has joined #openstack19:23
*** urchit has quit IRC19:25
*** burris has joined #openstack19:25
btorchwhat do I need to do to tell nova to use openstack api and not the ec2 one ?19:26
*** mgoldmann has quit IRC19:28
jaypipesmtaylor: around? I think I need help...19:31
bcwaldonbtorch: both endpoints should be available19:32
jaypipesbtorch: just query Nova on the openstack API point (8774 I believe?)19:32
*** kazu has joined #openstack19:33
*** urchit has joined #openstack19:34
*** dprince has joined #openstack19:36
jaypipesis anyone at all around that has access to the Jenkins box?19:36
*** jc_smith has joined #openstack19:40
*** jc_smith has left #openstack19:40
*** adiantum has quit IRC19:43
*** kazu has quit IRC19:43
*** reldan has quit IRC19:46
*** joearnold has quit IRC19:47
justinsbjaypipes: Suppose I (or someone else) fixed the index=True SQLite bug with SA-Migrate.  Would you then be OK with SA Migrate, or are there deeper issues?19:49
*** topper has joined #openstack19:53
topperheyo everybody19:53
vishyjustinsb, jaypipes: I think we should discuss whether a) downgrade is important at all b) the best way to test migrations19:53
vishyI'm having trouble finding a downgrade use-case that makes it worth our time to actually make downgrading work19:54
justinsbvishy: I think downgrade is important in theory, but in practice I suspect that unless it is automated it will not get as much "focus" as the upgrade script19:54
jaypipesjustinsb, vishy: testing that upgrades don't lose data would be enough for me. I've had it with SA-migrate. Wasted a ton of my time over the past couple weeks trying to get it to work with tests.19:54
edayvishy: production sites may try an upgrade, see things break, and need to downgrade until issues are resolved19:55
vishyeday: why can't they just backup restore?19:55
edayvishy: I've had to downgrade a number of times when things went wrong after an upgrade :)19:55
justinsbvishy: Presumably because they don't want to rollback on a live system19:55
vishyeday: the one use-case I can think of is that they upgrade production, use it for two weeks, see that it is broken, then try to rollback19:56
edayvishy: you could, but that's a bit heavy for a downgrade.19:56
justinsbI think if that ever happened with a release, we would _really_ have screwed up19:56
vishyeday: i think that is rare enough that I'm ok saying, listen we don't support downgrade so don't do it.19:56
jaypipesvishy: like I said, we don't even test upgrades in Nova... that would be a good first step.19:57
vishyin talking with people running large web systems with db migrations (in ruby for example), no one ever really uses downgrades19:57
vishyjaypipes: agreed, how exactly do we test these things though?19:57
mtaylorjaypipes: what's up bro?19:57
vishyjaypipes: throw some sample data into the old version, do the upgrade, and make sure that the new data makes sense?19:58
edayvishy: I would have to disagree, seamless downgrades have been very important for production systems im my experience, but I'll defer to the group :)19:58
jaypipesvishy: yep, that's pretty much what our test case does...19:58
vishyjaypipes: and what happens to these tests if we roll-up all of the migrations into one script after each release?19:58
jaypipesvishy: not sure. what do we do as far as the migration process in that case anyway?19:59
vishyjaypipes: the discussion was to take all of the changes done in the release and roll them up into one upgrade, I guess we could port any usable tests at that time?20:00
jaypipesvishy: to be honest, my post was really about Glance. I'd like to see some testing in Nova's migrations, but my big concern right now is the ton of stuff blocked in Glance because migrations don't work.20:00
vishyjaypipes: is the only issue with downgrade? or are there other blockers?20:00
jaypipesvishy: pls see the post. There are two blocking bugs in sa-migrate.20:01
justinsbjaypipes: It wasn't clear to me if the second one only came up because of the 1st?20:01
jaypipesjustinsb: the second one refers to when you use MySQL. The first's fix is for the SQLite engine, and the fix is to use SQL scripts, not Python scripts, but that breaks MySQL. :(20:02
justinsbjaypipes: OK, but if bug #1 was magically fixed, would you avoid hitting #2?20:02
*** dprince has quit IRC20:02
uvirtbotjustinsb: Error: Could not parse data returned by Launchpad: The read operation timed out20:02
jaypipesjustinsb: perhaps, not sure yet since I haven't been able to get our unit test to work either wya.20:03
vishyjaypipes: ah ok, I misread your initial post that the bugs were only when doing downgrades.20:03
justinsbjaypipes: Can I check out the code and experience the horror for myself?  Is this in trunk or one of your branches?20:03
*** omidhdl has quit IRC20:04
jaypipesvishy: no, AFAIR, I only mentioned that Nova doesn't have downgrade in a lot of the files... didn't mention there was a bug specific to downgrading.20:04
edayjaypipes: I know you've not tested it yet, but pg (and others) should really be included too, which means sticking with a choice leveraging the abstraction layer (ie, fixing sqlalchmey/sa-migration bugs) might be the best choice.20:04
jaypipesjustinsb: pls see the bug report, I attached all code there:
edayjaypipes: otherwise we'll end up rewriting our own ORM/migration package. :)20:05
jaypipesvishy, justinsb: to be completely frank, the lack of any responses on either bug report (one since November) doesn't make me think the software is either professional or quality. And after looking at the tests in the a-migrate package itself, that conviction got even stronger.20:06
vishyjaypipes: agreed, but do we really want to write it all from scratch?20:07
edayjaypipes: do you think it would be easier to fork/fix or start from scratch?20:07
jaypipesvishy: that was only one option in my email.20:07
btorchhas anyone seen this error while trying to upload an image using nova-manage image ?20:07
vishyjaypipes, eday, justinsb: In a meeting, but i'll keep up with the discussion20:07
*** hadrian has quit IRC20:08
kpepple_btorch: are you on trunk ?20:08
vishybtorch: don't pass data as kwargs20:09
vishyi.e: no kernel_id=XXXX just XXX20:09
jaypipesmtaylor: my current annoyance:
jaypipesmtaylor: its stuck doing python test and I have no idea what is happening with it... as typical, runs fine locally, breaks on Hudson.20:10
btorchvishy: cool let me try that20:11
mtaylorjaypipes: wow - that's special20:12
btorchvishy: kernel_id=006 didn't work either20:12
sorenjaypipes: Sort of.20:12
vishyi'm saying take out 'kernel_id='20:13
vishyjust use the number20:13
*** spartha has quit IRC20:13
*** rds__ has quit IRC20:13
vishythere should be no <key>=<value>20:13
vishyjust use <value> by itself20:13
btorchvishy: ok does the same apply for the other options within [] that "nova-manage image image_register" displays or just  [kernel_id=None] [ramdisk_id=None] ?20:15
btorchok it does :)20:16
mtaylorjaypipes: I get a bunch of this:
btorchcool getting it uploaded now vishy thanls20:16
*** fabiand_ has left #openstack20:18
sorenDoes this ring a bell to anyone?
sorenTowards the end, there's the output from a run of the test suite. It fails.20:20
sirp_nova-core: anyone with some spare cycles mind looking at this:  just needs one more +1 to be merged20:21
vishysoren: old buggy version of python-memcache?20:21
*** nelson has quit IRC20:21
*** nelson has joined #openstack20:21
vishysoren: actually nm, doesn't seem to be memcache related sorry20:22
vishysoren: that looks similar to what was happening with unicode error messages20:22
vishysoren: are you using your danish translations?20:23
sorenvishy: Ah, so there's really an underlying problem that's not being properly reported for some reason.20:23
sorenvishy: No, this is on a buildd.20:23
edaymtaylor: did you add the "0 pep8 violations" check to the burrow tarmac thingy too? if not, could you please? :)20:23
sorenvishy: build deamon.20:23
nelsonIs it just me, or does it seem a little dangerous for the 'delete container' or worse, st's --all operation to exist?20:23
justinsbjaypipes: Sorry was on the phone so missed the tailend of the conversation.  Can I check out the code and experience the horror for myself?  Is this in trunk or one of your branches?20:23
vishyjustinsb: he said the code is in the glance bug report20:24
sorenvishy: ...which means it's a very clean environment.20:24
justinsbvishy: OK - will look there!20:25
clayg_nelson: I've found both options useful on numerous occasions20:25
vishysoren: yes it looks very similar to an old bug where errors would come through as unicode and we would try to set the body with the error message20:26
clayg_kpepple_: I did a pull a few mins before that paste, my images "problem" was just an un-upgraded images dir20:26
nelsonclayg_: I'm just thinking that once we've got all of Wikipedia's images loaded into the cluster, I'm gonna be really REALLY grumpy if someone deletes them.20:26
kpepple_clayg_: yeah, the lp735641 bug actually fixes that for you ... but hasn't been merged into trunk yet20:27
nelsonthat's not waking up on the wrong side of the bed, that's waking up on the wrong side of the world!20:27
clayg_nelson: yeah, maybe you can give these "someone"s a read-only account if you don't trust them to not randomly be typing rm -fr / all over the place20:27
nelsonclayg_: on a bad, day, "someone" is me, so that's no help.20:27
vishysoren: we fixed the old issue on 363 of nova/api/ec2/__init__.py20:27
clayg_nelson: lol!20:27
vishysoren: perhaps there is a similar issue that has been introduced somewhere else20:28
sorenvishy: This only happens on Lucid, fwiw.20:28
clayg_nelson: fwiw the api doesn't support any recursive delete's st adds them as a convience to list -> iter -> delete obj -> delete container20:28
sorenMaybe that can provide a hint of some sort.20:28
nelsonst delete container/file    versus    st delete container /file   is WAY too similar.20:28
sorenvishy: It would be awesome if we could stop masking the error, though.20:28
creihtnelson: inside swift, the delete container command will not be successful... yeah what clayg_ said :)20:29
vishysoren: I'm guessing a dependency is missing somewhere, and the underlying issue is a missing import or some such20:29
nelsonthen maybe I need to put 'st' behind barbed wire?20:29
creihtnelson: I agree that there should be at least a warning in st if you are deleting a container that has data20:29
creihtnelson: file a bug?20:29
sorenvishy: Reasonable guess.20:29
vishysoren: I bet you can wrap the error in str somewhere and actually get a reasonable error message20:29
creihtnelson: or better yet, submit a patch? :)20:30
clayg_kpepple_: oic, because the old image names cause it to blow up during the convert instead of just ignoring them - gotca thanks20:30
creihtshouldn't be too difficult20:30
sorenvishy: Yeah, I'll dig around a bit. Thanks.20:30
*** Ryan_Lane_ has joined #openstack20:30
nelsonyeah, I'm thinking that st should respect some "locked" metadata marker, and just not operate on that container.20:31
nelsonOr maybe only delete a container if it has a magical metadata marker.  turn it into two steps.20:32
nelsonAs long as it's not done frequently, the steps won't get merged together, which is the danger with confirmation.20:32
*** Ryan_Lane has quit IRC20:32
clayg_vishy, kpepple_ : I just confirmed the nova-manage image convert in place images upgrade works a treat20:34
*** Ryan_Lane_ is now known as Ryan_Lane20:34
*** mray has quit IRC20:35
kpepple_vishy , clayg_ : do we have that documented anywhere ? i put something in the cactus release notes about the problem, but nothing with the convert solution. how are we going to handle that when cactus is released -- automatic convert or documented process for upgrade ?20:35
*** spectorclan_ has quit IRC20:35
clayg_oops ment to say it works once you merge in kpepple_'s fix for lp73564120:36
*** mray has joined #openstack20:37
topperanyone know if anyone is working on making swauth not use plaintext passwords? (especially for suerpadmin)?20:37
topperand if not... is swauth just considered "not ready for production" ?20:38
clayg_kpepple_: vish said he was going to update to do it automatically, for a packaged release I'm sure something will have to go in the release notes...20:38
sorenI'd really apprecite a quick review of
*** lvaughn has quit IRC20:38
openstackjenkinsProject nova build #677: SUCCESS in 2 min 4 sec:
creihttopper: It is somewhere in the middle at the moment20:38
openstackjenkinsTarmac: wrap and log errors getting image ids from local image store20:38
sorenIt's quite straight forward and works wonders for my integration tests.20:39
*** lvaughn has joined #openstack20:39
toppercreiht: I didn't see any blueprints for non plaintext20:39
kpepple_clayg_: what's the convert command that you used ? i'll paste that into the release notes so people at least have a clue ...20:39
creihttopper: It should be easy to add20:39
clayg_nova-manage image convert <path-to-images-dir>20:39
topperyeah - I'm considering doing it20:39
creihtNot sure if it is a big enough feature to require a whole blueprint20:39
toppercreiht: I just wanted to know if someone else is already doing it20:39
topperah - I'm new to the whole release cycle20:39
creihttopper: not that I am aware of20:40
topperso in production everyone is just writing their own auth servers?20:40
jaypipesmtaylor: no output anywhere? Is there any way to run a ./ -N for the branch?20:40
creihttopper: In general, we expect most places to already have some sort of auth system that they have to integrate with20:41
*** slyphon has joined #openstack20:41
kpepple_clayg_: got it. thx.20:41
jaypipesmtaylor: the functional tests create a config file for the servers to use (which is what the /tmp/tmpblah stuff is after the calls to the api and registry server binaries...20:41
toppercreiht: so swauth is more "for the future" then20:41
creihttopper: there is currently work being done to provide an auth solution for all of openstack (which will likely replace swauth for swift)20:41
topperah ha20:41
jaypipesmtaylor: there may also be some /tmp/test.[0-9]/ directories that contain log files?20:42
topperso then is modifying swift an exercise in futility then?20:42
creihtwe use swauth for dev, and for internal auth needs (like stats collection)20:42
creihttopper: not at all, some people may still want to use swauth :)20:42
creihtI'm not sure if there is a timeframe for openstack auth yet20:42
creihtswauth will still remain an option for swift, but just not the default20:43
creihtof course all of this is always subject to change :)20:43
topperso forgive my newbness... but proper way is to file a bug with a diff?20:43
creihttopper: file a bug at
creihtmay be of interest20:45
topperah - perfect - I was looking for that20:45
topperyou seem to be an expert... you see any drawbacks to just not using palintext *anywhere* - and using salted/hashed passwords?20:45
topperthe one that bothers me most is the .super_admin20:45
creihtheh... not really, but gholt  may have more opinions20:46
kpepple_clayg_, vishy : can you look at to make sure it is correct ?20:46
*** dendrobates is now known as dendro-afk20:46
creihttopper: well, now that I think about it, hashed passwords will not work for signed request (as with the s3 compatibility layer)20:47
*** azneita has joined #openstack20:47
*** drico has joined #openstack20:47
topperah - I was looking at that code and trying to see where the signing is done20:47
*** lvaughn_ has joined #openstack20:48
*** dendro-afk is now known as dendrobates20:49
topperor I guess the "signed check"20:49
*** johnpur has quit IRC20:49
*** lvaughn has quit IRC20:50
*** brd_from_italy has joined #openstack20:51
Ryan_Laneis 2011.1.1 ever going to get pushed to the release ppa?20:52
Ryan_Lanettx: ^^20:52
creihttopper: it is in the swift/common/middleware/swauth.py20:53
Ryan_Laneor are people expected to only use the tarball?20:53
topperyeah - i'm there and I see if s320:53
creihttopper: starting with line 20120:53
topperoh I see20:53
topper"using my auth account, get the user"20:53
topperand line 22620:54
clayg_kpepple_: ya, looks good20:54
*** mdomsch has joined #openstack20:54
topperok - thanks20:54
*** taihen has joined #openstack20:54
topperdah- securitah fail though - making swauth kinda unusable in production with plaintext passwords in the config file for "root" and then plaintext passwords put into replicas on the storage layer20:55
*** karmabot has quit IRC20:56
topperI wonder if you could have super_admin be salted password and then put some sort of private key encryption when storing users20:56
*** ctennis has quit IRC20:56
creihttopper: I'm beginning to see why this wasn't done the first time around :)20:57
*** _0x44 has quit IRC20:57
topperbasically we're still wrestling with "write our own" or "fix swauth"20:57
creihttopper: what type of timeframe?20:57
topperwe want to remain s3 compatible though20:57
* slyphon works with topper20:57
toppercreiht: eh "medium" :-) - yeah nowish to medium20:57
* slyphon and topper may or may not work for a major computer company that they're not exactly allowed to reveal the name of20:58
creihtso, you don't have any form of auth currently?20:58
creihtslyphon: you aren't the only ones :)20:59
*** donkdonk has joined #openstack20:59
toppercreiht: we do... but our goal is to be as "industry standard" as possible and "support as many clients already out there"20:59
slyphonok, i'll admit it20:59
* slyphon is actually Vint Cerf20:59
topperthis is a new product we're working on21:00
creihtinstalling swift on mars? :)21:00
Ryan_Lanesoren, vishy, ttx: anyone know if 2011.1.1 is going to get pushed to the release ppa?21:00
topperwell only in that little isolation box that simulated it21:00
sorenRyan_Lane: Erk, yes!21:00
Ryan_Laneit's not in the lucid one :(21:00
sorenRyan_Lane: Slipped my mind entirely.21:00
*** taihen has quit IRC21:00
Ryan_LaneI was wondering :D21:01
toppercreiht: thanks so much for your help on this!21:01
sorenRyan_Lane: Good!21:01
Ryan_Lanesoren: thanks :)21:01
topperwe have some stuff to ponder21:01
creihttopper: If you go the route of adding that to swift, the only request I have is to make plaintext still an option21:01
creihtand at least a plaintext->encrypted util21:02
sorenRyan_Lane: They're currently sitting in the nova-core/proposed ppa. If you could take them for a spin, that would be totally awesome.21:02
topperthat was the plan :-)21:02
creihtfor anyone who has a current install21:02
* slyphon winces at "plaintext"21:02
Ryan_Lanesoren: ah ,cool. will do21:02
creihttopper: gholt wrote swauth, so he may have some opinions as well21:03
toppercool - thanks21:03
sorenRyan_Lane: Unfortunately, I don't have a test suite that's loose enough that anything from the Bexar era will pass it.21:04
Ryan_LaneI can pretty much only test flatdhcp21:06
Ryan_LaneI also don't test 100% of the features21:06
*** hadrian has joined #openstack21:07
sorenRyan_Lane: No problem.21:08
*** ctennis has joined #openstack21:10
*** ctennis has joined #openstack21:10
*** h0cin has quit IRC21:10
*** Ephur has quit IRC21:10
*** MarkAtwood has joined #openstack21:11
*** bkkrw has joined #openstack21:11
*** joearnold has joined #openstack21:13
*** karmabot has joined #openstack21:13
*** _0x44 has joined #openstack21:13
*** ChanServ sets mode: +v _0x4421:13
*** RJD22 is now known as RJD22|away21:13
*** ThoughtWorks has joined #openstack21:13
vishyhas anyone noticed that greenthread isn't imported in
*** RJD22|away is now known as RJD2221:16
sorenvishy: Nope.21:23
sorenvishy: I guess live migration hasn't been tested much :(21:24
vishywe need a hudson pyflakes job21:25
*** troytoman is now known as troytoman-away21:25
sorenvishy: Doesn't pylint check this stuff, too?21:25
vishysoren: possibly, but pylint has been way to verbose.  Looks like that is getting fixed.  I have auto error highlighting with pyflakes that picks up this stuff21:26
sorenvishy: /me too.21:26
* soren hugs vim-pyflakes21:26
sorenIt's teh awesome.21:27
vishysoren: agreed, have you figured out how to turn off the few extraneous errors?21:27
sorenLike _?21:27
*** miclorb_ has joined #openstack21:28
vishysoren: like try: import x except: import y as x21:28
sorensirp_: If you're a vim user and not using vim-pyflakes, you're *SERIOUSLY* missing out.21:28
sorenvishy: Oh, that sort of thing.21:28
vishygives you a duplicate import error21:28
sorenvishy: No.21:28
sorenvishy: As in "No, I haven't worked that out".21:28
sirp_soren: yeah, im gonna check it out21:28
justinsbvishy, soren: I have cleaned up pylint so it's a bit more helpful... and eday set up an errors only job21:28
justinsbvishy, soren: Here's
justinsbvishy, soren: But I'd like to see pyflakes as well if it can catch different errors... I know pylint isn't 100%21:29
*** troytoman-away is now known as troytoman21:29
*** troytoman is now known as troytoman-away21:29
vishysirp_ check out for all sorts of vim goodness21:29
*** hub_cap has quit IRC21:29
sorenjustinsb: Nor is pyflakes. At all. pyflakes is just wicked fast.21:29
justinsbsoren: Can you do me a favor - if you find one that pyflakes catches that pylint doesn't, can you let me know... then I'll probably be motivated to add a job for it21:30
sorenjustinsb: So fast, in fact, that it can run every time your pause typing for more then a few hundred milliseconds without breaking a sweat.21:30
sorenjustinsb: Will do.21:30
justinsbsoren: Or we can just use another core for that, which is what we do in the Eclipse world :-)21:30
*** bkkrw has quit IRC21:30
*** vinay_ has joined #openstack21:31
justinsbsoren, vishy: I've requested a discussion at the design summit about greater usage of tools like pylint21:31
sorenjustinsb: Sounds good.21:31
vishyjustinsb: nice job.  Love to have those down to zero and enforce like we do with pep21:32
justinsbvishy: I think now that they are mostly "items of concern", if not actual bugs21:32
sorenI'm all for anything that means we can spend brain power on stuff that actually requires brain power.21:32
uvirtbotNew bug: #739743 in nova "greenthread not imported in libvirt_conn" [Medium,In progress]
*** troytoman-away is now known as troytoman21:32
justinsbvishy: I think once we get the feature backlog merged we should actually go through and clean them all up21:32
*** joearnold has quit IRC21:33
*** bcwaldon has quit IRC21:33
*** troytoman is now known as troytoman-away21:33
sorenvishy: Any comments on ?21:33
justinsbvishy: But we might need beers to ease the dullness of doing so :-)21:34
vishysoren: seems fine, a filelock may be a bit heavyweight, but if it works...21:35
*** adiantum has joined #openstack21:36
sorenvishy: A semaphore would work as well, but I really do wonder if we really need two distinct locking mechanisms.21:36
vishysoren: well i'm already using a semaphore in libvirt for image cache21:36
sorenvishy: I guess utils.synchronized could take an extra arg saying whether it needs to work across multiple binaries.21:36
sorenvishy: Yeah, we need to talk about that :)21:37
vishysoren: perhaps we should generalize it and stick it in utils?21:37
sorenvishy: Yeah, decorator style, I think.21:37
vishysoren: i guess the issue is that it is parameter based21:37
sorenvishy: It feels elegant.21:37
sorenvishy: Is that a problem?21:37
vishysoren: I don't know how to use the decorator syntax in that case, parameterized decorator?21:38
sorenvishy: It already is.21:38
sorenvishy: Or are we talking about different things?21:38
vishysoren: i tend to find them a bit annoying, because when you go back to remember how it works the syntax is ugly21:38
vishysoren: it locks based on image_id21:38
sorenvishy: Yeah. I think it's racy, to be honest.21:39
mtaylorjaypipes: not really sure what's going on - but sort of poking at things hoping to magically fix it21:39
vishysoren: which is a parameter that is passed to the method21:39
sorenvishy: I'm not convinced, at least, that your approach isn't racy. It certainly is if we ever go natively threaded.21:39
*** patcoll has quit IRC21:40
vishysoren: true, so perhaps we just file lock on everything?21:40
sorenvishy: Semaphores are fine. It's just the way you're doing it right there that doesn't really work.21:40
sorenvishy: There's a race between when it checks if there's already a semaphore and when it creates one.21:40
sorenvishy: Two threads coudl do that at once.21:40
vishysoren: right21:41
sorenI started out doing the same thing for the dnsmasq thing.21:41
soren...but then I realised I need to have  global semaphore to lock that dict.21:41
sorenso to speak.21:41
vishysoren: but it is way simpler to understand and works fine with greenthreads21:41
soren..but if I had a global semaphore anyway, I might as well just start dnsmasq while holding it. It's pretty cheap. It's some string operations and a fork.21:41
uvirtbotNew bug: #739750 in nova "XenAPI's Rescue/Unrescue needs unit tests" [Low,In progress]
sorenvishy: I'm not 100% convinced. I don't know if creating a semaphore yields.21:42
vishysoren: if you have an elegant way to switch the volume semaphore to use utils.syncronized, do it21:42
vishysoren: one locking mechanism is best, even if it is a bit less efficient21:42
sorenvishy: True. The file locks are needed in at least one place (iptables).21:42
sorenvishy: So we can't avoid that one.21:43
sorenvishy: ...and it turned out to be thread aware.21:43
*** joearnold has joined #openstack21:43
*** bkkrw has joined #openstack21:43
*** aryan has joined #openstack21:44
vishysoren: I have a pretty good test on the  image cache race, so you should know pretty quick if it works21:44
pvosoren: I'm rebuilding the nova packages for debian and removing dependencies we're not going to use. I'm stuck on the upstart jobs. I can't see where the control file or rules define the upstart-job requirement.21:46
pvoany pointers?21:46
*** patcoll has joined #openstack21:46
sorenpvo: It gets added automatically because there are upstart jobs in debian/21:46
vishysoren: it passed both unit test and functional test so the semaphore is definitely working.21:46
sorenvishy: Oh, no doubt about that!21:47
vishysoren: but like I said, I'm happy for it to be replaced by a file lock.  I just would like it to not block threads creating other images.  There is a test to make sure that doesn't happen as well21:48
openstackjenkinsProject nova build #678: SUCCESS in 2 min 4 sec:
openstackjenkinsTarmac: Wrap update_dhcp in utils.synchronized.21:48
*** allsystemsarego has quit IRC21:48
sorenvishy: Cool. CAn you file a bug about consolidating our locking primitives, perhaps? I'm kind of tied up right now.21:48
vishysoren: sure21:49
*** ptremblett has quit IRC21:52
*** joearnold has quit IRC21:52
*** enigma1 has quit IRC21:52
* soren does a happy dance21:53
sorenAll my integration tests now all work on the first try, too.21:54
pvosoren: I had to do this to get upstart-job to run. Do you think this will be a problem?
*** clayg_ has left #openstack21:54
vishysoren: bug filed21:55
pvoI got the same error as Aaron.21:55
sorenpvo: You had to do what?21:55
vishysoren: link to hudson report?21:55
pvopkg-divert --local --rename --add /sbin/initctl21:55
pvothen ln -s /bin/true /sbin/initctl21:55
sorenpvo: Uh.. Why are you running this in a chroot?21:55
*** donkdonk has quit IRC21:56
*** Jordandev has joined #openstack21:56
sorenpvo: And that is hardly "fixing" anything. The jobs will never run.21:56
vishysoren: nice, now you just need xml output so it can show failing tests!21:56
vishysoren: like so
sorenvishy: Are those smoketests out of date?21:58
vishysoren: that is an old result yes.21:59
sorenvishy: Ok.21:59
sorenvishy: I expect to leave my tests behind very shortly.21:59
sorenvishy: The smoke tests test everything my tests test :)22:00
*** mdomsch has quit IRC22:00
*** urchit has quit IRC22:00
*** nelson has quit IRC22:00
vishysoren: nose produces nice output for jenkins22:01
sorenvishy: it does, yes. My tests just don't use nose.22:01
sorenvishy: They're (very, very) crude bash scripts.22:01
vishysoren: gotcha22:01
uvirtbotNew bug: #739763 in nova "Consolidate locking mechanisms" [Undecided,New]
sorenvishy: moribund ones at that.22:01
*** deshantm has quit IRC22:02
sorenpvo: I'm also curious why you're running this on Debian?22:03
pvoits what the ops team chose to deploy on.22:04
pvonot *eveyone* runs ubuntu yet. : )22:04
sorenThey do realise we have a reference platform and it's something else, right?22:04
sorenJust saying.22:04
*** deshantm has joined #openstack22:04
pvoI know. I asked the same question.22:04
pvowe also assume with the ubuntu packaging that you're going to run kvm22:05
pvowhich is a dependency I'd need to strip out anyways22:05
*** dspano has quit IRC22:06
btorchjaypipes: cool thanks the novatools client works for getting image info that I have in glance :)22:08
*** azneita has quit IRC22:09
*** joearnold has joined #openstack22:10
*** patcoll has quit IRC22:11
*** Shentonfreude has quit IRC22:11
btorchI assume that the python-novaclient is xen specific ?22:12
sorenbtorch: Not at all.22:13
sorenpvo: Yeah, no XenServer in Debian nor Ubuntu.22:13
pvowe'll, you should be able to install the compute node without having to require kvm, no?22:14
btorchsoren: cool,  I've uploaded an ubuntu1010-UEC image/kernel with "nova-manage image" but wondering how the nova client will bring that up22:16
sorenpvo: Since we also support uml and lxc, yes.22:17
sorenpvo: "should" being the operative word, though.22:18
btorchsoren: nova image-list doesn't really display which is a kernel or image and what kernel belongs to an image .. I guess I'm still thinking ec2 style :)22:18
sorenbtorch: I'm not the right person for these questions :)22:20
*** hggdh has quit IRC22:21
*** hggdh has joined #openstack22:22
*** hggdh has joined #openstack22:22
MotoMilindHello Soren.  I was wondering if you could help me with this use-case: I want to spin up an instance, configure it for my app and then, from outside the instance (may be from controller node), create and register its updated image for future instance creation.  Is there a set of commands in nova (or even base Ubuntu) for that?  Something like your vmbuilder, but by pointing at a running instance, instead of being inside an instance22:25
MotoMilindThanks in advance22:25
*** jero has quit IRC22:26
sorenMotoMilind: The Rackspace API has support for that. I don't know if it's made it into OpenStack yet, to be honest.22:29
pvoMotoMilind: you should (very soon) be able to take a snapshot of your instance and have it pushed to glance22:33
*** brd_from_italy has quit IRC22:34
*** gasbakid has quit IRC22:34
pvowhich would push it to whatever back end you set your glance repository to. this could be ftp, local disk, s3 or swift22:34
*** topper has quit IRC22:36
*** Ryan_Lane_ has joined #openstack22:38
*** jc_smith has joined #openstack22:40
*** jc_smith has left #openstack22:40
sorenvishy: I think we need some combination of semaphores and filelocks.22:41
*** Ryan_Lane has quit IRC22:41
uvirtbotNew bug: #739788 in nova " -> get_info should only accept a vm opaque reference" [Undecided,New]
*** Ryan_Lane_ is now known as Ryan_Lane22:41
*** mray has quit IRC22:44
sorenvishy: Either that or changes to trunk are landing too fast for my test suite to finish, so stuff fails during the package upgrades.22:45
MotoMilindpvo: Is there a code snippet I can look at until your described functionality is available?  Doesn't matter if it is working or not.22:46
pvosirp_: may be able to help here22:46
pvohe is working on glance integration22:46
*** ThoughtWorks has quit IRC22:47
*** gasbakid has joined #openstack22:47
*** joearnold has quit IRC22:48
sirp_MotoMilind: if you're using the XenServer hypervisor and have python-novaclient handy, you can do something like novaclient image-create <SERVER ID> "My Image Name" to take snapshot of your running instance, you can then use novaclient create "My Server Name" --image <IMAGE ID> to boot an new server from that image22:51
*** MarkusT has joined #openstack22:57
MarkusTIs there another place where nova stores the MySQL password, besides nova.conf?22:57
*** pvo has quit IRC22:58
*** joearnold has joined #openstack22:58
*** joearnold has quit IRC23:00
*** CloudChris has quit IRC23:01
*** westmaas1 has joined #openstack23:11
*** dfg has quit IRC23:12
*** zaitcev has joined #openstack23:23
*** MarkusT has quit IRC23:24
uvirtbotNew bug: #739815 in nova "euca-authorize and euca-revoke throws unknown error for invalid port range" [Undecided,New]
MotoMilindOh wow, I will try that!  If that works, I will try to integrate it with Dashboard!23:26
vinay_is this the right forum to ask newbie questions??23:27
vinay_I am trying to do a single node installation of nova ... going through admin guide... and running into problems23:27
*** kapil has quit IRC23:28
*** rnirmal has quit IRC23:29
*** ThoughtWorks has joined #openstack23:30
uvirtbotNew bug: #739818 in nova "euca-authorize and euca-revoke gives Unknown error for invalid source subnet parameter" [Undecided,New]
kpepple_vinay_: sure. fire away.23:31
vinay_I started nova by running sudo ./ run23:33
vinay_but the nova-volume cmd is returning an error23:34
vinay_Error: volume group nova-volumes doesn't exist23:34
kpepple_vinay_: if you haven't created a volume group with the name "nova" it will do that.23:34
kpepple_vinay_: unless you need volume support, you can ignore it23:35
vinay_I see... since I am just starting to learn the openstack .. I guess I can ignore it..23:35
vinay_so for future .. the resolution to this problem is to "create a volume group with the name nova" ??23:36
Ryan_Lanesoren: I'm not seeing the nova-ajax-console-proxy package available in 2011.1.1...23:36
*** rds__ has joined #openstack23:36
Ryan_Laneor the service23:36
kpepple_vinay_: see for setting up the volume group (a simple way)23:36
vinay_okay.. thanks kpepple_23:37
vinay_The setting this link talks about .. which file does it go into?23:37
kpepple_vinay_: sorry, what do you mean by "setting" ? these are commands that you execute at the command prompt (it looks like the $ is missing)23:38
vinay_oh ok..23:39
vinay_I got it wrong then.. will try that :)23:39
vinay_thanks for now23:39
*** ThoughtWorks has left #openstack23:40
kpepple_vinay_: np23:40
Ryan_LaneI also don't see euca-get-ajax-console23:43
*** londo_ has joined #openstack23:43
*** MotoMilind has quit IRC23:44
*** MotoMilind has joined #openstack23:44
*** MotoMilind has quit IRC23:44
*** MotoMilind has joined #openstack23:45
*** gregp76 has quit IRC23:57
*** nelson has joined #openstack23:57
*** clauden_ has quit IRC23:59

Generated by 2.14.0 by Marius Gedminas - find it at!