Thursday, 2011-04-14

*** RobertLaptop has joined #openstack00:01
*** santhosh has joined #openstack00:01
*** ccustine has quit IRC00:01
*** jkoelker has quit IRC00:01
*** joearnold has quit IRC00:04
*** joearnold has joined #openstack00:04
*** joearnold has joined #openstack00:05
*** maplebed has quit IRC00:06
*** neogenix has joined #openstack00:10
*** kashyap has quit IRC00:14
*** rchavik has joined #openstack00:15
*** kashyap has joined #openstack00:16
*** dragondm has quit IRC00:21
*** ryker has left #openstack00:25
*** kashyap has quit IRC00:25
*** kashyap has joined #openstack00:26
*** santhosh has quit IRC00:27
*** filler has quit IRC00:27
*** hagarth has quit IRC00:28
*** adjohn has joined #openstack00:28
*** kashyap has quit IRC00:33
*** adjohn has quit IRC00:41
*** adjohn has joined #openstack00:43
*** joearnold has quit IRC00:43
*** zul has joined #openstack00:45
*** elambert has quit IRC00:47
*** kashyap has joined #openstack00:49
*** benbenhappy has joined #openstack00:51
*** mdomsch has joined #openstack00:54
*** mdomsch__ has joined #openstack00:54
*** mdomsch_ has joined #openstack00:54
*** mdomsch_ has joined #openstack00:55
*** hggdh has quit IRC01:06
*** hggdh has joined #openstack01:07
*** filler has joined #openstack01:07
*** joshfng has joined #openstack01:15
*** joshfng has quit IRC01:23
*** Ryan_Lane has joined #openstack01:24
*** gaveen_ has joined #openstack01:29
*** onlany has quit IRC01:31
*** gaveen has quit IRC01:32
creihtyay... seems to be down :/01:50
*** rds__ has quit IRC01:54
*** santhosh has joined #openstack01:55
*** santhosh has quit IRC02:04
*** Ryan_Lane has quit IRC02:04
*** gaveen_ has quit IRC02:14
*** jeffjapan has quit IRC02:16
*** elambert has joined #openstack02:17
*** jeffjapan has joined #openstack02:20
*** dysinger` has quit IRC02:22
*** mray has quit IRC02:39
*** j05h has quit IRC02:42
*** AimanA is now known as HouseAway02:50
*** dendro-afk is now known as dendrobates02:51
*** RJD22 is now known as RJD22|away02:54
*** sebastianstadil has joined #openstack02:55
*** RJD22|away is now known as RJD2202:59
*** shermanboyd has joined #openstack03:01
shermanboydhmm I don't understand the network diagram in the nova admin book (section 3.2.1)03:04
shermanboydwhy would you place the compute services node and db over the internet03:04
shermanboydinstead of locally with the compute nodes03:05
*** mahadev has joined #openstack03:08
*** sebastianstadil_ has joined #openstack03:09
*** sebastianstadil has quit IRC03:11
*** sebastianstadil_ is now known as sebastianstadil03:11
*** shermanboyd has quit IRC03:22
*** shermanboyd has joined #openstack03:25
*** guynaor has joined #openstack03:26
*** rsaidan has joined #openstack03:33
*** guynaor has left #openstack03:34
HugoKuo_hi vishy03:35
HugoKuo_I solve my problem :>03:35
*** Ryan_Lane has joined #openstack03:35
HugoKuo_thanks for you advice03:35
*** joearnold has joined #openstack03:36
*** reldan has quit IRC03:41
*** koolhead17 has quit IRC03:46
*** j05h has joined #openstack03:47
dsockwellshermanboyd: I'm no expert, but I imagine it's because you can.  there shouldn't be any reason not to put them on a LAN though.03:50
shermanboyddsockwell: thanks ...03:55
*** sebastianstadil has quit IRC04:03
*** mahadev has quit IRC04:04
*** clayg_ has joined #openstack04:04
*** sebastianstadil has joined #openstack04:10
*** smeier00 has joined #openstack04:12
*** adiantum has quit IRC04:13
*** paltman has quit IRC04:14
*** paltman has joined #openstack04:15
*** benbenhappy has left #openstack04:19
*** nwww has quit IRC04:19
*** adiantum has joined #openstack04:27
*** f4m8_ is now known as f4m804:44
*** miclorb_ has joined #openstack04:47
*** mdomsch__ has quit IRC04:53
*** mdomsch_ has quit IRC04:53
*** mdomsch has quit IRC04:53
*** clayg_ has quit IRC04:54
*** ianweller has quit IRC05:05
*** ianweller has joined #openstack05:05
*** ianweller is now known as Guest6828305:05
*** miclorb_ has quit IRC05:05
*** lionel has quit IRC05:14
*** lionel has joined #openstack05:14
winston-dBK_man : around?05:16
*** zenmatt has quit IRC05:25
*** joearnold has quit IRC05:27
winston-dHugoKuo_ : how did you solve ur problem?05:28
HugoKuo_winston-d : add iptable rule manually and open promics on br10005:34
winston-dr u using VlanManager or Flat?05:34
*** miclorb_ has joined #openstack05:35
winston-di see.  I'm using Vlan.05:36
HugoKuo_I finally success to separate nova-network05:36
HugoKuo_what's your prograss now ?05:36
winston-di'm struggle with Vlan. :/  I noticed that <filterref> part of <interface> caused unreachable network.05:41
winston-di'd like to confirm that it's the case for others.05:42
winston-dI think I understand why network is not working... hmmm05:48
*** guigui has joined #openstack05:49
*** santhosh has joined #openstack05:53
HugoKuo_so the instace can be run up ?05:57
HugoKuo_but can not be access through SSH or ping ?05:57
HugoKuo_right ?05:57
*** koolhead17 has joined #openstack05:58
*** santhosh has quit IRC06:00
*** smeier00 has quit IRC06:00
*** nwww has joined #openstack06:16
winston-dHugoKuo_ : right06:16
*** smeier00 has joined #openstack06:26
*** smeier00 has left #openstack06:29
*** flashn has quit IRC06:30
uvirtbotNew bug: #760422 in nova "exception.NotFound occurred when installing nova-objectstore in a different marchine from nova-compute" [Undecided,New]
*** elambert has quit IRC06:34
*** neogenix has quit IRC06:38
*** mramige has joined #openstack06:38
*** lionel has quit IRC06:42
*** lionel has joined #openstack06:43
ttxvishy: if you're still up, let me know if you're ok to freeze06:44
sorenttx: We still need the Final =TRue patch.06:51
sorenttx: I'll whip one up.06:51
sorenttx: Or do you want to?06:51
ttxI have branches ready06:51
ttxsoren: but apparently there is still one branch up the pipe06:52
BK_manwinston-d: yep, here06:52
ttxthough LP codereview seems to be busted06:52
ttxsoren: I'm trying to piece the puzzle together06:52
sorenttx: Which branch is in the piipe?06:52
winston-dBK_man : i was about to ask you some question related to Vlan manager.06:52
winston-dBK_man : but I think i've got it figured out.06:53
ttxsoren: vishy/jaypipes mention a new RC bug in the backlog, bug 76028906:53
uvirtbotLaunchpad bug 760289 in nova "All image types return as 'machine' when using glance client in ec2 api" [High,In progress]
BK_manYeeha! I'm not only one facing that issue:
uvirtbotLaunchpad bug 760422 in nova "exception.NotFound occurred when installing nova-objectstore in a different marchine from nova-compute" [Undecided,New]06:53
ttxsoren: though they didn't set the milestone clearly06:53
BK_manwinston-d: np, ask each time. I will answer when I can06:53
BK_manttx: is that a regression?
uvirtbotLaunchpad bug 760422 in nova "exception.NotFound occurred when installing nova-objectstore in a different marchine from nova-compute" [Undecided,New]06:54
*** mgoldmann has joined #openstack06:55
ttxBK_man: I doubt it is06:55
ttx<vishy> 22:46:32> jaypipes...i'd say it is critical06:55
ttx<jaypipes> 22:46:39> vishy: I would agree.06:55
* ttx grumbles. Targeting bugs to RC a few hours to RCFreeze is ok if you can get the patch in in time06:58
* ttx kicks in the balls06:59
ttx"builds, merge proposals and mailing lists having trouble (fallout from hardware failure) - back soon"06:59
ttxsoren: looks like we'll have a wait a few hours to cut the RC.07:00
*** zigo-_- has joined #openstack07:03
zigo-_-alekibango: Hi !07:03
*** rcc has joined #openstack07:04
winston-dBK_man : did u get glance work as image service?  I've got some issues.07:07
*** mihgen has joined #openstack07:08
winston-dOK. did you try it and failed or you just sticked to objectstore for now?07:09
BK_manI have two traces - one from Glance and another one from Local storage07:11
BK_manwinston-d: glance:
zigo-_-After a fresh pull, I got the following error:
zigo-_-What could that be???07:11
BK_manwinston-d: sorry, last link was from local storage07:11
BK_manwinston-d: glance:
zigo-_-"OSError: [Errno 2] No such file or directory"07:12
zigo-_-Which thing is it trying to access?07:12
BK_manzigo-_-: is your question about my trace?07:12
zigo-_-No, about mine! :)07:13
BK_manzigo-_-: show me yours07:13
zigo-_-Fresh pull from minutes ago...07:14
winston-dBK_man : both traces are failed.  for glance, that looks familiar to me.  I always get 400 response from glance server.07:14
zigo-_-Just reinstalled my test server from scratch, to test missing dependencies...07:14
BK_manzigo-_-: something strange happened.07:14
zigo-_-Also, something tried to write a lock file in /usr/lib/python2.6 ...07:15
zigo-_-I had to put some loose perms there.07:15
BK_manzigo-_-: I put debug code in Nova and discovered that cp failed in my code with No such file or directory. However, there is a file and permissions & owner are correct07:15
*** CloudChris has joined #openstack07:16
zigo-_-Oh, I think I got it.07:16
zigo-_-qemu-img is not there!07:17
zigo-_-Yet another missing dep...07:17
BK_manzigo-_-: you're lucky man! Mine problem is more complicated :(07:17
winston-dBK_man : in your bexar RHEL packages, 'nova-manage log request' would fail due to the modification of nova log settings.07:20
*** thor has quit IRC07:20
BK_manwinston-d: never tested it! thanks for reporting07:20
BK_manwinston-d: hold on switching to cactus. trunk is broken at the  moment07:21
winston-dBK_man : will u do 'C' packaging?07:21
winston-dBK_man : yeah, i'd wait until you release RHEL RPMs. :)07:21
BK_manwinston-d: yep, in fact we are doing trunk builds and trunk is almost cactus at the moment07:22
BK_manwinston-d: the only small problem is that trunk code is not working now :)07:22
winston-dBK_man : looking forward to a working version from u.07:23
zigo-_-Well, now got another one...07:23
zigo-_-Could it be that my libvirt doesn't have the needed patches? (it's from Debian Experimental, not from PPA...)07:25
zigo-_-BK_man: Are you doing the RPMs for Fedora? :)07:26
*** sebastianstadil has quit IRC07:26
BK_manzigo-_-: not now. But I can't see any stops compiling my RPMs on Fedora07:26
zigo-_-So in which distro?07:27
zigo-_-SUSE ?07:27
BK_manzigo-_-: my RPMs targeted specifically RHEL 6.0 at the moment07:27
zigo-_-Ok, good to know.07:27
zigo-_-I'm doing the Debian packaging ...07:28
zigo-_-Got a bunch of patches against the Ubuntu debs already.07:28
BK_manzigo-_-: please let me know if you obtain information about Nova-specific libvirt patches07:28
BK_manzigo-_-: forgot to ask - are you going to Summit?07:28
BK_manzigo-_-: my mail is in case if I absent on IRC07:29
*** nacx has joined #openstack07:29
zigo-_-Unfortunately, I have no budget for that, my passport is not a biometric one, and I have other stuff planned.07:29
zigo-_-I hope to go to the next one.07:29
*** lionel has quit IRC07:29
BK_manzigo-_-: where are you from?07:29
zigo-_-I'm a French guy living in Shanghai, China.07:30
HugoKuo_winston-d : my topology
*** lionel has joined #openstack07:30
BK_manzigo-_-: Ah, ok. I already asked you that question. :(07:30
zigo-_-People in here are interested a lot with RPMs for Openstack. Especially people from China Standard Software Co. (the biggest Unix distro in China)07:30
zigo-_-BK_man: And yourself, were are you from?07:31
winston-dzigo-_-: is benbenhappy from China Standard Software?07:31
zigo-_-He is my friend.07:32
zigo-_-I'll see him tonight.07:32
*** beudbeud has joined #openstack07:33
winston-dzigo-_- : OK07:33
BK_manzigo-_-: I'm from Russia07:34
winston-dHugoKuo_ : why do you want to separate nova-network ?07:34
HugoKuo_Does any know how to resolve  " need to NFS mount /var/lib/nova/images folder from nova-bojectstore to nova-compute ?07:34
HugoKuo_winston-d : bcz I want to try HA for nova-network07:35
*** beudbeud has left #openstack07:35
HugoKuo_I believe that while nova-network dead.....07:35
HugoKuo_we can not access into instance....07:35
HugoKuo_so that I separate nova-network from controller node :)07:36
HugoKuo_is this make sense ?07:36
winston-dHugoKuo_ : i think if one instance is up and running (network connection is OK), nova-network has nothing to do with client-to-VM traffic anymore.07:36
HugoKuo_how about nova-network dead ....07:37
*** Ryan_Lane has quit IRC07:38
winston-dif nova-network session/process/host is dead, the impact would be you won't be able to create new instance.  that's my understanding.07:38
HugoKuo_btw , you are right about that , but once that host dead.....07:38
winston-dunless you have VPN service in nova-network host07:39
HugoKuo_all instance traffice go through from nova-network07:39
zigo-_-Where may I find image tarballs that are known to work???07:39
HugoKuo_right ?07:39
HugoKuo_zigo-_- : wait07:39
zigo-_-It seems the one I used doesn't.07:39
HugoKuo_I'm using 64 bits now07:40
HugoKuo_did you mention that instance's gateway is on nova-network ..?07:41
zigo-_-Thanks, I will try that one.07:42
HugoKuo_zigo-_-: UEC image instead of Amazon EC2 image07:42
HugoKuo_maybe you should check the log file or console output ...07:42
zigo-_-It says 'i386' instead of x86_64 after the upload ...07:44
HugoKuo_IMAGE   ami-18b939f8    maverick/maverick-server-uec-amd64.img.manifest.xml             available       public          x86_64  machine aki-1fac05cb07:45
HugoKuo_mine is x86_64 .........07:45
HugoKuo_same image...07:45
HugoKuo_what's your nova version ?  host OS version ?07:45
zigo-_-Debian Squeeze.07:47
zigo-_-Last pull from bzr...07:47
winston-dHey guys, how long does it take in your setup for Nova to bring up a new instance?07:47
HugoKuo_depends on the size of image07:48
winston-dBK_man, HugoKuo_ : you are using KVM, right?07:48
HugoKuo_for exapmple tty image07:48
HugoKuo_winston-d : yes07:48
BK_manwinston-d: correct07:48
*** antenagora has joined #openstack07:48
*** CloudChris has left #openstack07:48
HugoKuo_first time to bring up a instance take more time07:49
HugoKuo_bcz nova-compute must load image and untar07:49
winston-dBK_man, HugoKuo_: do you guys have any special treatment for HVM guest?  Here it seem adding <kernel> <ramdisk> in libvirt.xml would cause instance failed to boot.07:50
HugoKuo_no experience :\07:50
BK_manwinston-d: nope. no special customization was required for KVM from my side07:50
*** daveiw has joined #openstack07:51
*** rcc has quit IRC07:51
winston-dBK_man : here's the console output of my newly created instance.  very long trace, you can skip the beginning and look for the end.
winston-dBK_man : I double checked that I used the correct kernel/ramdisk.  Don't know why it doesn't work.  Removing <kernel><initrd> should help.  Let me try.07:56
BK_manwhat version of RHEL in guest?07:56
winston-dThe guest is RHEL 607:58
BK_manMine is 5.6, 6.0 guest preparation in plans. let me check what's if /etc/fstab in my guest07:59
BK_manwinston-d: /dev/vda                /                       ext3    defaults        1 108:00
BK_manwinston-d: that's what I have in guest in /etc/fstab08:00
winston-dBK_man : i recalled CentOS 5.3 was OK while I tested 'A' release.08:00
BK_manwinston-d: initrd-2.6.18-238.el5-virtio.img - initrd with virtio driver08:01
winston-dBK_man : my guest was installed onto a Qcow2 virtio backed image, so I assume the ramdisk already includes virtio driver?08:02
BK_manwinston-d: indeed08:02
* BK_man finished with management call. uff...08:04
winston-dBK_man : haven't had lunch yet?08:04
BK_manwinston-d: nope. workday just started. We have shifted worktime to better match colleagues in US, CA08:05
winston-dBK_man : meaning getting up early and staying up late...08:06
BK_manwinston-d:  yep08:06
winston-dBK_man : unfortunately, the fstab for my guest is LVM based...08:09
winston-dis that the root cause?08:09
HugoKuo_hey man , I forgot the flag of nova-volume ....08:13
HugoKuo_assume VG= ubuntu2  nova-volume host ip=
*** sebastianstadil has joined #openstack08:19
ttxsoren: you should still be able to pull lp:~jaypipes/nova/bug760289 and judge its merits offline, so that it's ready to go in whenever LP recovers08:26
ttxI'll announce RCFreeze with a provision for merging that last one, just after coffee.08:26
*** bkkrw has joined #openstack08:29
*** allsystemsarego has joined #openstack08:31
*** kashyap_ has joined #openstack08:33
*** kashyap has quit IRC08:34
*** dodeeric1 has joined #openstack08:37
BK_manwinston-d: root=/dev/vda console=ttyS0 selinux=008:37
BK_manwinston-d: from libvirt.xml08:38
*** sebastianstadil has quit IRC08:38
winston-dBK_man : same here for that line08:38
winston-dBK_man : i think the root cause is not supporting LVM image with separate kernel & ramdisk.08:42
BK_manwinston-d: yep. how did you prepare your image?08:42
winston-dinstall a VM locally and then bundle/register via euca2ools08:43
*** jeffjapan has quit IRC08:45
*** sebastianstadil has joined #openstack08:49
*** onlany has joined #openstack08:50
*** adjohn has quit IRC08:56
*** headoff has joined #openstack08:59
*** flopflip_ has joined #openstack09:01
HugoKuo_    connect refused during create volume prograss ...... any suggestion ?09:03
*** flopflip has quit IRC09:04
ttxsoren: looks like it's working -- please review
ttxand accept if you're alrigth with it, Vish reviewed it offline and acked it09:04
*** arun_ has joined #openstack09:06
HugoKuo_forget about that question , I just forgot to enable iscsi09:07
*** dirkx_ has joined #openstack09:10
*** DigitalFlux has quit IRC09:10
*** miclorb_ has quit IRC09:12
ttxsoren: is next :)09:14
sorenttx: Yeah, I saw :)09:15
BK_mansoren: ttx: really???09:22
sorenBK_man: Really what?09:23
* BK_man thinks that RC buggy09:23
ttxBK_man: I won't wait for 150 bugs to be fixed to release it.09:23
BK_mansoren: ttx: localstorage is broken09:23
ttxBK_man: did it use to work ?09:24
BK_manttx, yep, couple days ago09:24
openstackjenkinsProject nova build #817: SUCCESS in 2 min 29 sec:
openstackjenkinsTarmac: Rework GlanceImageService._translate_base() to not call BaseImageService._translate_base() otherwise the wrong class attributes are used in properties construction...09:24
BK_manttx: however, i can't run any instance on Glance also09:24
ttxBK_man: looks like a packaging problem to me09:25
BK_manttx: commented today on Ubunty - exact problem09:25
BK_manttx: let me search a link for you09:25
BK_manttx: soren:
uvirtbotLaunchpad bug 760422 in nova "exception.NotFound occurred when installing nova-objectstore in a different marchine from nova-compute" [Undecided,New]09:26
ttxBK_man: if you don't file a bug, it's a bit difficult for us to know about it.09:26
* ttx sighs09:26
BK_manttx: now hardy working to determine which rev introduced a regression09:26
ttxBK_man: you said it used to work... when was that ?09:27
uvirtbotNew bug: #760484 in nova "Suspend/Pause not supported for libvirt" [Undecided,New]
sorenI think this is a known issue.09:27
BK_mantxx: bug is still in bzr977, checking another rev09:27
ttxBK_man: I think it never was supported.09:27
HugoKuo_my nova-objectstroe is not stay with compute node , is that the reason why I have to NFS mount /var/lib/nova/images to compute node ?09:27
ttxBK_man: ISTR Vish saying that you should use Glance if you use multiple machines09:28
BK_manttx: I'm not using multiple hosts at the moment.09:28
ttxBK_man: but if you can point to a recent release that worked... than I'd agree it's a regression09:28
sorenI wonder if setting it to an S3-like backend helps.09:28
HugoKuo_ttx : thanks , I'll try glance09:28
BK_manttx: but let me provide you another trace using Glance...09:28
BK_manttx: soren: this is using Glance:
BK_man# ls -l /var/lib/nova/instances/_base/09:32
BK_mantotal 468409:32
BK_man-rw-r--r-- 1 nova nobody 2094684 Apr 14 13:30 0000000109:32
BK_man-rw-r--r-- 1 nova nobody 2698697 Apr 14 13:30 0000000209:32
BK_man-rw-r--r-- 1 nova nobody       0 Apr 14 13:30 00000003_sm09:32
* BK_man above is too small for, sorry guys :)09:32
sorenToo small?09:32
sorenHow so?09:32
ttxsoren: the ls -l output, I suppose09:33
BK_mansoren: so, it can't copy rootfs from unexistant file09:33
ttxsmall enough to not warrant a pastebin09:33
ttxBK_man: your last paste looks like a different issue09:34
*** ramkrsna has joined #openstack09:34
*** nwww has quit IRC09:35
ttxBug 760422 looks a lot by design.09:37
uvirtbotLaunchpad bug 760422 in nova "exception.NotFound occurred when installing nova-objectstore in a different marchine from nova-compute" [Undecided,New]
ttxsince we now defaults to local storage09:38
ttxBK_man: in all cases I'd be surprised if it worked recently. At least not since objectstore was refactored09:40
ttxBK_man: please file a separate bug for your Glance-based issue09:40
BK_manttx: I don't think it's Glance issue since it look almost the same for Glance & local storage09:41
* BK_man debugging _cache_image in libvirt_conn.py09:41
BK_manshould _cache_image get 0000003 instead of 0000003_sm for rootfs?09:42
ttxthey look different to me. In your last trace you should get more info in Glance logs09:44
zigo-_-How can I do a forced push with bzr?09:44
zigo-_-I rewrote my branch...09:44
ttxsince the glance client complains ab out the Glance server status line09:44
zigo-_-Now, I'd like that what's online on launchpad be ovewritten by my new commits.09:44
zigo-_-So, what's the equivalent of "git push --force" ?09:45
BK_manis that a known issue with glance?09:45
ttxBK_man: what glance version are you running ?09:46
BK_manttx: trunk autobuilds. let me check09:46
BK_manttx: bzr11309:46
sorenzigo-_-: Why?09:47
sorenzigo-_-: There's a --overwrite, but it's thoroughly discouraged.09:47
zigo-_-soren: Because you didn't take my changes, and it diverged ...09:48
zigo-_-And also, my commit history wasn't clean at all.09:48
zigo-_-So I wanted to push something with a nicer history.09:49
zigo-_-Can you pull now?09:49
sorenzigo-_-: Noone looks at the history anyway.09:49
sorenIn bzr, you don't (you can, but you don't) rebase. You just merge.09:49
sorenbzr is not a patch management system. It's a revision control system.09:50
zigo-_-I'd appreciate if you'd just pull my changes now, so it wont diverge again.09:51
zigo-_-Plus, we already discussed the changes...09:52
sorenDiverging branches is not a problem.09:52
sorenIn itself.09:52
zigo-_-Why do you always love to discuss? :)09:53
zigo-_-Just pull, ack, move forward ...09:53
sorenBecause, like you, I'm not just going to take other people's changes and blindly apply them to my code and upload anywhere.09:53
zigo-_-I never said that.09:54
sorenSure you did.09:54
ttxBK_man: strange bug, pretty sure it's a corner case09:54
sorenzigo-_-: "Moreover, I hope my work wont just be "downstream", I hope we will constructively collaborate and work together! If what you expect me to do is just download your work, do a build, and upload blindly to Debian, that wont work, I'll be of no use. "09:55
zigo-_-Which is why I am sending you patches.09:56
sorenRight. And it's why I'm not just pulling, acking, and moving forward. I have to review this stuff.09:56
soren..and now that you've rebased, I have to go through *everything* again, because I can't just look at your most recent chagnes.09:56
* ttx disappears to get his daughter from school, back in ~90min09:57
sorenttx: Have fun.09:57
sorenzigo-_-: I can't help but notice that you've documented *nothing* in debian/changelog.09:58
zigo-_-soren: I can do that if you want...09:59
zigo-_-But more or less, it's going to be a cut/past from "bzr log".09:59
sorenzigo-_-: That's fine.10:01
zigo-_-doing it...10:01
*** headoff has quit IRC10:04
sorenThe man pages still talk about a "cloud controller", apparently.10:06
zigo-_-That can be fixed later on, no?10:07
sorenWhy the rush to get in?10:07
zigo-_-It's been quite some time these are ready!10:08
sorenIT's been quite some time since I pointed this out to you. And now you want me to accept the patch instead of you fixing it. That's "rush" in my book.10:09
zigo-_-So, because of man pages that are only stubs to be filled when we got time, you are going to reject ALL ?10:11
sorenYou are proposing this as one big patch.10:12
sorenSo yes.10:12
zigo-_-Take all the commits but the man pages then...10:12
sorenI'm not done reviewing it again.10:13
zigo-_-I did a clean history for this very reason!10:13
sorenAs I said: You rebased, so I have to start all over.10:13
zigo-_-Anyway, I'm fixing...10:13
zigo-_-Pushed up to revision 145.10:16
zigo-_-There was other mistakes too that I fixed in these man pages.10:16
* soren also never subscribed to the idea that a stub man page is better than no man page, but meh.10:16
BK_manttx: false alarm10:16
BK_manttx: with Glance10:16
zigo-_-Well, the man pages that I wrote at least document the --flag-file parameter that is used, and what the daemons are for.10:17
BK_manttx: soren: sleep 15 between euca-upload-bundle (the final one) and euca-run-instances solved my problem10:17
zigo-_-So it's better than nothing.10:17
BK_manttx: soren: uploading of rootfs.qcow2 takes ~10 secs10:17
zigo-_-I got to go.10:18
BK_manttx: soren: and I think it's async rather than sync10:18
zigo-_-Shanghai Linux User Group hacking thursday time ... :)10:18
*** zigo-_- has quit IRC10:18
alekibangosoren: stub man page is something to bug about :)   missing man page looks like intent10:20
*** adiantum has quit IRC10:25
HugoKuo_is the difference between Glance and SWIFT , Glance provide image service for instance ?10:31
BK_manHugoKuo_: Glance also can be a client for Swift10:41
* BK_man running to lunch10:41
*** dirkx_ has quit IRC10:57
HugoKuo_what if I just want to use glance for compute node to retrive image , is that very hard ?11:03
*** kakoni has joined #openstack11:05
*** smaresca has quit IRC11:20
mihgenCan be Nova configured to use Swift directly, without Glance?11:20
sandywalsh_soren, heh ... good idea putting "" in the sig of the review day email11:21
HugoKuo_mihgen : in my knowing "can not"11:22
mihgenHugoKuo_: That's what I thought. Have you succeed to connect glance->swift?11:24
sorensandywalsh_: Still, almost everyone responds to it. Seriously :)11:24
HugoKuo_mihgen : sorry I did not use swift now . I play with Swift 4 months ago ....11:24
sandywalsh_soren, I was about to until I saw it11:25
sorensandywalsh_: Well, you still responded. Just on IRC :)11:26
*** miclorb_ has joined #openstack11:26
sandywalsh_soren last time ... promise11:26
sorenOh, it's fine! :)11:27
sorenIt's just mildly amusing.11:27
*** patcoll has joined #openstack11:27
* onlany is waiting for cactus release very much :)11:28
*** DigitalFlux has joined #openstack11:28
*** DigitalFlux has joined #openstack11:28
*** nwww has joined #openstack11:30
*** smaresca has joined #openstack11:33
ttxNova RC milestone is up11:34
*** ctennis has quit IRC11:34
ttxGlance RC milestone is up11:35
ttxnotmyname: waiting on you to push the 1.3-rc versioning branch at
ttxnotmyname: or let me know if I can release RC with "1.3-dev"11:36
rsaidanttx: can you tell me what ppa should we be using for the cactus-rc release?11:36
ttxrsaidan: the RC should appear on the nova-core/trunk PPA11:37
ttxas "bzr988"11:37
* ttx looks up if it's there already11:37
rsaidanttx: thx.. trunk is the dev branch right? do we use this until cactus is officially released?11:38
ttxrsaidan: yes11:39
rsaidanttx: great11:39
ttxsoren: looks like glance bzr113 never made it to PPA11:39
ttxsoren: maybe a sideeffect of the outage11:39
*** zigo-_- has joined #openstack11:40
* ttx tags11:40
sandywalsh_interesting discussion of the AGPL
*** ctennis has joined #openstack11:48
*** ctennis has joined #openstack11:48
*** gaveen has joined #openstack11:56
*** gaveen has joined #openstack11:56
*** adiantum has joined #openstack11:57
*** irahgel has joined #openstack11:59
sorenttx: /me investigates12:08
*** rds__ has joined #openstack12:12
*** patcoll has quit IRC12:15
*** h0cin has joined #openstack12:17
sorenttx: PPA builds on their way.12:25
sorenIf not done already.12:25
*** mramige has quit IRC12:25
ttxok, numbers coming up. Based on the title, so I might be wrong.12:28
ttx6 22 56 64 67 80 85 101 10212:29
ttx+ yours obviously, but I guess you upvoted that one already :)12:29
ttxmaybe 312:30
sandywalsh_question about lp merges *not* to lp:nova ... myself and dabo are working on a branch and issuing merge-props into our shared dev branch. If I final-approve these will LP handle the merge? Or does it need to be done manually by one of us?12:30
ttxwrong chan12:31
ttxsandywalsh_: no, you need to manually push. Or setup a jenkins/tarmac job to do it for you12:31
sandywalsh_ttx, k thanks again!12:34
doudeHi all, I use 2011.2~gamma2~bzr988-0ubuntu0ppa1~lucid1 with KVM on 2 compute nodes (A and B). I can migrate a instance from B to A but not from A to B. When it fails, I cannot see anything in log files (Nova or libvirt). I check my libvirt configuration on the two hosts and they seem similar. Can you help me ?12:36
n1md4# euca-authorize -P icmp -t -1:-1 default12:36
n1md4[Errno 111] Connection refused12:36
n1md4? :(12:36
*** adiantum has quit IRC12:37
doudeThe log of the two compute nodes :
*** dendrobates is now known as dendro-afk12:38
*** miclorb_ has quit IRC12:38
*** dendro-afk is now known as dendrobates12:39
*** adiantum has joined #openstack12:49
sandywalsh_I'm looking at all these docstring merge-props ... what's the rule on using " vs. ' ... is there a new rule in place?12:58
notmynamettx: done12:59
*** ppetraki has joined #openstack13:01
ttxnotmyname: cool. Will push RC as soon as a new tarball is out of the Jenkins mill.13:02
*** adjohn has joined #openstack13:03
*** zigo-_- has quit IRC13:05
*** mray has joined #openstack13:10
jaypipessandywalsh_: specifically for docstrings? I think """ only.13:11
sandywalsh_jaypipes, for example ... most is changing "foo" to 'foo'13:11
jaypipessandywalsh_: heh, I've been staying out of those reviews... I don't really care about the ' vs ", to be honest. It's not a top priority of mine ;)13:13
sandywalsh_jaypipes, 100% agree ... just curious13:13
jaypipessandywalsh_: I know of no documentation on what the preferred style is.13:13
sandywalsh_jaypipes, that's what I thought too ... personal preference.13:14
*** bcwaldon has joined #openstack13:15
*** matclayton has joined #openstack13:16
*** hadrian has joined #openstack13:16
sandywalsh_ttx is there a way we should mark the merge props that have enough approves, but waiting on diablo?13:17
sandywalsh_ttx to make searching easier later?13:17
openstackjenkinsProject swift build #238: SUCCESS in 32 sec:
openstackjenkinsTarmac: Bump version to 1.3-rc13:17
ttxsandywalsh_: not really, you can comment on them that they are good to go... We'll go through them as part of the branch rollover13:17
ttxsandywalsh_: maybe set a review type (like "diablo") so that it shows at the top13:18
*** sideone has joined #openstack13:18
*** sideone has left #openstack13:19
dabosandywalsh_: For docstrings, """ is an almost universal standard. For normal string quoting, either is fine, but consistency is important.13:19
*** irahgel1 has joined #openstack13:19
*** irahgel has quit IRC13:19
dabosandywalsh_: most people use the single quote for strings, but I think that's because they are too lazy to hit the shift key :)13:19
sandywalsh_dabo, yeah, I'm cool with the docstring """ ... it was the other change that confused me13:19
sandywalsh_seemed like a lot of effort for no clear reason13:20
*** adjohn has quit IRC13:21
dabomixing quote styles is just one of those things that makes code in a language that prides itself on good-looking, readable code a little less aesthetically pleasing13:22
ttxSwift RC out at
shermanboydI'm trying to put together a Openstack proof of concept cluster, should I start with the Cactus RC?13:25
ttxshermanboyd: pending acts of god, it should be what the release will be.13:25
*** patcoll has joined #openstack13:25
*** pandemicsyn has quit IRC13:26
*** pandemicsyn has joined #openstack13:26
*** ChanServ sets mode: +v pandemicsyn13:26
*** notmyname has quit IRC13:26
*** jheiss has quit IRC13:27
*** notmyname has joined #openstack13:27
*** ChanServ sets mode: +v notmyname13:27
shermanboydI'm having trouble understanding the overarching architecture, I'm finding lot's of component level documentation, but not a good doc explaining what each component is and how it works with the others ... is there a doc like that out there?13:28
*** f4m8 is now known as f4m8_13:30
*** jheiss has joined #openstack13:33
*** Guest68283 is now known as ianweller13:34
*** ianweller has joined #openstack13:34
sandywalsh_shermanboyd, the intro to my zone spec might be of use? (not the latter parts, since it's a spec)13:35
shermanboydsandywalsh_: thanks I'll take a look13:35
shermanboydon a different note, I have a friend who says he accidentally "brought down" rackspace cloud by spinning up 500 instances at once.  I'm wondering if openstacks architecture would handle that better ...13:38
*** Zangetsue has joined #openstack13:39
sandywalsh_shermanboyd, heh ... that's the plan13:39
blamarshermanboyd: That could be a good thing to shoot for...but I'm not sure it's quite there yet.13:39
shermanboydI'm working on a 4 node cluster13:40
sandywalsh_shermanboyd, node = compute node?13:40
*** dspano has joined #openstack13:40
shermanboydyeah but they will have to do double duty13:40
*** dspano has quit IRC13:40
sandywalsh_shermanboyd, yeah, most do :)13:41
shermanboydsandywalsh_:  I have 4 servers w/ 64G ram and 3TB drive arrays13:41
*** santhosh has joined #openstack13:41
sandywalsh_which hypervisor?13:41
shermanboydsandywalsh_: I like KVM13:41
shermanboydsandywalsh_: but if cactus supports container virtualization I'd like to play with that too13:42
sorenIt does.13:42
*** smeier00 has joined #openstack13:42
*** jero has joined #openstack13:43
rsaidanis anyone here using postgres database? I seem to be getting programming errors, is I switch to mysql and restart nova-* everything works fine.13:43
rsaidanthe errors i'm seeing are:13:43
rsaidan(nova): TRACE: ProgrammingError: (ProgrammingError) operator does not exist: character varying = integer13:43
rsaidan(nova): TRACE: LINE 2: ...AS instance_types_1 ON instances.instance_type_id = instance...13:43
ttxnotmyname: I prepared for you -- please accept at the very end of your day, when you're OK for release.13:43
*** jero has left #openstack13:43
rsaidan TRACE: HINT:  No operator matches the given name and argument type(s). You might need to add explicit type casts.13:43
notmynamettx: will do13:44
ttxjaypipes: I prepared for you -- please accept at the very end of your day, when/if you're OK for release.13:44
ttxvishy: I prepared for you -- please accept at the very end of your day, when/if you're OK for release.13:45
* ttx can't wait for more projects to join to continue copypasting stuff.13:45
jaypipesttx: okey dokey.13:45
*** zaitcev has joined #openstack13:45
*** kashyap_ has quit IRC13:46
*** kashyap has joined #openstack13:46
ttxnotmyname, jaypipes, vishy: in the mean time, don't hesitate to contribute any known bug that users should be aware of to
ttxplease also check that the feature description makes sense :)13:47
jaypipesttx: will do.13:48
*** rchavik has quit IRC13:48
*** smeier00 has quit IRC13:54
*** ryker has joined #openstack13:56
*** jero has joined #openstack13:58
*** jero has left #openstack13:59
*** kashyap has quit IRC14:00
*** ryker has quit IRC14:00
*** ryker has joined #openstack14:01
shermanboydsandywalsh_: I like your proposal, I think rackspace does something like this, they call them puddles14:03
*** j05h has quit IRC14:03
*** nwl has left #openstack14:04
*** j05h has joined #openstack14:04
*** Shentonfreude has joined #openstack14:06
*** mdomsch has joined #openstack14:07
*** jkoelker has joined #openstack14:08
sandywalsh_shermanboyd, huddles :)14:09
*** smeier00 has joined #openstack14:09
blamaralthough I like the 'puddles' concept...14:09
shermanboydsandywalsh_: ahh hah hah14:09
*** zenmatt has joined #openstack14:10
*** Zangetsue has quit IRC14:10
*** Zangetsue has joined #openstack14:10
shermanboydI got that second hand from they guy that spun up 500 instances on rackspace14:10
*** antenagora has quit IRC14:11
shermanboydI've been calling my redundant hardware zones puddles ever since14:12
uvirtbotNew bug: #760717 in nova "Parameter mismatch when calling to_xml() from spawn() in" [Undecided,In progress]
*** dragondm has joined #openstack14:18
*** kashyap has joined #openstack14:23
*** dendrobates is now known as dendro-afk14:26
*** dendro-afk is now known as dendrobates14:28
*** ryker has quit IRC14:30
*** ryker has joined #openstack14:36
*** daveiw has left #openstack14:44
*** ramkrsna has quit IRC14:45
*** ramkrsna has joined #openstack14:45
*** mahadev has joined #openstack14:47
*** smeier00 has quit IRC14:48
*** zigo-_- has joined #openstack14:49
*** j05h has quit IRC14:50
*** j05h has joined #openstack14:50
*** dendrobates is now known as dendro-afk14:53
*** clayg_ has joined #openstack14:54
*** smeier00 has joined #openstack14:54
*** dodeeric1 has left #openstack14:55
*** mahadev has quit IRC14:59
*** j05h has quit IRC14:59
*** santhosh has left #openstack15:01
*** santhosh has joined #openstack15:01
*** xavicampa has joined #openstack15:02
*** j05h has joined #openstack15:02
*** mahadev has joined #openstack15:06
*** guigui has quit IRC15:07
* soren wanders off15:07
shermanboydis there anything I should know when installing the Ubuntu 10.4.2 LTS in preparation for Openstack?  What filesystem should I use, anything I should know about partition layout etc?15:13
*** dspano has joined #openstack15:20
*** iammartian has joined #openstack15:25
*** mray has quit IRC15:25
*** enigma1 has joined #openstack15:27
doudeHi, I've got an allocated floating IP address associated to an instance and if I've release it before disassociate it, the floating IP stills associated to the instance and not listed by describe address command. Do you know this bug ?15:30
*** ccustine has joined #openstack15:30
*** kbringard has joined #openstack15:32
*** mahadev has quit IRC15:32
BK_mandoude: hi. my pals are playing a lot with floating IPs.15:33
BK_mandoude: they are off for a dinner at the moment.15:33
BK_mandoude: ask adiantum for help for floating IPs15:33
kbringardI've done some work with floating ips, I may be able to help15:33
doudeBK_man: ok15:34
clayg_shermanboyd: you might do maverick instead of lucid... is this a development installation (i.e. "all in one") or a production node (compute/controller/volume/api/etc?)15:36
*** bkkrw has quit IRC15:38
*** dragondm has quit IRC15:45
*** dragondm has joined #openstack15:50
shermanboydclayg_: It's a little bit of both, I've got 4 servers and I expect they will each have to bear multiple roles15:51
shermanboydclayg_: I'm still trying to figure out exactly what the best strategy would be15:52
shermanboydclayg_: but the intention is to run production on these same servers15:53
*** ramkrsna has quit IRC15:54
*** sebastianstadil_ has joined #openstack15:54
*** chuck_ has joined #openstack15:54
*** mahadev has joined #openstack15:54
*** zul has quit IRC15:54
*** chuck_ is now known as zul15:54
*** zul has joined #openstack15:54
shermanboydclayg_: right now I'm thinking server1(compute,controller,volume,api)  server2(compute,volume)  server3(compute,volume)  server4(compute,volume)15:55
*** matclayton has left #openstack15:55
*** sebastianstadil has quit IRC15:58
*** sebastianstadil_ is now known as sebastianstadil15:58
clayg_shermanboyd: that's probably pretty good, no failover on the api server tho...15:58
clayg_the volume manager will want some disk carved out to be added to the nova-volume group, that's all I can think of on partitioning - filesystem should be whatever you're most comfortable with15:58
*** zigo-_- has quit IRC15:59
shermanboydclayg_: server2(compute,controller,volume,api)15:59
kbringardspeaking of API failover... I tried to setup the API on 2 backend servers and load balance them, but it didn't work because the signatures include the host you're connecting to in them16:00
kbringardso when you say api failover, do you mean manual failover?16:00
*** mihgen has quit IRC16:01
kbringardor has that been updated? (it's been awhile since I tested it)16:01
*** kakella has joined #openstack16:02
clayg_kbringard: oh idk I haven't personally tested a production nova cluster - I was just mapping over my experience with swift proxies.  If you can't lb the api servers - that's a real problem :\16:02
kbringardyea, like I said, the issue I had was that the signatures were generated based on the hostname of the LB16:03
shermanboydclayg_: how big can I expect the controller db to be?16:03
*** smeier00 has quit IRC16:03
*** kakella has left #openstack16:03
kbringardso when it'd send the connection back to the actual API server, the sig was wrong16:03
*** maplebed has joined #openstack16:03
*** kbringard_ has joined #openstack16:03
*** kbringard has quit IRC16:03
*** kbringard_ is now known as kbringard16:03
kbringardthere's probably a way around it by messing with hostnames and stuff though16:03
* ttx disappears for next 2 hours16:04
*** mahadev_ has joined #openstack16:04
*** mahadev has quit IRC16:04
*** MarkAtwood has joined #openstack16:07
shermanboydclayg_: also why would you go with maverick?16:07
*** dspano has quit IRC16:07
*** mdomsch has quit IRC16:09
clayg_meh, either is probably fine, but I've had better luck with newer releases.  For example the 10.04 python-routes package (1.11?) isn't compatible with trunk, but python-routes 1.14 is already included in maverick16:10
*** elambert has joined #openstack16:11
*** maplebed has quit IRC16:12
shermanboydclayg_: ok, I generally stick to to LTS for production , but openstack is still a fast moving target16:14
shermanboydI should just deploy on gentoo16:15
*** irahgel1 has quit IRC16:16
*** irahgel has joined #openstack16:16
xavicampahi! I'm setting up a private cloud across a few nodes. The ubuntu installation guide in openstack doc states you don't need nova-api nor nova-scheduler in the compute nodes, but what about nova-objectstore and rabbitmq-server?16:17
*** benbenhappy has joined #openstack16:23
*** rnirmal has joined #openstack16:24
*** smeier00 has joined #openstack16:25
*** mahadev_ has quit IRC16:26
*** dubsquared has joined #openstack16:27
*** crago has quit IRC16:28
xavicampaI assume the cloud controller requires all components to be running (a bit unsure about the need for nova-compute), while the compute node only requires nova-compute and nova-network. Can anybody confirm this is correct in a multiple server setup?16:28
kbringardcompute only requires compute16:29
*** mahadev has joined #openstack16:29
kbringardnetwork should only be running on the network controller16:29
kbringardsince it handles routing and dhcp16:29
kbringardxavicampa: you can have anything running anywhere, so long as it knows where the rabbitmq server is16:30
kbringardand of course the DB16:30
dubsquaredI just updated to the gamma code, and now my nodes aren't checking-in as expected…when i run nova-manage service list they are showing up as XXX… is there anything I need to do to 'reinit' these?16:31
kbringarddubsquared: don't think so... are the times synced?16:31
kbringardI had that problem before16:31
*** dendro-afk is now known as dendrobates16:31
dubsquaredi didn't check that...16:31
kbringardgenerally I run a quick ntpdate on all of them16:32
kbringardor whatever ntp server you want to use16:32
kbringardalso, did you update the code on the nodes as well?16:33
*** aliguori has quit IRC16:34
*** kakella has joined #openstack16:34
*** j05h has quit IRC16:34
*** j05h has joined #openstack16:34
*** mramige has joined #openstack16:36
xavicampakbringard, thanks. I'm trying to run the compute nodes as light as possible, to maximize the resources available to VMs, and at the same time I try to understand well the roles of the different components by installing them in different machines16:37
xavicampathat's why I'm try to find out what needs to be running where, and how many instances can be in the whole network setup16:37
dubsquaredheh, holy crap16:37
dubsquaredntpdate worked...16:37
dubsquaredive actually seen that said, but first ive ever ran into it myself16:38
kbringardno prob16:38
*** mdomsch has joined #openstack16:38
kbringardI'd probably make sure the ntpd is running, so the times stay synced16:38
dubsquaredyeah, just pushed down a ntp.conf to all the nodes16:39
kbringarddon't want clock drift taking your nodes offline, hehe16:39
*** kakella has left #openstack16:39
kbringardcool, hopefully that doesn't happen again then :-D16:40
*** irahgel has left #openstack16:52
shermanboydIs the openstack summit info here still accurate:
shermanboydLocation: Hyatt Santa Clara, CA16:54
shermanboydApril 26 - 29, 201116:54
*** j05h has quit IRC16:54
kbringardshermanboyd: that's where I'm going, so I hope so :-)16:54
shermanboyddo I need to register or anything to attend?16:55
*** j05h has joined #openstack16:55
kbringardyea, there is a registration link somewhere16:55
*** RickB17 has joined #openstack16:55
kbringardI'm looking for it16:56
blamarI think registration closed last week? But I'm not certain...16:56
*** smeier00 has joined #openstack16:56
*** mray has joined #openstack16:56
*** benbenhappy has left #openstack16:57
kbringardhmm, yea, it says it's closed16:57
kbringardsorry shermanboyd16:57
*** joearnold has joined #openstack16:58
vishyapt-get install ntp is safer17:01
*** mahadev has quit IRC17:03
*** maplebed has joined #openstack17:03
*** dragondm has quit IRC17:04
*** mahadev has joined #openstack17:05
*** aliguori has joined #openstack17:07
kbringardjeebs, it's like tornado windy outside17:09
*** smeier00 has quit IRC17:10
*** joearnold has quit IRC17:13
*** joearnold has joined #openstack17:13
*** hagarth has joined #openstack17:15
dubsquaredkbringard:  have you seen this? only happening on the nodes...17:23
*** ctennis has quit IRC17:23
*** smeier00 has joined #openstack17:23
dubsquaredand yes, virt is enabled in bios :D17:23
kbringardheh, uhm..17:24
kbringardwhat image store are you using?17:24
*** mahadev has quit IRC17:24
dubsquaredwith….not swift17:24
kbringardI recall having similar issues... although I don't recall it being that specific error17:25
kbringardbut I think what happened was that the images that were cached in base were broken17:25
uvirtbotNew bug: #760921 in nova "nbd device busy and inject key fails" [Undecided,New]
dubsquaredive been dealing with that far too much latelt17:26
kbringardlike, what had happened to me was17:26
dubsquaredlemme try to reup17:26
kbringardI had removed the images from glance17:26
kbringardthen uploaded them again17:26
kbringardor rather, I switched glance to mysql17:27
kbringardand reupped the images17:27
kbringardso there was an image name conflict17:27
dubsquaredi see17:27
kbringardso it was trying to load image 1, which it found in _base, but it was the wrong image, or something like that17:27
dubsquaredyeah, i ran into a similar problem ...17:27
dubsquaredbut my instances boot on one server17:28
dubsquared(where api, network, etc live)17:28
dubsquaredjust not on the nodes that just have compute only17:28
kbringarddo they know where the DB and rabbitmq are?17:28
kpeppledubsquared: the specific error you are getting ( is trying to save iptables ...17:28
kbringardyea, maybe iptables isn't installed on your compute nodes?17:28
dubsquaredroot@colo08:~# dpkg -l | grep iptables17:29
dubsquaredii  iptables                        1.4.4-2ubuntu2                                  administration tools for packet filtering an17:29
dubsquaredand yes, rabbit and sql_conn are set in nova.conf17:29
kbringardcan you run iptables-save manually on that compute node?17:30
*** ctennis has joined #openstack17:30
*** ctennis has joined #openstack17:30
kbringardmaybe a bug... sorry, I've not seen that particular error :-/17:31
kpeppledubsquared: it's catching an exception on "sudo iptables-save -t filter"  ... long shot -- did you update the code on all of the nodes ? I've seen that type of error when the DB gets updated but the nova-compute doesn't17:31
*** MarkAtwood has quit IRC17:32
*** mahadev has joined #openstack17:32
dubsquaredYeah, code is uptodate on all the nodes, and the 'controller'17:32
kpeppledubsquared: so other possibility is the eventlet ... is it up to 0.9.12 ?17:33
dubsquared 0.9.12-0ubuntu2~lucid117:34
dubsquaredhmm, im going to get some food for my brain…I'll be back in a bit, and maybe it will be magically fixed when i come back, or it will come to me while i am stuffing my face17:35
dubsquaredthanks for the help guys, much appreciated! bbiab!17:36
*** jero has joined #openstack17:36
*** jero has left #openstack17:36
*** ryker1 has joined #openstack17:36
*** smeier00 has quit IRC17:36
*** ryker has quit IRC17:39
*** kbringard has quit IRC17:40
*** kbringard has joined #openstack17:41
*** Ryan_Lane has joined #openstack17:41
kpeppledubsquared: i think you've run into the eventlet bug because you are using the stock eventlet ... not the patched one from our ppa. see
*** CoreyM has quit IRC17:47
*** ryker1 has quit IRC17:55
*** ryker has joined #openstack17:56
*** mdomsch has quit IRC17:59
*** enigma2 has joined #openstack18:00
notmynamettx: got some questions for ya about the summit18:00
*** mahadev_ has joined #openstack18:01
*** mahadev has quit IRC18:01
*** dinnerjacket has joined #openstack18:03
*** enigma1 has quit IRC18:03
dinnerjackethey guys, I just upgraded an installation of trunk bzr655 to trunk_safe bzr641, and I'm wondering about that new 'zones' table in the DB18:05
*** MarkAtwood has joined #openstack18:05
dinnerjacketit's currently empty; anythig I need to put there?18:05
*** nacx has quit IRC18:10
*** ryker has quit IRC18:12
kpeppledinnerjacket: look at doc/source/devref/zones.rst ... but AFAIK nothing unless you want to have multiple availability zones18:12
dinnerjacketkpepple: ok, thanks, I'll check that out.  currently having an issue with some instances 404'ing on the .../placement/availability-zones URL during init, and thought that might be worth looking at18:14
*** mramige has quit IRC18:15
*** pquerna has quit IRC18:17
*** pquerna has joined #openstack18:17
notmynamettx: are the sessions only to be about future development or should we include sessions on teaching/lessons-learned/lightling-talks/etc?18:18
*** lionel has quit IRC18:19
*** KnightHacker has joined #openstack18:20
*** lionel has joined #openstack18:20
*** mahadev_ has quit IRC18:21
*** mahadev has joined #openstack18:21
ttxnotmyname: theoretically the design summit sessions are about future features... we have lightning talks every day18:23
ttxnotmyname: the "conference" had a technical track for more classic presentations.18:24
ttxnotmyname: that said, if there are holes in the schedule, why not18:24
ttxnotmyname: last time we did tutorials/demos on the empty slots on the last day.18:25
ttxnotmyname: I have no idea how full the schedule will be, we should have more clue mid-next-week.18:26
ttxnotmyname: no need to submit a blueprint for those, we'll add them statically in the schedule if there is room.18:26
ttxnotmyname: makes sense ?18:27
*** blamar has quit IRC18:28
gholtttx, notmyname: Coolio, I was actually the one bugging him about that, hehe.18:28
gholtMaybe not this summit, but at some point I think cool "teaching" sessions would be good.18:28
notmynamettx: ya, sorry. got pulled away for a moment18:28
gholtI know that last time we had installfest type things going on, which was cool.18:29
notmynamettx: we were thinking of teaching sessions of both cool stuff in the code (like the rings in swift) and for lessons-learned type stuff (like things RAX has learned trying to make swift scale)18:29
ttxnotmyname: understood. The 'technical track' of the conference was supposed to give us room for those kind of sessions... but there were not so many slots18:30
ttxand maybe they weren't advertised to developers as much as they should have.18:31
*** MarkAtwood has quit IRC18:33
notmynamettx: ok. just trying to figure out which sessions I'm going to be adding for swift. Unless I hear differently from you, I will not add any teaching/lessons-learned sessions18:37
ttxnotmyname: maybe I should set up some etherpad or wiki page to track proposals18:38
ttxnotmyname: let me do that.18:39
notmynameforgive my ignorance, but isn't that what blueprints are supposed to do?18:39
notmynamemake the blueprints for all the proposed things, and then select the ones for the schedule18:39
ttxnotmyname: oh, I meant for teaching / lessons-learned ines18:40
notmynamethose aren't managed with blueprints too?18:40
ttxnotmyname: hm... we could.18:40
notmynameI'm not necessarily a fan of lp or blueprints, but it seems to make sense to use only one tool18:41
ttxnotmyname: these teaching/lessons-learned sessions are a new concept18:41
*** blamar has joined #openstack18:42
ttxnotmyname: I'll give it a thought and keep you posted18:42
ttxmaybe we can file regular BPs and set implementation to "informational"18:42
ttxjust need to check the summit scheduling would not ignore them.18:42
notmynameas openstack gets more mature, I can imagine that the summits can become a place for both planning and info dissemination. like a pycon but just for openstack :-)18:44
btorchso nova uses xen-api-server or xend-http-server ?18:44
medberrynotmyname, I think that that should be distinct from the summit. There's room for more than one meeting. If (and I'm speculating) OpenStack drops to two releases/year and summits drop to twice a year, there is probably planning/person power to handle an additional users conference (possibly in conjunction but possibly at another time)18:47
*** rsaidan has quit IRC18:49
*** arun_ has quit IRC18:49
*** photron has joined #openstack18:52
*** xavicampa has quit IRC18:54
*** MarkAtwood has joined #openstack19:01
uvirtbotNew bug: #761014 in nova "X-Server-Management-Url is filled from http header" [Undecided,New]
*** Ryan_Lane is now known as Ryan_Lane|food19:07
*** joearnold has quit IRC19:08
*** joearnold has joined #openstack19:09
*** vernhart has joined #openstack19:09
*** winston-d has quit IRC19:22
*** gaveen has quit IRC19:24
*** mgoldmann has quit IRC19:26
*** enigma2 has quit IRC19:27
*** sebastianstadil has quit IRC19:28
*** dinnerjacket has quit IRC19:30
*** enigma1 has joined #openstack19:33
_vinayI am using the EC2 api to run instances19:33
_vinayit returns a type of boto.ec2.instance.Reservation19:34
_vinaymy question is how to get the instance_id from this reservation ?19:34
kbringardare you writing your own tools, or using the euca2ools?19:34
_vinaywriting my own tool19:34
*** mdomsch has joined #openstack19:35
kbringardI'd look at the source of the euca2ools then19:35
kbringardsee how ecua-describe-instances works19:35
_vinayI could look at that19:35
kbringardor, if you can dump the data structure of the boto object that is returned19:35
kbringardthen determine the path to the instance id19:35
_vinayI did that (dump) :)19:35
_vinayit seems to be a simple string19:36
kbringardI want to say it's probably XML?19:36
_vinayso I found it weird that the api will require me to do another round trip19:36
_vinaythis is the output:19:36
_vinaytype: <class 'boto.ec2.instance.Reservation'> Reservation:r-f4o8prwk19:36
_vinayinst = self.conn.run_instances(u'ami-3f3lc0fb')         print >> sys.__stdout__, 'type: %s' % (type(inst))         self.pp.pprint(inst)19:37
kbringardyea, sorry, I don't know19:38
_vinayanyways will dig deeper into the Reservation class...19:38
_vinayno issues :)19:38
kbringardI've not done any coding on my own19:38
kbringardto that level19:38
kbringardare you using the boto classes?19:38
*** bkkrw has joined #openstack19:38
_vinayok thats fine.. just thought someone on this forum might know off the top of his head... ;-)19:38
_vinayyes ... boto classes19:39
kbringardI would have thought that they'd have all the logic for pulling the correct info out of the API19:39
kbringardbut again, I really have no idea19:39
_vinayI thought so too...19:39
_vinayno probs ... thanks again.. will look into source code of class ...19:40
kbringardnp, good luck19:40
kbringardI'd be interested to hear what you find19:40
*** HouseAway is now known as AimanA19:40
*** kashyap has quit IRC19:43
*** rsaidan has joined #openstack19:45
*** sebastianstadil has joined #openstack19:47
*** MarkAtwood has quit IRC19:47
*** sebastianstadil has quit IRC19:55
*** Ryan_Lane|food is now known as Ryan_Lane19:55
photron_vinay: in case you didn't figure it out yet, a reservation has a instances member that is a list of boto.ec2.instance.Instance objects. Instance has an id member, that's you instance id19:58
*** sebastianstadil has joined #openstack19:59
*** maplebed has quit IRC20:00
*** adiantum has quit IRC20:00
*** maplebed has joined #openstack20:00
*** dendrobates is now known as dendro-afk20:02
photron_vinay: the official api documentation of boto is pretty sparse and doesn't list possible properties of objects, it just documents the methods. giving the api user a hard time using it20:02
markwashanyone from ntt on here? I have some questions about nova.db.api.instance_get_fixed_address_v620:03
*** clauden has joined #openstack20:05
*** Shentonfreude has quit IRC20:08
*** jonkelly has joined #openstack20:13
*** arun_ has joined #openstack20:13
*** dendro-afk is now known as dendrobates20:13
*** Shentonfreude has joined #openstack20:14
Ryan_Laneif I run more than one network node, will each network node pick up the NAT changes from the queue, or will they interfere with one another?20:18
*** MarkAtwood has joined #openstack20:21
*** MarkAtwood has left #openstack20:21
*** hggdh has quit IRC20:24
*** joearnold has quit IRC20:25
*** hggdh has joined #openstack20:25
*** mihgen_ has joined #openstack20:27
*** dweimer has quit IRC20:28
*** dweimer has joined #openstack20:28
*** photron has quit IRC20:36
*** allsystemsarego has quit IRC20:37
*** mahadev has quit IRC20:37
*** sebastianstadil has quit IRC20:55
*** joearnold has joined #openstack20:58
daboAm I missing something? How are compute nodes supposed to "register" themselves so that their presence and capabilities are reflected in the compute_nodes table?21:00
*** h0cin has quit IRC21:03
*** ctennis has quit IRC21:03
dubsquaredi just check it with nova-manage service list, or check in mysql21:03
dubsquarednot sure about the compute_nodes table angle21:04
Ryan_LaneI've just imagined a fair amount of "black magic" when it comes to that service list21:04
daboRyan_Lane: I had assumed that the same "black magic" applied to the compute_nodes table, too, but mine are empty even with several nodes present.21:05
dabodubsquared: yeah, but I need the capabilities in the compute_nodes columns for the scheduler. I had assumed that they were auto-magically populated.21:06
dubsquaredthere is lot that nova doesn't do automagically D:21:07
dubsquaredbut it should, and will!21:07
Ryan_Lanewhich brings me to a question I asked earlier… :)21:07
dabodubsquared: ok, thanks. Guess I have to provide the magic21:07
Ryan_Laneif I run more than one network node, will each network node pick up the NAT changes from the queue, or will they interfere with one another?21:07
* Ryan_Lane wants to run more than one network node21:08
daboRyan_Lane: not sure - I haven't been digging into the network code, just the compute21:08
Ryan_LaneI guess I can just test it21:08
dabothat might work. :)21:08
dubsquaredwith the state of the networking in nova today… i wish you the best of work21:09
Ryan_LaneI want to have static NAT mapping on each node21:09
Ryan_Laneand set the gateway as my router21:09
Ryan_Laneand have each compute node also run the network node21:09
Ryan_Laneif they all get the NAT mappings, and the DHCP mappings, it should work perfectly21:09
Ryan_Laneif they pick up and remove items from the queue, it will work very poorly :)21:10
*** dendrobates is now known as dendro-afk21:10
Ryan_Lanethis way the network-node won't be a single point of failure for the instances21:10
dabodepends how the queue is configured21:10
Ryan_Laneyeah. should be easy to test I guess21:11
dabowe had to add fanout queues for multi-zone21:11
*** smeier00 has joined #openstack21:11
Ryan_LaneI may also add pybal support for BGP load balancing21:11
kbringardi totally read that as paypal support21:12
Ryan_Lanedefinitely not paypal :)21:12
kbringardgood call21:12
Ryan_Laneintegrating evil into a project isn't good21:12
kbringardsomeone needs to tell google that21:13
deepyThere's no 'evil' in Python21:15
kbringardanyone seen this error before?21:16
kbringardError: operation failed: failed to retrieve chardev info in qemu with 'info chardev'21:16
*** ctennis has joined #openstack21:17
*** ctennis has joined #openstack21:17
kbringardit's only happening with one image... and I can launch this particular image on the same host that has glance running, but when it tries to go to a compute node that isn't the glance host, it fails with that error21:17
kbringardI'm running ppa bzr981 on all my nodes21:18
kbringardthat's the whole trace21:19
*** OpenStack083 has joined #openstack21:30
n1md4kbringard: Yes, I've had this.21:31
*** JuanPerez has quit IRC21:31
kbringardn1md4: any ideas what you did to fix it?21:31
n1md4I could only solve this by crudely scp instances/base_ to each node.21:31
kbringardlast time this happened, it was because the images in _base didn't match21:31
OpenStack083can someone tell me how to delete user account in SWIFT21:31
kbringardn1md4: ok, that makes sense then21:31
*** mdomsch has quit IRC21:32
kbringardanyone know if there is any harm in removing images from _base/ while there are instances running on the compute host?21:32
*** adiantum has joined #openstack21:32
*** smeier00 has left #openstack21:34
n1md4kbringard: I21:35
*** OpenStack083 has left #openstack21:35
n1md4kbringard: I have 1 node and have just moved the _base file away from instances, and it's still up.21:35
n1md4sorry, 1 instance.21:36
kbringardok, cool21:36
kbringardI'll probably just blast everything in _base21:36
kbringardand try launching instances again21:36
kbringardthanks n1md421:36
n1md4I'm nost sure what would happen if I tried to launch another instance now though.21:36
n1md4I'll try for you if you like?21:36
kbringardin theory it grabs the image from glance21:36
kbringardor well, from the image store21:36
kbringardand plops it in _base21:36
kbringardeh, no worries, but I do appreciate it21:37
n1md4is that object-store?21:37
n1md4or nova-objectsore even.  Why do images not travel well, with a vanilla openstack install?21:38
n1md4..across multiple nodes.21:38
*** bkkrw has quit IRC21:39
*** mihgen_ has quit IRC21:40
*** ppetraki has quit IRC21:42
kbringardI'm just going to rsync _base from the working node over and see if that resolves it21:46
kbringardthanks again n1md421:46
*** romain_lenglet has joined #openstack21:50
*** romain_lenglet has left #openstack21:50
*** dendro-afk is now known as dendrobates21:52
*** bkkrw has joined #openstack21:53
*** stewart has joined #openstack21:53
*** sebastianstadil has joined #openstack21:54
*** gondoi has quit IRC21:55
*** rds__ has quit IRC21:57
*** adjohn has joined #openstack21:58
n1md4kbringard: no worries.  I'm still learning myself, I'm finding the curve steep too.22:01
kbringardhmmm, yea, same error :-/22:02
kbringardbut it's only this one image22:02
*** patcoll has quit IRC22:02
n1md4tried rsyncing from where the first image was published?22:04
n1md4Oh, you're using different images?22:04
kbringardso, I have a bunch of images22:04
kbringardand they all work22:04
kbringardexcept this one image... this one image works on the compute node that's on the same physical host as my glance api/registry22:05
kbringardbut not on any other compute nodes22:05
kbringardall other images so far work on all the other compute nodes22:05
kbringardso, I deleted everything in _base on one of the compute nodes where it's failing22:05
kbringardand then rsync'd _base from the compute node where it works22:05
*** aliguori has quit IRC22:06
kbringardbut I'm still getting the same error22:06
*** NovSlave has joined #openstack22:06
kbringardrunning the same compute code on all the machines22:06
kbringardit's very bizarre22:06
*** elambert has quit IRC22:06
kbringardsame version of qemu22:06
kbringardsame version of qemu-nbd22:06
kbringardI would just say the image is messed up, except that it works perfectly fine on that one compute node22:06
kbringardit's a head scratcher to be sure22:07
NovSlaveI ve got something really odd here if someone could help ! my nova instances status is always NETWORKING ! T_T22:07
*** mray has quit IRC22:08
*** adjohn has quit IRC22:08
n1md4kbringard: Changing the subject, have you tried working with multiple users and projects?22:08
*** rnirmal has quit IRC22:09
kbringardI have22:09
n1md4NovSlave: Assuming you're running a single controller/node, check /var/log/nova/nova-compute.22:09
n1md4hmm I keep getting wrapped up in credential problems.22:10
NovSlavethats another odd thing the log file is EMPTY :x22:11
kbringardwe're using some custom single signon code, that uses our company SSO to auth22:11
kbringardand if there is no user it creates creds for them22:11
kbringardand roles22:11
kbringardbut in truth I don'22:11
kbringardI have not delved into that code much22:11
*** adjohn has joined #openstack22:11
n1md4NovSlave: Sounds like you have something pretty major going wrong here.  Are you following a guide?22:12
*** bcwaldon has quit IRC22:13
*** lstoll has quit IRC22:13
*** vishy has quit IRC22:14
*** vishy has joined #openstack22:14
NovSlavethe main tutorial yeah , in my first installation everythig went fine , the day after i couldn t ping or ssh my instances so assuming it was a networking issue ,i changed the nova networking manager (i tried both flatdhcp and flanetwork after using the default Vlan mode)22:14
*** matiu has joined #openstack22:14
*** lstoll has joined #openstack22:14
NovSlavethan i tried different images22:15
NovSlavethe status became NETWORKING22:15
*** clayg_ has left #openstack22:15
kbringardNovSlave: what does your nova-network.log say?22:15
n1md4NovSlave: Do you have anything in /var/log/nova/ ?22:16
*** KnightHacker has quit IRC22:16
*** JordanRinke has quit IRC22:16
*** KnightHacker has joined #openstack22:16
*** JordanRinke has joined #openstack22:16
*** Xbss has joined #openstack22:16
Xbssbut i might tell you what my nova-network.log says22:17
Xbssif i only get access to the machine right now22:17
Xbssi ll try saving all my logs tomorrow morning22:17
kbringardI need to he head out, thanks for the attention n1md422:18
kbringardgood luck NovSlave22:18
*** kbringard has quit IRC22:18
Xbssthx i already installed Eucalyptus and openNebula but damn openstack-nova is killing me xD22:19
*** NovSlave has quit IRC22:19
*** Xbss has left #openstack22:19
*** sebastianstadil has quit IRC22:20
*** bkkrw has quit IRC22:31
*** Shentonfreude has quit IRC22:32
*** dubsquared has quit IRC22:36
*** JuanPerez has joined #openstack22:40
*** elambert has joined #openstack22:40
*** grapex has joined #openstack22:42
*** stewart has quit IRC22:42
*** grapex has quit IRC22:50
*** enigma1 has quit IRC22:51
*** grapex has joined #openstack22:54
*** grapex has quit IRC22:58
*** santhosh has quit IRC22:59
*** Zangetsue has quit IRC23:01
*** grapex has joined #openstack23:02
*** jonkelly has quit IRC23:04
*** zigo-_- has joined #openstack23:05
*** n1md4 has quit IRC23:08
*** matiu has quit IRC23:09
*** jkoelker has quit IRC23:10
*** adjohn has quit IRC23:12
*** n1md4 has joined #openstack23:14
*** matiu has joined #openstack23:15
*** dendrobates is now known as dendro-afk23:22
*** grapex has quit IRC23:23
*** joearnold has quit IRC23:32
*** adjohn has joined #openstack23:36
*** nelson has quit IRC23:45
*** Zangetsue has joined #openstack23:45
*** nelson has joined #openstack23:45
*** bcwaldon has joined #openstack23:49
*** adjohn has quit IRC23:51
*** KnightHacker has quit IRC23:51
*** rsaidan has quit IRC23:53
*** KnightHacker has joined #openstack23:53
*** JuanPerez has quit IRC23:56
*** maplebed has quit IRC23:57
*** clauden has quit IRC23:57
*** rchavik has joined #openstack23:58
*** jeffjapan has joined #openstack23:58

Generated by 2.14.0 by Marius Gedminas - find it at!