Thursday, 2011-02-17

*** dendro-afk is now known as dendrobates00:03
*** lvaughn_ has quit IRC00:05
uvirtbotNew bug: #720459 in glance "Remove validation on image_type" [Medium,In progress]
*** adiantum has quit IRC00:07
*** lvaughn has joined #openstack00:08
*** Gh0sty has quit IRC00:12
*** adiantum has joined #openstack00:12
*** kashyapc has quit IRC00:18
*** kashyapc has joined #openstack00:19
*** dendrobates is now known as dendro-afk00:21
*** eikke has quit IRC00:26
*** desai has joined #openstack00:26
openstackhudsonProject nova build #526: SUCCESS in 1 min 32 sec:
openstackhudsonTarmac: Fixes ldapdriver so that it works properly with admin client.  It now sanitizes all unicode data to strings before passing it into ldap driver.  This may need to be rethought to work properly for internationalization.00:27
*** drico has quit IRC00:28
*** kashyapc has quit IRC00:29
*** husimon has quit IRC00:30
*** ppetraki has quit IRC00:31
*** husimon has joined #openstack00:33
*** adiantum has quit IRC00:38
*** adiantum has joined #openstack00:43
*** kashyapc has joined #openstack00:43
*** sebastianstadil has quit IRC00:45
*** pvo has joined #openstack00:46
*** dragondm has quit IRC00:48
*** sebastianstadil has joined #openstack00:53
*** pvo has quit IRC00:53
*** adiantum has quit IRC00:53
*** adiantum has joined #openstack00:59
*** joearnold has quit IRC00:59
*** Ryan_Lane|dc has quit IRC00:59
*** vbn has joined #openstack00:59
*** adjohn has joined #openstack01:07
*** westmaas1 has joined #openstack01:16
*** aimon has joined #openstack01:17
*** clauden has quit IRC01:18
*** desai has quit IRC01:21
*** burris has quit IRC01:22
*** miclorb has joined #openstack01:25
*** maplebed has quit IRC01:30
*** desai has joined #openstack01:34
nelsoncreiht: I have plausible code to open a connection, return an object, which you can call obj.write(data) and then obj.close(data).01:36
nelsonUnfortunately, it's not fooling the proxy into actually WORKING. will have to drop back and punt.01:36
nelsonTomorrow, tomorrow, it will love me tomorrow, tomorrow is always a day away.....01:37
*** lvaughn has quit IRC01:39
*** lvaughn has joined #openstack01:40
*** lvaughn_ has joined #openstack01:43
*** lvaughn has quit IRC01:46
*** aimon has quit IRC01:53
*** dweimer has joined #openstack01:55
*** aimon has joined #openstack02:06
*** hadrian has quit IRC02:07
*** GasbaKid has quit IRC02:10
*** Brainspa1kle is now known as Brainspackle02:20
*** Brainspackle has joined #openstack02:21
*** Brainspackle is now known as viewlogic02:22
*** vbn has quit IRC02:23
*** jc_smith has quit IRC02:27
*** dendro-afk is now known as dendrobates02:42
*** dendrobates is now known as dendro-afk02:42
*** Pentheus has quit IRC02:43
*** pvo has joined #openstack02:45
*** dweimer has quit IRC02:49
*** dendro-afk is now known as dendrobates02:58
*** pvo has quit IRC03:03
*** asalkeld has joined #openstack03:18
*** alekibango has joined #openstack03:29
*** littleidea has quit IRC03:34
*** aimon has quit IRC03:40
*** pvo has joined #openstack03:42
*** adiantum has quit IRC03:44
*** adiantum has joined #openstack03:49
*** bird has joined #openstack03:57
*** resker has joined #openstack04:02
*** aimon has joined #openstack04:03
*** littleidea has joined #openstack04:14
*** sebastianstadil has quit IRC04:15
*** desai has quit IRC04:20
*** desai has joined #openstack04:21
*** joearnold has joined #openstack04:23
*** adiantum has quit IRC04:23
*** adiantum has joined #openstack04:28
*** desai has quit IRC04:53
*** kashyapc has quit IRC04:53
*** sebastianstadil has joined #openstack04:54
*** Ryan_Lane has joined #openstack04:55
*** sebastianstadil_ has joined #openstack04:56
*** sebastianstadil has quit IRC05:00
*** sebastianstadil_ is now known as sebastianstadil05:00
*** littleidea has quit IRC05:01
*** Ryan_Lane has quit IRC05:03
*** adjohn has quit IRC05:15
*** pvo has quit IRC05:18
*** omidhdl has joined #openstack05:20
*** adjohn has joined #openstack05:22
*** Pentheus has joined #openstack05:25
*** dendrobates is now known as dendro-afk05:27
*** kashyapc has joined #openstack05:30
*** clauden_ has joined #openstack05:31
*** ccustine has quit IRC05:31
*** resker has quit IRC05:35
*** adiantum has quit IRC05:38
*** f4m8_ is now known as f4m805:45
*** adiantum has joined #openstack05:45
*** clauden_ has quit IRC05:46
*** Code_Bleu has joined #openstack05:51
Code_Bleudoes openstack require a Intel VT or AMD SVM processor?05:52
*** joearnold has quit IRC06:02
alekibangoCode_Bleu: only for certain hypervisors...06:03
alekibangoit can run many of them , so answer is NO06:05
*** littleidea has joined #openstack06:37
*** exitdescription has joined #openstack07:08
*** Pentheus has quit IRC07:11
*** Pentheus has joined #openstack07:11
*** miclorb has quit IRC07:16
*** allsystemsarego has joined #openstack07:17
*** allsystemsarego has joined #openstack07:17
*** adiantum has quit IRC07:17
*** Pentheus has quit IRC07:19
*** Pentheus has joined #openstack07:19
*** CloudChris has quit IRC07:21
*** adiantum has joined #openstack07:23
*** thimble has joined #openstack07:25
*** omidhdl has quit IRC07:34
*** omidhdl has joined #openstack07:38
*** adiantum has quit IRC07:41
*** adiantum has joined #openstack07:47
*** CloudChris has joined #openstack07:51
*** omidhdl has quit IRC07:52
*** ramkrsna has joined #openstack07:56
*** ramkrsna has joined #openstack07:56
*** DigitalFlux has joined #openstack07:57
*** omidhdl has joined #openstack07:58
*** adiantum has quit IRC07:58
*** thimble has quit IRC08:00
*** jtimberman has quit IRC08:01
*** thimble has joined #openstack08:02
*** jtimberman has joined #openstack08:02
*** adiantum has joined #openstack08:04
*** skiold has joined #openstack08:15
*** aimon has quit IRC08:18
*** exitdescription has quit IRC08:20
*** exitdescription has joined #openstack08:20
*** exitdescription has quit IRC08:21
*** exitdesc1iption has joined #openstack08:25
*** miclorb has joined #openstack08:28
*** Nacx has joined #openstack08:39
*** rcc has joined #openstack08:39
*** omidhdl has quit IRC08:44
*** adiantum has quit IRC08:45
*** adiantum has joined #openstack08:48
*** omidhdl has joined #openstack08:49
*** Nacx has quit IRC08:52
*** ibarrera has joined #openstack08:52
*** MarcMorata has joined #openstack08:58
*** adiantum has quit IRC08:59
uvirtbotNew bug: #720589 in nova "nova.db.api.project_get_network has unused parameter." [Undecided,New]
*** adiantum has joined #openstack09:04
*** littleidea has quit IRC09:08
thimbleGood morning here. Is there any way to advise the scheduler to spawn num[x] instances at the same time? I cannot find any flag for this.09:11
dsockwellit should be part of the euca-run-instances command09:12
dsockwellso at the very least you can open up the python on that and see how it's handled09:13
dsockwellare you trying to script this?  or do you just want your N instances?09:14
*** calavera has joined #openstack09:14
thimbleno, thats not what I mean. When using -n in euca-run, the instances are spawned one after the other09:14
*** calavera has quit IRC09:14
dsockwellyou want them all brought up at the same moment?09:15
thimbleI have serveral nodes and I really have the hardware to spawn more than one instance per node.09:15
*** calavera has joined #openstack09:15
thimblelet´s say I´ll spawn 120 instances, I will spawn e.g. 5 instances per time on each node, the next schedule cycle should spawn the next 5 on each node and so on09:15
thimbleit seems like instance spawning is queued somehow, but I cannot find something in the source09:16
dsockwelli'm from eucalyptus land, honestly, but the way eucalyptus spawns instances is before it can spawn an instance, it has to create the virtual disk09:16
dsockwellso all the time that it takes to create an instance is taken up by the disk, making loop devices09:17
*** TheHutch has joined #openstack09:18
dsockwelli don't know of any way to do that faster other than to snapshot volumes, and i'm pretty sure that for integrity and compatibility reasons, openstack won't do that09:18
dsockwellso the delay you're experiencing is -probably- the writing of 5-10GB per instance onto your storage09:18
dsockwelldepending on the instance size09:18
dsockwelland afaik that's unavoidable09:19
dsockwellso unless i'm gravely mistaken, your nodes are spawning as fast as they can09:19
*** skiold has left #openstack09:20
*** skiold has joined #openstack09:20
thimblejust a moment, I´ll have a look at the timestamps when the compute node receives the cast command09:20
dsockwelloh, if it's waiting for one instance to finish before scheduling to the next node then that's a problem09:22
dsockwelland one that I don't have control over.  unfortunately it's still very early, it's likely anyone who matters is still asleep.09:23
thimblethat´s what I mean :)09:24
thimbleI´ve just reduced the setup to two nodes for easy watching and the compute-apis receive the commands with a very little time difference. But then they´re only processing one instance. There´s no advice to create a second on the same node at the same time.09:25
dsockwellit would be foolish to do that, the limitation is the write speed of your storage09:26
dsockwellif you do two at once it will take twice as long09:26
dsockwellcheck out iostat on the nodes and you'll see what i'm talking about09:27
*** ChrisAM has quit IRC09:29
*** mattinator has quit IRC09:29
*** mattinator has joined #openstack09:30
*** ChrisAM1 has joined #openstack09:32
*** ramkrsna has quit IRC09:37
*** perestrelka has quit IRC09:40
*** adiantum has quit IRC09:43
thimbledsockwell, you are absolutly right.. it takes 28 secondes to write the whole image we´re using. I thank you :)09:43
dsockwellmy pleasure.  by the way, what OS are you running openstack on?09:44
DigitalFluxGuys, # comments are allowed in nova.conf, correct ?09:45
*** omidhdl has left #openstack09:45
ttxsoren: Any chance you could set up a tarball job for nova-bexar on Jenkins ? I'd like that set up before we start approving the post-release BMPs09:46
sorenIn a minute.09:46
dsockwellI'm trying to install on Centos 5.5 as per .  Am I wrong to want to do this?09:46
sorenI'm fixing that libvirt bug. Almost done.09:46
sorendsockwell: ttx and I are both Ubuntu developers. I think our answer will be biased :)09:47
ttxdsockwell: it's not wrong, it'"s evil !09:47
dsockwelli'll admit, i'm biased, too.  ubuntu has historically ground my teeth on the desktop, so i'm reluctant to use it on a server.09:48
*** adiantum has joined #openstack09:48
DigitalFluxsoren: Thanks for the fosdem talk man .. nice presentation ;)09:48
dsockwellbut since it's just this one app i'm sure it can't hurt too much09:49
*** miclorb has quit IRC09:49
*** ramkrsna has joined #openstack09:51
sorenDigitalFlux: I'm glad you liked it :)09:51
DigitalFluxttx: about the Glance issue i was having09:53
DigitalFluxttx: I think i missed that there was an s3_host flag in nova.conf, this might have affected Glance not being used, but normally nova-api should not start or something when there is conflicting parameters in the config09:54
DigitalFluxttx: still didn't test the Glance scenario yet .. will tell you when i'm finished ..09:54
ttxDigitalFlux: ok09:54
DigitalFluxSo does # comments work in nova.conf ?09:55
ttxDigitalFlux: I think so, yes09:55
DigitalFluxOkay, lemme check it09:55
dsockwellso while i'm downloading the Ubuntu install media, if it's not a Waste of Everyone's Time, I'd like to try to build on centos.  I followed the guide I pasted, and downloaded nova-2011.1.tar.gz09:55
dsockwellbut when I go to run, this happens:09:56
dsockwell[root@roy nova-2011.1]# python26 ./ build09:56
dsockwellNo handlers could be found for logger "nova.root"09:56
dsockwell being the pre-build process09:56
dsockwelli don't suppose that's hopelessly out of date?09:57
DigitalFluxdsockwell: I think you should take a look at this
DigitalFluxdsockwell: Bexar release for RHEL 610:00
dsockwelloh, cool10:00
dsockwelli don't suppose there's a centos6 anywhere to match RHEL610:02
sorenttx: nova-bexar-tarball job created (and run for the first time). Enjoy.10:06
sorenttx: Crap, I shouldn't have run it.10:07
sorenttx: /me fixes10:07
dsockwellsince i'm going to burn that centos machine anyway, it will be fun to see what those packages do to it.10:08
*** adjohn has quit IRC10:09
sorenttx: Fixed.10:09
sorenttx: Now enjoy.10:09
ttxsoren: thanks!10:10
*** Jordandev has joined #openstack10:10
*** drico_ has joined #openstack10:13
*** drico has joined #openstack10:13
*** drico__ has joined #openstack10:13
*** drico has quit IRC10:13
*** miclorb has joined #openstack10:15
thimbledsockwell: we are running on SLES11SP1 acutally10:17
thimbleand please ignore any errors related to my keyboard ;)10:20
dsockwellit's OK, i've been drinking too10:20
dsockwelli already burned my RHEL6 trial, apparently10:21
dsockwelldon't remember that one10:21
thimblejust order a new one ;)10:22
dsockwellthen i'd have to falsify a bunch of stuff and I could just install ubuntu and wait for centos anyway10:22
dsockwelldo we have an idea of how painful it would be to move a cloud controller (and associated services) from an ubuntu host to RHEL/CentOS 6?10:24
thimbleAFAIR someone was working on rhel6 packages in this channel here. But I don´t rember the name10:28
dsockwellyes, i've seen them, i know i could install the software on the RHEL machine, but migrating the data would be the tough part.  i'm wondering if there's an official path for that.10:29
*** bird has quit IRC10:32
sorenThere is *no* such thing as cloud controller.10:32
sorenSo what do you actually mean?10:32
*** bird has joined #openstack10:32
dsockwellah -- i must have misread the schematic here:
dsockwelli suppose i mean the API server, is that something that exists?10:35
dsockwellsorry if i'm being dense.  what's the name for the computer that euca2ools talks to when everything is working?10:43
*** adiantum has quit IRC10:44
thimbleyou mean the host running the service: nova-api10:47
*** TheHutch has quit IRC10:47
dsockwellthank you10:47
sorenYeah, that's nova-api.10:48
sorenThis is not Eucalyptus :)10:48
ttxsoren: could you look into so that we get some testable tarball with the fixes in ?10:48
dsockwellperhaps I need some colored blocks to play with and work this stuff out =\10:49
*** adiantum has joined #openstack10:49
ttxsoren: The translations bmps should still land in cactus first, so maybe keep out of this one for now10:49
*** thimble has quit IRC10:50
sorenttx: Do we still just require 2 approves?10:50
sorenttx: For "SRU"'s?10:50
ttxsoren: I think we need to do this in two steps. One is getting a tarball built, the other is getting that tarball released. Think -proposed and -updates10:51
ttxOnce we have a candidate tarball i'll run test cases to reproduce bugs and check that they are fixed10:51
sorenIs that a "yes"?10:51
ttx"yes" :)10:52
sorenCool :)10:52
sorenI can't approve.10:52
ttxonly hudson can ?10:52
soren...and I'm listed as "community".10:52
* ttx looks10:52
*** miclorb has quit IRC10:52
sorenmtaylor: needs to fix this10:52
sorenmtaylor: By setting the review team for nova/bexar to nova-core.10:53
sorenmtaylor: Well, something other than hudson itself.10:53
ttxsoren: only mtaylor has the keys to ~openstack-hudson on LP ?10:56
sorenSo make sure he doesn't get run over by a train.10:57
ttxmtaylor: looks like we should increase the bus factor on that10:57
*** CloudChris has quit IRC10:58
*** CloudChris has joined #openstack10:58
dsockwelloh -- there's everything I wanted to know, in the documentation.  thanks for your patience, soren.11:01
ttxsoren: updated with what I think is the "next commit" you were mentioning, please doublecheck11:12
ttxas it seems a bit wider than just "apply the same to call()"11:12
*** adiantum has quit IRC11:13
*** adiantum has joined #openstack11:20
sorenttx: What I wanted you to do was to do the exact same to the call method that you did to the cast method. That would be a more contained patch than my "next commit".11:21
sorenttx: ...but this is fine.11:21
ttxsoren: oh, ok, I thought "copy my fix from a few commits later" :)11:22
ttxsoren: i prefer to have the same fix in cactus and bexar.11:22
ttxfor obvious "more people test trunk" reasons.11:23
DigitalFluxdsockwell: what do you mean by "migrating the data" ?11:23
*** Jordandev has quit IRC11:25
dsockwellI wanted to install nova on ubuntu now because that's what works, and then later when centos6 comes out, run it on that11:26
dsockwellwithout any data loss11:26
dsockwellsince i've found what I wanted in the documentation, I know that most of the volatile configuration stuff lives on a mysql database, and that should be easy to dump and restore, or run on centos to begin with11:28
dsockwellsince centos 5 runs databases just fine11:28
*** eikke has joined #openstack11:28
DigitalFluxdsockwell: Got it11:29
dsockwellso that leaves the object store (the machine images, etc), is there a procedure for moving those with a minimum of disruption to users?11:29
*** uksysadmin has joined #openstack11:30
dsockwellalso i'm sure I butchered the project's jargon there, forgive me11:30
DigitalFluxI have been working on a puppet module for quite sometime, if it's good enough it should be simple, modular so we can make the required modifications that every env need (addition of backups, abstracted monitoring and so on ) ..11:31
uksysadminhi all... I'm banging my nut against a brick wall here trying to set up a multi-node cluster under virtualbox (I like a challenge) and was wondering if someone can help11:31
DigitalFluxHowever i have been kind of lazy :)11:31
DigitalFluxBut the model should install Nova on Ubuntu/CentOS (once the packages are released of course) ..11:31
dsockwellwhich packages?11:32
uksysadminI want to set up floating IPs so I can access my instances from the host running vbox [my lappy].  how do I do this - the documentation really sucks.11:32
sorenttx: Yeah, that's sensible. Hence my "this is fine" :)11:32
uksysadminI've got a private network and I've an instance now running with vlan manage network and I can access this using from the compute node11:33
DigitalFluxuksysadmin: me needs answer to that too11:33
sorenWhat's the problem, exactly?11:34
DigitalFluxsoren: routing i guess, your instances network is different than that on the laptop you are using11:34
DigitalFluxsoren: so may be you can launch instances using eucatools and stuff like that .. but not ssh into them11:34
sorenThat's what floating IP's will let you do.11:35
sorenOr vpn.11:35
DigitalFluxAh sorry, that's not my situation, uksysadmin: does soren's answer solves your problem ?11:35
uksysadminI've got a vlan100 interface that is bridged to br100, but the logs say (quite rightly I guess): Command: sudo ip addr add dev vlan1  Stderr: 'Cannot find device "vlan1"\n'11:36
uksysadminNo - I can't even get the association to work11:36
sorenuksysadmin: You need to set the public_interface flag.11:36
sorenuksysadmin: Set it to (you guessed it) your host's public_interface.11:37
soren(This is a nova-network setting, by the way)11:37
uksysadmincheers @soren - looking in nova.conf I've not set that explicitly11:38
sorenvlan1 is an *exceptionally* poor default, though.11:40
*** h0cin has joined #openstack11:42
*** h0cin has joined #openstack11:42
*** adiantum has quit IRC11:45
* soren lunches11:45
*** adiantum has joined #openstack11:51
*** adiantum has quit IRC11:58
*** adiantum has joined #openstack12:04
uksysadminRight - that's now working but seems like a fudge - I set in /etc/nova/nova.conf: --public_interface=vmnet0 - I can ssh to an instance using the public floating ip now... but euca-describe-instances doesn't show this and in the logs I now get Error: IP leased that isn't associated12:04
uksysadminwhere vmnet0 is bridged and associated with the instance...12:05
*** metoikos has joined #openstack12:05
*** berendt has joined #openstack12:05
uksysadminweirdly, spinning up another instance, where it gets I can assign a seperate floating IP, but that gets vmnet1 associated to it [but an ifconfig on the compute host doesn't show this floating ip associated with it]12:05
uksysadminWhat's with the syntax of nova-manage floating create <host> <ip range>.... what's host?  no matter what I change this to it doesn't seem to affect anything.12:10
sorenIt's the name of the host where nova-network runs.12:11
uksysadminwhat's the purpose of it? is it so that relevant natting can occur in a multi-node environment?12:13
uksysadminthanks again though - got over this annoying little hurdle of having a mini multi-node test environment under vbox where I could only access the instances from the compute node.  at one point I managed to achieve what seemed like the impossible - access the instances from the private address from my host running vbox:  host->vbox->openstack->instance... but like anything that keeps it bods employed, I couldn't get it working a12:15
*** perestrelka has joined #openstack12:16
sorenThe current implementation of floating ip's assumes that a single host will act as the gateway for this traffic.12:16
sorenSo a floating ip belongs to a network host.12:17
uksysadminthis is important  as we're a vmware house hosting a large international website... I've got the task of making sense of openstack and how we can possibly replace vmware12:17
dsockwellehh -- starting with the automated install scripts now12:19
dsockwellsomething tells me i did a good job running apt through a caching proxy12:19
*** adiantum has quit IRC12:21
*** uksysadmin has quit IRC12:22
*** uksysadmin has joined #openstack12:24
*** uksysadmin has left #openstack12:25
*** uksysadmin has joined #openstack12:26
*** adiantum has joined #openstack12:27
*** desai has joined #openstack12:35
*** westmaas1 has left #openstack12:39
*** CloudChris has quit IRC12:43
*** adiantum has quit IRC12:43
*** CloudChris has joined #openstack12:43
*** adiantum has joined #openstack12:48
*** thimble has joined #openstack12:50
uvirtbotNew bug: #720685 in nova "vlan1 is a poor default choice for FLAGS.public_interface" [Undecided,New]
*** hadrian has joined #openstack13:08
*** johndoe_ has joined #openstack13:11
*** adiantum has quit IRC13:11
*** yt_ has joined #openstack13:11
DigitalFluxCool !13:15
DigitalFluxI see there is a libvirt.xml.template around the code13:15
DigitalFluxso if i want to edit the VM instance config, i can directly edit this file ?13:16
*** adiantum has joined #openstack13:17
ttxDigitalFlux: you should rather provide your own version of the file and pass it as --libvirt_xml_template13:18
*** drico_ has quit IRC13:18
*** drico_ has joined #openstack13:18
DigitalFluxttx: That's pretty awesome :)13:20
DigitalFluxttx: so on the compute node, i should just put libvirt_xml_template=/some/file in nova.conf ?13:21
DigitalFluxand that's it ?13:21
ttxDigitalFlux: never tried it, but looking at the code that should do it.13:22
DigitalFluxttx: will test it indeed13:23
*** ChrisAM1 is now known as ChrisAM13:24
*** johndoe_ has quit IRC13:30
*** adiantum has quit IRC13:35
*** eikke has quit IRC13:36
uksysadminJust want to say thanks again @soren - I've got a demo environment (single host) with a working vm instance in under 20 mins under virtualbox now.  Now onto multi-node clusterdom.13:38
*** Ryan_Lane has joined #openstack13:39
*** adiantum has joined #openstack13:42
*** maplebed has joined #openstack13:42
*** guigui has joined #openstack13:44
*** dprince has joined #openstack13:44
*** nelson has quit IRC13:46
*** nelson has joined #openstack13:47
*** Ryan_Lane has quit IRC13:47
*** maplebed has quit IRC13:48
*** CloudChris has quit IRC13:51
*** dendro-afk is now known as dendrobates13:51
uksysadminJust thinking about this - I can now set a "public" floating IP to my instance... am I doing something that's not necessary: i.e. my steps:  euca-run-instance ... && euca-associate-address - or is this how it is for the time being?13:53
*** maplebed has joined #openstack13:54
*** maplebed has quit IRC13:55
sorenuksysadmin: With the VlanManager model, that's how you do it (if you're not using vpn).13:55
uksysadminyeah not using vpn - I'll take a look, ta. and noted on the patch, will take a look too.13:56
*** adiantum has quit IRC13:57
*** eikke has joined #openstack13:59
uksysadminbtw - I tried setting eth0 in nova.conf as the public_interface and computer failed with an interface in use.  I've got --public_interface=vnet0  this was the only way I got floating IPs working.  Sure, eth0 is a better default though! ;-)14:00
*** adiantum has joined #openstack14:02
*** ppetraki has joined #openstack14:04
berendti used a string in a patch and i have to i18n it.. anybody knows how to do this?14:07
* ttx thinks jaypipes should whip up a reference wikipage about that14:09
* ttx wonders if Jenkins rejecting a branch shouldn't set it back to "Work in progress" instead of "Needs review" since in 90% of the cases it needs action from the proposer, rather than from reviewers.14:13
berendtI added _(...) now14:14
berendtI think nova.pot is changed automatically14:14
*** desai has quit IRC14:17
*** mray has joined #openstack14:18
*** ramkrsna has quit IRC14:20
*** adiantum has quit IRC14:22
*** CloudChris has joined #openstack14:24
wedjathi everyone :)14:27
*** adiantum has joined #openstack14:28
wedjati've noticed that in swift, when your nodes get some big files (1-4 Gbytes) the swift processes make the iowait on all the storage nodes increase really quickly14:28
wedjatdo you have any options or workaround to improve perfs ?14:29
*** mray has quit IRC14:45
*** mray has joined #openstack14:51
*** Mussi has joined #openstack14:51
*** Mussi has quit IRC14:53
*** adiantum has quit IRC14:56
*** Pentheus has quit IRC15:01
*** adiantum has joined #openstack15:02
*** gondoi has joined #openstack15:04
*** Ryan_Lane has joined #openstack15:07
*** Ryan_Lane is now known as Ryan_Lane|dc15:07
*** DanEllis has joined #openstack15:14
*** msassak has joined #openstack15:15
*** mdomsch has joined #openstack15:22
rcchey guys15:24
*** kashyapc has quit IRC15:24
rccwhat's openstack stance for ec2 api errors?15:24
rccdo we have the same 1-1 behaviour as ?15:25
*** DigitalFlux has quit IRC15:25
*** adiantum has quit IRC15:25
rcci have a couple of bugs assigned regarding errors, and sure, we print out exceptions, but i don't know if we're compliant with ec2 api15:25
*** DanEllis has quit IRC15:31
*** adiantum has joined #openstack15:32
*** rds__ has joined #openstack15:32
*** resker has joined #openstack15:34
jaypipesttx: quoi?15:36
jaypipesberendt: need assistance on i18n'ing?15:36
*** joearnold has joined #openstack15:37
ttxjaypipes: I think we could use a page that explains how to do i18n right. I'll try to write one but might need your help validating it15:38
*** adiantum has quit IRC15:38
*** joearnold has quit IRC15:40
*** hvaldivia has joined #openstack15:40
*** adiantum has joined #openstack15:45
*** hub_cap has joined #openstack15:46
*** dragondm has joined #openstack15:46
jaypipesttx: gotcha. will do.15:49
*** imsplitbit has joined #openstack15:51
*** adiantum has quit IRC15:52
*** adiantum has joined #openstack15:57
hvaldiviaDid anyone use hybridfox for connecting to openstack?15:59
*** burris has joined #openstack15:59
*** rnirmal has joined #openstack16:00
*** guigui has quit IRC16:00
*** troytoman-away is now known as troytoman16:02
hvaldiviaDoes anyone use hybridfox?16:02
*** asalkeld has quit IRC16:02
hvaldiviaI want to use hybridfox for connecting to my nova-cloud, But I get an error16:03
*** adiantum has quit IRC16:05
*** desai has joined #openstack16:05
xtoddxhvaldivia: can you pastie the error you're getting?16:07
*** asalkeld has joined #openstack16:08
hvaldiviaxtoddx: I get this error in hybridfox: EC2 responded with an error for DescribeSecurityGroups16:09
ttxjaypipes: about the translations branch(es), should we approve them as-is or wait for a way to fix the boilerplate ?16:10
*** adiantum has joined #openstack16:10
hvaldiviaxtoddx : I get this error in hybridfox  : EC2 responded with an error for DescribeSecurityGroups16:10
*** littleidea has joined #openstack16:11
jaypipesttx: I was hoping mtaylor would have an answer to that. Perhaps you, me, and mtaylor can discuss when he gets in today?16:12
ttxmtaylor: definitely. We are blocked on that and on nova-core not being able to review proposals to lp:nova/bexar16:12
ttxoops jaypipes ^16:12
hvaldiviain my novarc says that NOVA does not use user id16:12
jaypipesttx: why is nova-core not able to review stuff for bexar? you mean that the approval process isn't automated merge like it is for trunk?16:13
ttxjaypipes: no "review team" specified.16:13
ttxjaypipes: "openstack hudson" owns the branch, he is the only one that can approve until a review team is set.16:14
ttxjaypipes: "openstack hudson" being a cover name for Monty, only him can fix that.16:14
ttxjaypipes: and yes, we need to increase bus factor on that one.16:15
jaypipesttx: heh, ok. yes, let's talk to Mr. Hudson about that :)16:15
hvaldiviaDid anybody use hybridfox? . Well, I use hybridfox for connecting to my eucalyptus cloud. I configure hybridfox with my EC2_ACCESS_KEY,  EC2_SECRET_KEY, EC2_URL and EC2_USER_ID variables. I wanted to use the same approach for my nova-cloud, since the novarc file has the same variables referred before16:22
dprincejaypipes: glance-control is sweet. Where do I specify the log file when using it? In the config file? Which section?16:23
hvaldiviaBut hybrifox returns an error16:23
hvaldiviaAny idea?16:24
jaypipestr3buchet: englilish. nice.16:24
j05hhvaldivia: have you tried other tools?  like euca2tools?16:25
j05hdoes your account have the proper permissions? i've found permissions is usually my problem when API calls fail.16:25
jaypipesdprince: Thanks :) It's mostly swift-init, though, so I can't take all the credit ;) logging can be configured via a separate Python logging config file. I'll work on some documentation for that, sorry it is missing.16:26
jaypipesdprince: would you mind adding a bug to Glance saying "logging configuration not documented" and assign to me, please?16:26
creihtwedjat: well the io wait increases becaues it has to stream a lot of data to disk, and we also try to fsync the upload periodically to prevent a really long fsync16:26
dprincejaypipes: Sure. I can do that. NP16:27
jaypipesdprince: cheers mate16:27
creihtwedjat: what version of swift are you running16:27
*** uksysadmin has quit IRC16:27
creihtI believe that we moved the fsyncs to a thread in 1.2 so that they wouldn't block16:28
jaypipessandywalsh_, dabo: thx for stepping up to the plate with reviewing. much appreciated.16:28
dabojaypipes: part of the job, imo16:28
jaypipestr3buchet: FYI, set your merge prop to Work In Progress while you change stuff, then after you bzr push your changes to LP, set the merge prop back to Needs Review. This will fire off an email to all prior reviewers to re-check your merge proposal.16:29
jaypipesdabo: ya, but it's still good to have your help. thx.16:30
jaypipesdabo: you said it best on the ML. more eyes looking at stuff, the better.16:30
tr3buchetjaypipes: alrighty16:30
dabojaypipes: exactly16:30
* creiht wishes async file io wasn't such voodoo magic16:31
Kegarooops..I was about to ask a question...but I think I found the answer.  :-)16:37
jaypipeshvaldivia: unfortunately, I don't think many folks on the channel (including me!) know much about hybrid fox or have used it.16:38
*** bird has quit IRC16:40
*** grapex has joined #openstack16:40
*** olivier_ has quit IRC16:41
uvirtbotNew bug: #720816 in glance "logging configuration not documented" [Undecided,New]
dprincejaypipes: The glance logging docs bug is created. Apparently I don't have permissions to assign to you.16:42
*** olivier_ has joined #openstack16:43
jaypipesdprince: cheers, I took care of that, thx16:44
*** skiold has quit IRC16:45
wedjatcreiht: i'm running the 1.2.0 version of swift16:45
wedjatcreiht: seems that the auditor processes cause a long iowait on big files, right ?16:45
creihtyeah the auditors can cause a lot of iowait16:46
openstackhudsonProject swift build #200: SUCCESS in 28 sec:
openstackhudsonTarmac: rename log-processing.conf to log-processor.conf16:46
creihtthough we've been working on making that better16:46
wedjati understand16:47
creihtwedjat: you can slow it down my setting max_files_per_second and max_bytes_per_second in the config16:48
*** metoikos has quit IRC16:48
creihtthat may help bring the iowait down16:49
*** MarkAtwood has quit IRC16:49
*** johnpur has joined #openstack16:50
*** ChanServ sets mode: +v johnpur16:50
wedjatcreiht: thanks ! i will try that :)16:50
creihtwedjat: how many drives do you have on a node?16:51
wedjat1, some have 216:51
creihtahh, well when you get more drives in a machine it isn't as bad16:52
creihtwhen you have only one drive16:52
creihtthen all operations are all working on that one drive, instead of distributed accross many16:53
*** berendt has quit IRC16:53
wedjatyeah you're right. it's better when several drives are suffering, instead of having only one16:54
wedjatso my auditor processes will harass several drives :)16:54
creihtonly one drive at a time16:54
tr3buchetwhat is a floating IP?16:54
mtaylorjaypipes, ttx: what do we want to discuss?16:55
jaypipesmtaylor: we need to complete the translation process for nova. and there is a problem: the only one who can approve the merge proposal of ~jaypipes/nova/translations into lp:nova/bexar is the Openstack Hudson LP user, which is you :)16:56
*** thimble has quit IRC16:56
*** hggdh has quit IRC16:56
mtaylorjaypipes: hehe. I should fix that16:56
mtaylorjaypipes, ttx: ok. who should be branch reviewer be? nova-core?16:58
jaypipesmtaylor: well, we need to ensure that the automation picks up the Approved status, but merges it into the correct branch (i.e. NOT lp:nova).16:58
*** hggdh has joined #openstack16:59
*** CloudChris has quit IRC16:59
*** MarkAtwood has joined #openstack17:00
*** CloudChris has joined #openstack17:00
*** mdomsch has quit IRC17:01
jaypipessoren: please, please help figure out what's up with blamar's Author/Email signature here:
jaypipessoren: I'm unsure what the problem is.17:02
hvaldiviaj05h: Hi, I used euca2ools. it works fine with nova.17:03
KegaroI have been having some trouble setting up a swift system.  here is the command i am trying to run, and the syslog output:   Can someone show me where to start looking?17:03
*** adiantum has quit IRC17:03
hvaldiviaJ05h: I will try to see the account permissions17:03
ttxmtaylor: nova-core, yes17:04
ttxmtaylor: about the translations branch(es), did you find a way to fix the boilerplate ?17:04
*** littleidea has quit IRC17:05
ttxmtaylor: or should we just commit as-is17:05
*** adiantum has joined #openstack17:05
*** isvara has joined #openstack17:06
*** CloudChris has quit IRC17:06
*** calavera has quit IRC17:08
mtaylorttx: I think we should just commit as is for now. I think I need to go but pitti about that17:08
ttxjaypipes: about blamar branch: I think the .mailmap is inverted17:09
mtaylorttx: ok. nova-core should now be able to approve merges17:09
blamarttx, I've tried it both ways :( I'm pretty sure17:09
jaypipesttx: we've tried both ways. no go.17:12
*** adiantum has quit IRC17:12
*** photron has joined #openstack17:14
*** adiantum has joined #openstack17:14
*** kashyapc has joined #openstack17:15
*** littleidea has joined #openstack17:17
sandywalsh_grapex, did you have a question about novatools?17:21
*** KnightHacker has joined #openstack17:23
*** that__guy has joined #openstack17:26
*** ibarrera has quit IRC17:26
*** KnightHacker has left #openstack17:29
that__guyhi guys.  Clearing out the contents of /var/lib/nova/instances/_base/ after stopping the nova-compute service (to get rid of cached amis) shouldn't cause problems, right?17:29
*** justinsb_ has joined #openstack17:31
uvirtbotNew bug: #720850 in nova "No handlers could be found for logger "nova.root"" [Undecided,New]
*** mdomsch has joined #openstack17:32
jaypipesdevcamcar: see that__guy's ? above ^^17:32
*** fysa has joined #openstack17:32
*** justinsb has quit IRC17:34
*** mdomsch has quit IRC17:35
*** dendrobates is now known as dendro-afk17:37
that__guyjaypipes: you guys have a pool going on the number of stupid questions per day?17:38
*** vvuksan has joined #openstack17:38
*** rds__ has quit IRC17:41
*** adiantum has quit IRC17:44
*** omidhdl has joined #openstack17:45
jaypipesthat__guy: no such thing as a stupid question.17:46
jaypipesthat__guy: unless I ask it.17:46
*** mdomsch has joined #openstack17:46
jarrodjust to verify, volume snapshots are unsupported at this time for libvirt/kvm ?17:46
*** dendro-afk is now known as dendrobates17:51
*** adiantum has joined #openstack17:51
jaypipesjarrod: AFAIK, that is correct, yes.17:55
jarrodok, so i move to XEN17:56
*** imsplitbit has quit IRC18:00
justinsb_Hi - I'd like to get this merged, but we need to decide on whether to call metadata 'metadata' or 'properties'.  'metadata' might cause confusion in the DB migration file; 'properties' would require changing the API.  I tried the mailing list, but this seems a better forum for reaching consensus...18:01
justinsb_eday, dprince, jaypipes: I'm looking at you here :-)18:01
dprincejustinsb_: Hey. You got this coded up quick!18:02
openstackhudsonProject nova build #527: SUCCESS in 1 min 36 sec:
openstackhudsonTarmac: example:18:02
openstackhudson# nova-manage network list18:02
openstackhudsonCIDR                    netmask         dhcp_start      DNS18:02
openstackhudsonxx.xx.35.0/25  xx.xx.35.2      None18:02
openstackhudsonxx.xx.36.0/25  xx.xx.36.2      None18:02
openstackhudson(DNS set to None because networks not yet used..)18:02
dprincejustinsb_: I know I voted for 'instance_properties' the other day. After pondering it a bit more I'm liking calling it 'instance_metadata' a bit more.18:03
jaypipesjustinsb_: yes, I know... been swamped, sorry.18:03
jaypipesjustinsb_: I also voted for instance_properties and the model named InstanceProperty.18:03
jaypipesjustinsb_: and I believe I responded that to your ML thread ;)18:04
dprincejustinsb_: So. You have my blessing whichever way we go.18:04
justinsb_dprince, jaypipes: No worries about delays etc.  I think real-time chat is just much easier.18:04
justinsb_jaypipes: My concern with properties is that the existing API uses 'metadata'.18:05
jk0kpepple: unit tests passing in your branch?18:05
justinsb_jaypipes: eday then rejected the patch because it got awkward with both metadata and properties depending on which code area we were in (API vs internals)18:05
jaypipesjustinsb_: I understand your concern, but vishy and others explained on the merge proposal ( that metadata in the API is meant for client-side metadata, not server-side.18:05
justinsb_jaypipes: I didn't really get vish's point - I just replied to that on the main thread.  I'm just implementing the CloudServers API & associated blueprint here.18:06
justinsb_jaypipes: What we do with that metadata is another patch :-)18:06
*** adiantum has quit IRC18:06
*** MarcMorata has quit IRC18:09
jaypipesjustinsb_: well, like I said on the merge prop, whatever the list decides is fine with me, I'll go along. Is there a clear consensus on the ML thread?18:09
jarrodAre there some instructions for installing Xen with openstack?18:09
openstackhudsonProject nova-tarmac build #62,070: FAILURE in 4 min 2 sec:
kpepplejk0: away from machine. give me 15.18:10
edayjaypipes, justinsb_ : I think since the API doc uses 'metadata', we might just want to use that all the way down. I'm not concerned with overlap with sqlalchemy 'metadata', there are always going to be overlaps across contexts. within the same context (nova), I would prefer to keep it consistent (don't switch from md -> props)18:11
justinsb_jaypipes: You & dprince replied on the mailing list liking properties.  I coded properties, and eday then rejected it because the API called it metadata, and switching was confusing.18:11
jaypipeseday: metadata != properties, just saying. metadata is data about data. properties are just data.18:11
justinsb_jaypipes: I dislike properties because we have properties that are not in this collection (e.g. InstanceType)18:12
jaypipesjustinsb_: Tags then?18:12
jaypipesjustinsb_: technically, InstanceType is an attribute...18:12
jaypipesbut I digress..18:12
justinsb_jaypipes: Sure - I like 'tags'. But then we'd have to change the API for consistency.  Am I allowed to do that?18:13
jaypipesjustinsb_: the API as in the OpenStack API? no.18:13
*** adiantum has joined #openstack18:13
justinsb_jaypipes: I'm digging out my dictionary to learn what the difference is between properties & attributes & metadata :-)18:13
justinsb_jaypipes: If we're not allowed to change the API, then I think we have to go with eday's suggestion of calling it Metadata all the way down, or live with inconsistencies.18:14
jaypipesjustinsb_: it's a shame that the term metadata was used in the OpenStack/CloudServers API. But since it was, I suppose I will side with eday in the consistency battle...18:14
jaypipesjustinsb_: k, ++. sold me.18:14
openstackhudsonProject nova-tarmac build #62,071: STILL FAILING in 3 min 52 sec:
dprincejustinsb_: I agree with eday as well.18:15
justinsb_jaypipes, eday, dprince: Hooray!  Metadata it is, for consistency.  Thanks.  I'll call the table instance_metadata, the model InstanceMetadata, and the property 'metadata'18:15
dprincejustinsb_: "Metadata" initially through me off. There were other blueprints and code that mention it but now that I get it consistency makes the most sense.18:15
jaypipescool. agreement then :)18:16
justinsb_dprince: Hopefully calling it InstanceMetadata in the DB model will make it clearer18:16
dprincejustinsb_: Sure. It will.18:16
jaypipesjustinsb_: I set the merge prop back to Work In Progress while you make those (hopefully final!) changes. :)18:16
dprincejustinsb_: So I added one more comment on the merge review. How hard is it to add a test for the upper limit of the quota?18:16
*** eikke has quit IRC18:17
dprincejustinsb_: Seems like something like that could be easily broken and an unlimited quota is kind of a security concern.18:17
justinsb_dprince: I could add a unit test, but I was a bit disappointed in my unit tests before, because the unit tests all passed but then everything failed when I tried it for real18:17
justinsb_dprince: I wrote an integration test that tested it properly, but I don't know how to submit that...18:17
sirp_jaypipes: as an alternative to removing validation entirely, i could just add 'vhd' to the list of approved types (had considered that first, but objected on philosophical grounds, but since type is going away any way, guess it doesn't matter)18:17
justinsb_dprince: But I can submit a unit test, just remember that it probably won't actually test anything until we have integration testing18:18
dprincejustinsb_: Sure. That is probably fine I guess. Is that in smoke tests then?18:19
*** adiantum has quit IRC18:19
jaypipessirp_: I actually would prefer that approach. it would make it easier on me (and the bug704854 merge proposal) and like you say, can easily be refactored a bit with the image_format blueprint.18:19
openstackhudsonProject nova-tarmac build #62,072: STILL FAILING in 4 min 5 sec:
justinsb_dprince: It's the Cloudservers/OpenStack API (for now), so not in Smoke Tests.  It was just a crappy test (bash script).  I really long for a real test suite!18:20
sirp_jaypipes: cool, ill make that change and re-propose18:20
edayjaypipes: perhaps properties is more appropraite for what it actually is, I'm just a big fan of consistency :)18:21
*** gregp76 has joined #openstack18:21
*** dendrobates is now known as dendro-afk18:21
jaypipessirp_: thx mate. sorry for not getting back to you on that sooner... been swamped trying to get through email backlog and reviews..18:22
Code_Bleuinstalling openstack on one pc using the manual mode, and everything has went well, up until the enabling access to VMs on compute node part.  Here is the error im getting.  Does anyone know what might be the problem?
*** CloudChris has joined #openstack18:22
edayjaypipes: and changing the patch and nova code is probably easier than the API spec at this point18:22
jaypipeseday: I am too. consistency should rule in this case.18:22
*** sebastianstadil has quit IRC18:23
openstackhudsonProject nova-tarmac build #62,073: STILL FAILING in 3 min 56 sec:
*** jaypipes is now known as jaypipes-afk18:25
xtoddxCode_Bleu: can you post logs from your api controller?18:25
*** adiantum has joined #openstack18:26
*** photron has quit IRC18:26
openstackhudsonProject nova-tarmac build #62,074: STILL FAILING in 3 min 55 sec:
*** CloudChris has quit IRC18:30
*** rcc has quit IRC18:30
Code_Bleuxtoddx: here it is
*** adiantum has quit IRC18:34
openstackhudsonProject nova-tarmac build #62,075: STILL FAILING in 4 min 13 sec:
*** dirakx has joined #openstack18:35
xtoddxCode_Bleu: looks like an auth failure.  i'd nova-mange users list (to make sure your account exists), nova-mange project list USERNAME (to make sure you have at least one project you're active in) and nova-mange project zipfile PROJECT USER to get new credentials, source them, and try again.18:37
Code_Bleuxtoddx: ok. i did have an issue using the variables $NOVA_PROJECT and $NOVA_USERS during the creation of the certification, but i just put in the same name i did when i created the user and project instead and it seemed to work.  I will check what you said and see18:39
openstackhudsonProject nova-tarmac build #62,076: STILL FAILING in 3 min 58 sec:
*** adiantum has joined #openstack18:41
*** Kegaro has quit IRC18:43
openstackhudsonProject nova-tarmac build #62,077: STILL FAILING in 4 min 31 sec:
Code_Bleuxtoddx: here is the results again:
*** mray has quit IRC18:46
xtoddxCode_Bleu: you need to cd to /root and unzip the, then '. novarc' to source the new credentials into your shell environment18:47
*** isvara has left #openstack18:48
*** rds__ has joined #openstack18:48
Code_Bleuxtoddx: sorry..i forgot that part again....will do and let you know18:49
*** adiantum has quit IRC18:49
openstackhudsonProject nova-tarmac build #62,078: STILL FAILING in 4 min 12 sec:
*** mray has joined #openstack18:52
*** kbringard has joined #openstack18:54
kbringardhey guys, I have a question about load balancing...18:54
openstackhudsonProject nova-tarmac build #62,079: STILL FAILING in 3 min 51 sec:
*** adiantum has joined #openstack18:56
Code_Bleuxtoddx: still no dice
kbringardI have nginx setup to load balance between 2 API frontends, but that seems to cause the signature generated by euca-* to break18:56
*** resker has quit IRC18:59
openstackhudsonProject nova-tarmac build #62,080: STILL FAILING in 4 min 12 sec:
xtoddxCode_Bleu: echo $EC2_ACCESS_KEY and compare to what is in novarc, and what you get from `nova-manage user exports admin`19:01
*** blueadept has joined #openstack19:03
Code_Bleuxtoddx: 76c4b624-eb43-436b-b1a6-fd3e53b87946:adminproject19:03
xtoddxCode_Bleu: is that the same in all 3 places19:04
*** burris has quit IRC19:04
Code_Bleuxtoddx: sorry...yes19:04
openstackhudsonProject nova-tarmac build #62,081: STILL FAILING in 3 min 59 sec:
xtoddxCode_Bleu: I also should have mentioned that if you're running nova-api with flags or a flagfile, you need to run nova-mange with the same ones, otherwise they might be trying to look in different databases for users.19:05
*** adiantum has quit IRC19:06
Code_Bleuxtoddx: sorry, im not following you.  all the nova-manage commands are running fine19:08
xtoddxCode_Bleu: when you ran nova-api, did you pass a --flagfile or any other arguments?19:08
sorenjaypipes-afk: Looks like Brian worked it out.19:09
openstackhudsonProject nova-tarmac build #62,082: STILL FAILING in 3 min 56 sec:
*** hvaldivia has left #openstack19:10
*** burris has joined #openstack19:11
*** kashyapc has quit IRC19:12
tr3buchetjaypipes, sandywalsh_, dabo, _cerberus_:
tr3buchetlunch time!19:14
openstackhudsonProject nova-tarmac build #62,083: STILL FAILING in 3 min 48 sec:
Code_Bleuxtoddx: no, not that i recall...ive been following the document online..step by step19:14
*** alekibango has quit IRC19:16
*** adiantum has joined #openstack19:19
openstackhudsonProject nova-tarmac build #62,084: STILL FAILING in 4 min 0 sec:
*** drico_ has quit IRC19:20
*** drico__ has joined #openstack19:20
*** jc_smith has joined #openstack19:21
*** hggdh has quit IRC19:22
*** hggdh has joined #openstack19:22
xtoddxCode_Bleu: what document, are you using contrib/ ?19:23
*** Jordandev has joined #openstack19:23
xtoddxCode_Bleu: try running ./bin/nova-manage --flagfile=bin/nova.conf user list19:24
openstackhudsonProject nova-tarmac build #62,085: STILL FAILING in 4 min 1 sec:
xtoddxCode_Bleu: I suspect your nova-manage and nova-api aren't looking at the same flagfile19:26
Code_Bleuxtoddx: that returned the same results: admin19:26
*** adiantum has quit IRC19:26
xtoddxCode_Bleu: does using that flagfile with the command 'user exports admin' return the same access key?19:27
Code_Bleuxtoddx: although i just ran 'nova-manage --flagfile=/etc/nova/nova.conf user list'19:27
zedashey is there docs on how to actually work with swift ACLs? specifically setting/changing X-Container-Read and X-Container-Write ACLs?19:28
xtoddxCode_Bleu: thats right, assuming your nova-api ran with that flagfile (which it would if you set it up with
daboStill looking for another nova-core dev to review/approve:
Code_Bleuxtoddx: yes, same access key19:29
xtoddxCode_Bleu: can you 'ps auxwww | grep nova-api' and see what arguments nova-api is running with19:29
*** CloudChris has joined #openstack19:30
Code_Bleuxtoddx: su -c nova-api --flagfile=/etc/nova/nova.conf nova19:31
Code_Bleuxtoddx: /usr/bin/python /usr/bin/nova-api --flagfile=/etc/nova/nova.conf19:31
openstackhudsonProject nova build #528: SUCCESS in 1 min 36 sec:
openstackhudsonTarmac: Add a command to nova-manage to list fixed ip's.19:32
*** perestre1ka has joined #openstack19:33
xtoddxso running "./bin/nova-mange --flagfile=/etc/nova/nova.conf user exports admin" gives you the same access key that is already in your environment?19:33
*** adiantum has joined #openstack19:33
xtoddxCode_Bleu: ^^19:33
*** littleidea has quit IRC19:33
Code_Bleuxtoddx: yes19:34
*** fabiand_ has joined #openstack19:35
xtoddxCode_Bleu: can you make sure your flagfile has --verbose in it, and pastie the nova-api log again?19:35
*** perestrelka has quit IRC19:36
xtoddxCode_Bleu: I gotta run out for a bit, bb soon19:37
Code_Bleuxtoddx: ok, thanks19:37
openstackhudsonYippie, build fixed!19:38
openstackhudsonProject nova-tarmac build #62,086: FIXED in 12 min:
openstackhudsonProject nova build #529: SUCCESS in 1 min 35 sec:
openstackhudson* Tarmac: I have a bug fix, additional tests for the `limiter` method, and additional commenting for a couple classes in the OpenStack API. Basically I've just tried to jump in somewhere to get my feet wet. Constructive criticism welcome.19:39
openstackhudson* Tarmac: added labels to networks for use in multi-nic19:39
openstackhudsonadded writing network data to xenstore param-list19:39
openstackhudsonadded call to agent to reset network19:39
openstackhudsonadded reset_network call to openstack api19:39
*** sebastianstadil has joined #openstack19:40
*** Jordandev has quit IRC19:41
openstackhudsonProject nova build #530: SUCCESS in 1 min 36 sec:
openstackhudsonTarmac: Update the Openstack API so that it returns 'addresses'.19:44
openstackhudsonThis branch should resolve nova bug #713144 (
uvirtbotLaunchpad bug 713144 in nova "Openstack API:  GET on /servers/<id> returns empty 'addresses'" [High,In progress]19:44
uvirtbotLaunchpad bug 713144 in nova "Openstack API:  GET on /servers/<id> returns empty 'addresses'" [High,In progress]
Code_Bleuxtoddx: i dont understand...i added --verbose to the file...restarted nova-api...ran a tail -f /var/log/nova/nova-api.log and started another ssh session and ran the auth command and it worked.19:44
*** adiantum has quit IRC19:44
*** adiantum has joined #openstack19:46
*** dendro-afk is now known as dendrobates19:52
kpepplejk0: sorry, my meeting went waaay long. revno 660 of my branch runs unittests clean.19:53
jk0no problem19:53
*** MarcMorata has joined #openstack19:54
*** jbryce has joined #openstack19:59
*** adiantum has quit IRC20:05
*** grapex has left #openstack20:06
creihtmtaylor: ping20:06
*** adiantum has joined #openstack20:06
*** fabiand_ has quit IRC20:06
*** brd_from_italy has joined #openstack20:07
*** csackmann has joined #openstack20:10
*** jaypipes-afk is now known as jaypipes20:14
*** vvuksan has quit IRC20:15
*** MarcMorata has quit IRC20:15
mtaylorcreiht: pong-diddly-pong20:15
jaypipesblamar: so what was the Authors issue in the end?20:16
zedasalright so I figured out how to do ACL changes via curl:  Question is: why was this removed from the API docs on in Jan 2011?  Is there a change coming?20:16
*** dendrobates is now known as dendro-afk20:18
openstackhudsonProject nova build #531: SUCCESS in 1 min 33 sec:
openstackhudsonTarmac: Hi guys20:22
openstackhudsonTaking Jay suggestion, i completely got rid of BadParameter and just used ValueError, no point of using a new exception class when there's already something available in Python for that purpose.20:22
Code_Bleuanyone know how i can check this? Unable to run euca-describe-images.  Is euca2ools environment set up?20:23
creihtmtaylor: do we have tarmac setup for the 1.2 series for swift yet?20:23
creihtzedas: which docs?20:23
creihtthere is no change20:23
creihtmost likely a mistake20:23
*** desai has quit IRC20:24
creihtActually, I'm not sure they were eve in the api docs yet, since those are derrived from the cloud files api docs20:24
*** adiantum has quit IRC20:24
creihtand cloud files doesn't have that feature yet20:24
zedascreiht: PDF linked off here has a note "Jan. 12, 201120:26
zedas"Removed references to ACL (Access Control List).20:26
*** csackmann has quit IRC20:26
zedasand there's no mention of ACL changes, X-Container-Read, X-Container-Write20:26
*** omidhdl has quit IRC20:27
*** desai has joined #openstack20:27
creihtzedas: Not sure off the top of my head, other than the current cloud files doesn't support ACLs20:27
creihtso either it was something really old, or somehow some reference had gotten in there20:28
creihtbut yes, the ACL stuff needs to be added to the openstack dev guide located here:20:28
zedascreiht: ok great, so i can expect this to stay in Swift and not suddenly change later.  very cool, and thanks.20:28
creihtzedas: indeed20:28
creihtand np20:28
zedashell i can add it to the docs.20:28
creihtzedas: and welcome to the group :)20:28
creihtannegentle: -^ :)20:28
*** sebastianstadil has quit IRC20:28
creihtzedas: We are trying to get that together now20:29
zedascreiht: thanks, still wrapping my head around it but so far pretty fun.20:29
*** CloudChris has quit IRC20:29
creihtfeel free to let us know if you have any questions20:29
zedasi will. mostly there's nothing new.  i think when i start seriously trying to get it to run on OSX then i'l have real questions.20:30
creihtzedas: if you have any comments on docs, feel free to ping annegentle, as she is heading up that effort20:30
zedasok will do.20:30
*** adiantum has joined #openstack20:31
uvirtbotNew bug: #720941 in nova "Set up a XenServer test cluster and integration test" [High,Confirmed]
*** johnpur has quit IRC20:32
Code_Bleupublish tarball error: any ideas?20:32
creihtzedas: I just heard that it was some CDN acl stuff that was removed20:32
blamarjaypipes, yeah, I think so, combination of other Brians/Authors/mailmap... thanks for the assistance !20:33
jaypipesblamar: np, but I think it was all soren :)20:34
blamarjaypipes, soren: I guess I'll have to send two gift baskets then20:34
jaypipesblamar: hehe20:35
*** adiantum has quit IRC20:39
*** dendro-afk is now known as dendrobates20:41
zedascreiht: ah ok, thanks.20:43
openstackhudsonProject nova build #532: SUCCESS in 1 min 34 sec:
openstackhudsonTarmac: Re-alphabetise Authors, move extra addresses into .mailmap.20:44
*** adiantum has joined #openstack20:44
*** spectorclan has joined #openstack20:46
openstackhudsonProject swift build #201: SUCCESS in 29 sec:
openstackhudsonTarmac: fixed glob pattern matching index offset error in stats' LogUploader20:47
jk0kpepple: just fired off a merge prop your way20:47
openstackhudsonProject nova build #533: SUCCESS in 1 min 38 sec:
openstackhudsonTarmac: Make eth0 the default for the public_interface flag.20:48
mtaylorcreiht: yes!20:50
creihtmtaylor: awesome thanks!20:50
blamardoes anyone know anything about plans to support multiple user types for OpenStack as a whole? I guess to replace is_admin?20:52
creihtmtaylor: would you have any idea why it doesn't let us approve a merge prop for the 1.2 series?20:55
justinsb_What's the plan for DB migrations now?  I saw a 003_ just got merged; I also had a 003_.  Do I need to rename to 004_?  Isn't this going to be problematic?20:55
creihtmtaylor: for example:
*** miclorb_ has joined #openstack20:55
sorencreiht: Because the hudson user is the default reviewer for the target branch.20:55
sorencreiht: ...which means only it can approve stuff.20:55
soren...and only mtaylor can act as openstack-hudson.20:55
redbopeople keep lighting up my irc by talking about my branches20:56
creihtmtaylor: would you mind fixing that please :)20:56
*** jmckenty has joined #openstack20:58
*** h0cin has quit IRC20:59
jarrodmr jmc20:59
*** adiantum has quit IRC20:59
*** joearnold has joined #openstack21:00
*** vvuksan has joined #openstack21:01
openstackhudsonProject swift build #202: SUCCESS in 34 sec:
openstackhudsonTarmac: ring: pickles now use only stdlib objects; old and really old pickles can still be read21:01
jarrodkvm snapshots seem broke, perhaps better to roll with lvm ?21:02
*** jesse_ has joined #openstack21:02
*** asalkeld has left #openstack21:03
*** adiantum has joined #openstack21:06
*** naehring has joined #openstack21:08
*** naehring has quit IRC21:09
jaypipesdabo: hmm, I noted in my review that you weren't in the Authors list... you said you fixed that :) looks like you didn't?21:09
daboI fixed it using 'bzr whoami' to use the name in Authors. Tests were passing on my end.21:10
sandywalsh_jaypipes, I'm trying to add localization to some new code. Is this page still accurate?21:12
sandywalsh_'pygettext -p locale -d nova nova/*.py' doesn't get all the files21:12
sandywalsh_jaypipes, do you use a different incantation?21:12
sandywalsh_jaypipes, sorry,
jaypipessandywalsh_: not quite sure what you're trying to do... are you attempting to build the message catalogs, compile them, or?21:13
sandywalsh_rebuild the .pot21:14
sandywalsh_to include my new file21:14
jaypipessandywalsh_: you shouldn't need to do any of that yourself any more... Launchpad does that now.21:14
sandywalsh_ah... tests are failing in the meanwhile21:14
jaypipessandywalsh_: all you're responsible for is adding the _() stuff :)21:14
dabojaypipes: also, pygettext is deprecated21:14
jaypipesdabo: we don't use pygettext...21:14
jaypipessandywalsh_: what tests are failing?21:14
dabojaypipes: just reading from the wiki page21:14
sandywalsh_there's a localization test21:15
jaypipesdabo: ah.21:15
jaypipesthere is?21:15
dabosandywalsh_: the multiple string formatting?21:15
dabosandywalsh_: yeah, that one21:15
daboyou need to use mappings if you have more than one replacement in a localized string21:15
jaypipessandywalsh_: that shouldn't have anything to do with building the pot file.21:15
jaypipesdabo: we REALLY need to create the documentation for i18n'ing stuff for devs.. ;)21:16
jaypipescan't really expect people to understand it...21:16
dabowell, the test does explain a bit when it fails. :)21:16
jaypipessandywalsh_: what dabo is saying is that you have introduced an i18n'd string that contains >1 %s...21:17
annegentlezedas: great catch on ACL in the docs - feel free to get a copy of the source in openstack-manuals and edit, or I can do it.21:17
jaypipessandywalsh_: and that those %s's should be turned into %(var1)s, %(var2), etc21:17
sandywalsh_gotcha ... easy fix :)21:18
sandywalsh_cryptic error21:18
jaypipessandywalsh_: k, cool :)21:18
sandywalsh_thx guys21:18
dabosandywalsh_: what else should I add to the error message to make it less cryptic?21:21
sandywalsh_what jay said21:22
*** adiantum has quit IRC21:24
*** pvo has joined #openstack21:25
*** Pentheus has joined #openstack21:26
*** allsystemsarego has quit IRC21:27
*** CloudChris has joined #openstack21:28
* soren fixes more concurrency problems21:28
*** adiantum has joined #openstack21:31
*** dprince has quit IRC21:32
*** jmckenty has quit IRC21:35
* creiht pokes mtaylor again21:36
zedasannegentle: ok21:36
annegentlezedas: now that I'm looking at it, it might be easiest for me to put it in since I have easy access to the removed bits :)21:37
annegentlezedas: but you are welcome to contribute docs!21:37
*** adiantum has quit IRC21:40
jarrodubuntu 10.04 doesnt even support xenserver anymore21:41
creihtannegentle: those removed bits don't need to be put back in21:42
*** miclorb_ has quit IRC21:42
creihtthey were for a cloudfiles feature that was depricated, and then removed21:42
*** miclorb_ has joined #openstack21:43
kpepplejk0: merged but we seemed to pick up another migration file from trunk ... which hoses our migration file. hold on before you pull.21:44
*** miclorb_ has quit IRC21:44
j05hannegentle: do you know if there's something we can do on the wiki to flag not working docs? it could help when folks find issues methinks.21:44
*** miclorb_ has joined #openstack21:44
annegentlej05h: yeah I have a page insertion that I use... let me find it21:44
*** adiantum has joined #openstack21:45
annegentlej05h: ok, you can insert <<Include(DeprecatedPage)>> at the top of a wiki page to see the text in this page:
annegentlej05h: text can be tweaked21:47
*** dendrobates is now known as dendro-afk21:47
annegentlej05h: but it does give a reader a heads-up21:47
*** filler has quit IRC21:47
*** lvaughn_ has quit IRC21:47
*** lvaughn has joined #openstack21:48
*** filler has joined #openstack21:48
annegentlecreiht: ok, looking.... yeah I got ACL confused with public containers21:49
creihtthe only docs we have so far on ACLS are at:
kpepplejk0: ok, lp:~ken-pepple/nova/dynamicinstancetypes is back working at revno 66221:50
uvirtbotNew bug: #720984 in nova "nova-manage network create passes a label kwargs that VlanManager doesn't accept" [Undecided,New]
kpepplejk0: i'm working docs and then should be ready for code review if you're done21:51
jk0sounds good man21:52
*** mdomsch has quit IRC21:53
*** arun_ has quit IRC21:55
*** lvaughn_ has joined #openstack22:01
openstackhudsonProject swift build #203: SUCCESS in 30 sec:
openstackhudsonTarmac: Make swift-auth-to-swauth work with really old devauth dbs.22:01
openstackhudsonUpdate swauth to accept non-alnum chars in account and user names.22:01
openstackhudsonProject nova build #534: SUCCESS in 1 min 40 sec:
openstackhudsonTarmac: Added http://mynova/v1.0/zones/ api options for add/remove/update/delete zones. child_zones table added to database and migration. Changed novarc vars from CLOUD_SERVERS_* to NOVA_* to work with novatools. See python-novatools on github for help testing this.22:02
*** lvaughn has quit IRC22:03
jk0kpepple: I'm ready to review whenever you are22:07
*** desai has quit IRC22:09
*** adiantum has quit IRC22:09
sandywalsh_thanks for the reviews everyone!22:11
openstackhudsonProject nova build #535: SUCCESS in 1 min 36 sec:
openstackhudsonTarmac: Add **kwargs to VlanManager's create_networks so that optional args from other managers don't break.22:12
*** hub_cap has quit IRC22:14
*** troytoman has quit IRC22:14
*** troytoman-away has joined #openstack22:15
*** arun_ has joined #openstack22:20
sorendabo: What do you mean it won't let you push it?22:20
dabosoren: bzr merge error, due to branches "diverging"22:21
*** adiantum has joined #openstack22:22
*** spectorclan has quit IRC22:25
*** CloudChris has quit IRC22:27
*** mray has quit IRC22:28
dabosoren: jaypipes just showed me how to correct that. My bzr-fu is not very strong.22:30
daboso forget about the new branch.22:31
jaypipesdabo: it's exactly the same as git. git push --force == bzr push --overwrite...22:32
*** dfg has joined #openstack22:33
dabojaypipes: I never seemed to have as many problems with git. Bzr seems to object to a lot more. Don't know if that's good or bad, but it is frustrating at times.22:33
*** mray has joined #openstack22:35
jaypipesdabo: sure, understood.22:35
jaypipesdabo: it's partly because the underlying representation for "what is a changeset" is very different between bzr and git...22:35
jaypipesdabo: and those differences manifest themselves in ways like this :)22:35
dabojaypipes: understood. I just can't foresee when they will cause problems22:36
*** desai has joined #openstack22:37
jaypipesdabo: hehe, that's why your buddy jay is here ;)22:37
dabojaypipes: exactly! Now I have a retort when everyone says that you're not good for anything. ;-)22:38
*** kbringard has quit IRC22:38
dabobtw, do I need to 'resubmit proposal'?22:38
*** ppetraki has quit IRC22:40
jaypipesdabo: hehe22:40
jaypipesdabo: did you push --overwrite to that original LP branch?22:41
dabojaypipes: yep22:41
jaypipesif so, simply set the merge prop to Work in Progress, then back to Needs Review...22:41
jaypipesdabo: and you can delete the other branch.22:41
daboalready deleted it22:41
jaypipesdabo: k. should be all set then.22:41
dabojaypipes: ok, status has been toggled.22:42
jaypipesdabo: sweeet, I'll approve it now.22:42
jaypipesdabo: you'll love my review comment.22:43
dabojaypipes: so true...22:44
sorenHaving to "bzr push --overwrite" is an indication that you've done something else very wrong, I'm afraid.22:45
*** sandywalsh_ is now known as sandywalsh22:45
sorendabo You shoudn't be able to end up having diverged with your own branch. Have you uncommitted anything, for instance?22:45
dabosoren: You mean besides doing some development on a machine that had a different 'bzr whomai' setting?22:46
*** joearnold has quit IRC22:46
dabosoren: yes, I had to uncommit several commits, as I was working on a different machine at the time22:46
*** desai has quit IRC22:46
daboand the author info for the commits was different than the Authors document22:46
sorendabo: Ok. Uncommitting stuff is generally a bad idea.22:49
dabosoren: of course, but what other option is there in this case?22:49
sorendabo: That's what .mailmap is for.22:49
dabosoren: now you tell me. :)22:50
*** MarkAtwood has quit IRC22:50
daboI had no idea that that file was even there22:51
sorendabo: You ask, I answer :)22:51
daboyeah, but i asked a few hours ago :(22:51
openstackhudsonProject nova build #536: SUCCESS in 1 min 36 sec:
openstackhudsonTarmac: Added support for feature parity with the current Rackspace Cloud Servers practice of "injecting" files into newly-created instances for configuration, etc. However, this is in no way restricted to only writing files to the guest when it is first created.22:52
sandywalshEverytime a LP diff takes a long time to update god kills a puppy22:52
sandywalshMulti-Cluster/Zones Phase 2 -
*** jbryce has quit IRC22:59
*** desai has joined #openstack23:00
jaypipessandywalsh: and kittens cry.23:00
*** desai has quit IRC23:04
jaypipesjustinsb_: nice work on the latest metadata patch.23:05
*** MarcMorata has joined #openstack23:06
*** desai has joined #openstack23:07
*** brd_from_italy has quit IRC23:08
jaypipesjustinsb_: however, there is a problem due to your use of git...23:08
*** desai has quit IRC23:09
jaypipesjustinsb_: never mind, looks like you found the problem and pushed to a metadata3 branch...23:10
vishyalthough there are a bunch of merge errors there :)23:11
uvirtbotNew bug: #721004 in nova "s3_host and ec2_url set wrongly in nova.conf when using OpenStack NOVA Installer Scripts" [Undecided,New]
justinsb_jaypipes: I think it might actually be a bazaar / launchpad bug.  I thought it was git as well...  I might need some help here.23:11
justinsb_jaypipes: I was using git-bzr-ng, and was getting the error, which I presumed was a git/bzr problem23:12
justinsb_jaypipes: So I checked out lp:nova, copied the raw files and did one clean check-in23:12
jaypipesjustinsb_: no, it's not. it has to do with in your local git branch, you did a: mv
*** gondoi has quit IRC23:12
jaypipesjustinsb_: and git thought that was because it was the same file after you merged trunk...23:13
justinsb_jaypipes: Hold on, I think we're talking about 2 different issues.  Let me check the current merge request patch23:14
jaypipesjustinsb_: I've seen this issue now about 5 times with the git-bzr-ng bridge. Each time it is because a file was mv'd to replace an existing file. bzr and git have different ideas of what that means to a changeset, which results in problems when pushing that look like a file was deleted and re-added.23:14
justinsb_jaypipes: But I didn't use the git bazaar bridge for this checkin (...metadata3)23:15
mtaylorcreiht: what now??? :)23:15
justinsb_jaypipes: I gave up and did it by hand23:15
jaypipesjustinsb_: nothing wrong with git OR bzr. It's just the two tools have different ideas of what a mv of a file means to the changeset...23:15
jaypipesjustinsb_: right, which is why metadata3 is correct and metadata2 is not.23:15
vishyjustinsb_: metadata3 looks like it just needs a merge trunk23:16
jaypipesjustinsb_: git-bzr-ng bridge works fine *until* someone mv's a file :)23:16
jaypipesvishy: yes, it does, but that's a different problem than what was happening with metadata2 :)23:16
uvirtbotNew bug: #721005 in nova "Incompatible comparison on flavor_id when using the Openstack API with XML request body" [Undecided,New]
jk0kpepple: ready to rock?23:17
termiejaypipes: i have run into similar problems without any gitness involved23:17
termiejaypipes: a couple times i have had a "conflict contents"23:17
jaypipestermie: and that has to do with justinsb_ how?23:17
justinsb_vishy, jaypipes: Let's ignore metadata2 for the moment - why does metadata3 have the weird "+<<<<<<< TREE" and "+>>>>>>> MERGE-SOURCE" stuff in it?23:18
jaypipesjustinsb_: that is a merge conflict that needs to be resolved.23:18
vishyjustinsb_: yeah merge trunk and fix conflicts23:18
termiejaypipes: just saying i think there is some way of generating the issue outside of git23:18
jaypipestermie: depends on what the "issue" is :)23:18
*** jmckenty has joined #openstack23:18
jaypipestermie: and right now I really don't feel like getting into a bzr vs git debate :)23:19
jaypipesjust trying to solve justinsb_'s current issue.23:19
termiejaypipes: i wasn't, i was just pointing out that i think there is a way of running into that outside of git, so the solution to that may be applicable23:19
justinsb_vishy, jaypipes: Cool - didn't realize someone had committed something in the meantime!  That's a speedy trunk :-)23:19
jaypipesjustinsb_: heh, yes it is. :)23:20
*** MarcMorata has quit IRC23:20
justinsb_On the rename thing, I think that the 'micro' migrations are going to cause a nightmare ... I was 003_ and then someone else committed 003_ first, so now I'm 004_, but if someone else beats me to the punch...23:20
jaypipestermie: not sure it's an issue that would have a resolution. bzr and git simply have different ideas of what constitutes a changeset when it comes to copying over files in a source-controlled directory...23:20
justinsb_termie: I still think git-bzr-ng is great :-)23:21
jaypipesjustinsb_: agreed, it's a pain, but not sure there's a way around that. termie, do the migration files have to start with a number? I was assuming they did..23:21
vishyjustinsb_: BASIC naming ftw23:21
*** adiantum has quit IRC23:22
jmckenty10 REM I like Openstack23:22
termiejaypipes, justinsb_: i don't know a specific spot in the docs that claims it as a requirement23:22
jmckenty20 PRINT "OpenStack FTW"23:22
jmckenty30 GOTO 1023:23
* soren wonders if vim has syntax highlighting for BASIC23:23
termiejaypipes, justinsb_: but all the docs and tooling use that naming scheme23:23
termiesoren: yup23:23
vishytermie: can we just name them all 003_?23:23
jaypipestermie, justinsb_: hmm, well perhaps we can backburner a "fix" for that naming for a later day...23:23
termiesoren: just vim test.bas23:23
vishyso that we know that they are all part of cactus?23:23
sorenThey just need to sort correctly.23:24
termievishy: do we want migrations to work per-revision?23:24
jaypipesvishy: possible, but how would migrations know to update the database version on a new rev in trunk?23:24
kpepplejk0: sorry, keep getting interrupted. 30 minutes.23:24
sorentermie: I'd say yes.23:24
termiesoren: me too23:24
* termie universe implodes23:24
jaypipestermie: yeah, I think definitely..23:24
sorentermie: Come release time maybe we can cold them all into one.23:24
creihtmtaylor: seems we don't have access to approve merges to the 1.2 series23:24
sorentermie: ...but during the dev cycle, we shouldn't punish people for running trunk.23:25
jk0kpepple: all of my work is in your branch, so we just need to propose the merge to trunk23:25
jk0kpepple: unless there was some other stuff you wanted to go over as well?23:25
justinsb_I know you can't name them all 003_ because that's what I had and the unit tests failed immediately...23:25
mtaylorcreiht: dammit man. I fixed that already for nova today - aren't you guys the same thing?23:25
vishysoren: lol @ footnote in email23:25
*** sebastianstadil has joined #openstack23:25
mtaylorcreiht: try again23:26
*** eikke has joined #openstack23:26
jaypipesvishy: yeah, soren the librarian...23:26
* jaypipes takes note to start using footnotes in his emails... just looks like he knows what he's doing...23:27
termiejustinsb_: so the issue for the moment is that we have to name the files sequentially23:27
jaypipesvishy: but yes, the monkey-patching of eventlet would indeed be very Inception.23:27
termiejustinsb_: i am not sure whether i would want to have and just keep updating it or make a bunch of files23:28
vishytermie, justinsb_: I think having to rename the files if someone else gets something in first is not too big of an issue23:28
justinsb_vishy: I'm happy to do it for now, but I don't think it's a brilliant solution23:29
termievishy, justinsb_: my concern is in a situation where we revert what something should do, we have to migrate to something and then away from it23:29
termievishy, justinsb_: if we go the separate file route23:29
vishywe can combine them all before the cactus release23:30
termievishy, justinsb_: if we do it the other way it will be trickier for people to migrate a trunk23:30
justinsb_Who proposed the separate files again?  I think they said Glance had solved this problem... (?)23:31
*** mdomsch has joined #openstack23:31
sorenvishy: \o/ I was wondering if anyone would bother reading that far down in that e-mail :)23:32
*** adiantum has joined #openstack23:32
termiejustinsb_: glance looks to be doing the separate files approach23:33
sorenmtaylor: If you want to distribute the load for this stuff a bit, I wouldn't mind having the credentials for the openstack-hudson lp user.23:33
jaypipesjustinsb_: glance has identical issue.23:33
*** mray has quit IRC23:34
jaypipesjustinsb_: meaning, contributors may need to rename if there is a merge logjam...23:34
jaypipesjustinsb_: we just don't have that many contributors ;)23:34
*** that__guy has quit IRC23:35
jmckentyjaypipes: define many23:35
justinsb_jaypipes, termie, vishy: Well it sounds like termie has a separate issue, but for now I guess we just keep going with 'pick next number and hope!'23:35
jaypipesjmckenty: 2? :)23:35
jmckentyhey, the Inception joke is mine. I'm going to start collecting royalties23:35
jmckentyno, no. It's 1, 2, many23:36
jaypipestermie: your concern above about reverting... could you elaborate? couldn't one write a migration .py file that simply removed the added tables or columns?23:37
termiejaypipes: you could do exactly that, it just sounds like it would create a somewhat silly situation23:37
termiejaypipes: not a technical problem23:38
*** gregp76 has quit IRC23:38
vishytermie, jaypipes: but we can clean them up when we release23:38
jaypipesya, agreed.23:39
termievishy: that doesn't really solve the issue of updating from trunk to a release, but i don't know how big of an issue that is23:39
kpepplejk0: did you see 721005 ?23:40
vishytermie: ah i see.  Well it happens in version control all the time anyway23:40
jk0kpepple: I don't think so23:40
vishybug 72100523:40
uvirtbotLaunchpad bug 721005 in nova "Incompatible comparison on flavor_id when using the Openstack API with XML request body" [Undecided,New]
sorentermie, vishy: What we do in Ubuntu for these sorts of things is to keep them up until release. Right after release, we'd fold them into one.23:41
sorentermie, vishy: Updating from a dev version of cactus to cactus final should work. From a dev version of cactus to diablo final is not a supported upgrade scenario.23:41
*** mray has joined #openstack23:41
vishythat makes some sense23:42
zedasannegentle: ok go ahead. i do have other work to do first.23:42
sorenSo released cactus would contain the lot, but post-cactus-release, we fold all the ones from the cactus cycle into a great big cactus migration script.23:42
justinsb_But SQLAlchemy isn't smart enough to cope with changing version scripts in development?23:44
jaypipessoren: so Cactus release rev + 1 would be the consolidation of the migration scripts?23:44
jaypipessoren: and Cactus release rev + 1 == Diable rev 0 ....23:45
sorenjaypipes: Yup.23:45
*** jesse_ has quit IRC23:46
sorenjustinsb_: Not sure I understand.23:46
jaypipessoren: k. that sounds reasonable to me.23:46
justinsb_soren: If we just had 'one big script' per version, even in trunk/active development, could SQLAlchemy Migrate still do our migrations for us in development?23:46
vishyjustinsb_: i don't think so23:47
sorenjustinsb_: No.23:47
vishyjustinsb_: because if someone had a partial version, then it would fail23:47
sorenjustinsb_: If it could, we could also just have "one big script" for migrations from release to release.23:48
sorenjustinsb_: It's really no different.23:48
rds__hi guys, i'm trying to install nova using the novascript but i have some problems, can you help me please23:48
justinsb_soren, vishy: True.. but if it actually worked, it would be 'enterprise' and we'd have to convert it to Java :-)23:48
justinsb_soren, vishy: I'm wondering if, for development, we could just rebuild the DB every time or do something like that23:49
sorenjustinsb_: Again, I'd hate to punish people for running trunk.23:49
justinsb_soren: Fair point23:50
sorenjustinsb_: Having to blow away your DB every once in a while is rather discouraging, I think.23:50
*** rnirmal has quit IRC23:50
vishysoren, justinsb_: doesn't really affect me, because i don't try to upgrade my db.  I just recreate it each time via or chef23:50
sorenvishy: I do both.23:50
justinsb_vishy: I think I use your script and do the same thing :-)23:50
*** jmckenty has quit IRC23:51
*** mray has quit IRC23:51
sorenvishy: I first upgrade, run my tests, then blow everything away, install from scratch, run tests again.23:51
*** dendro-afk is now known as dendrobates23:51
justinsb_Bugger... the zone tables migration just stole my #004 :-(23:51
sorenvishy: I want both scenarios to work.23:51
justinsb_I can confirm that having two things numbered the same _definitely_ fails23:52
sorenvishy: Ideally, I want to have a way to ensure that upgrades from any previous revision works, but from rev=rev-1 is as good as it gets right now.23:52
soren...and with that, I bid you all a good night.23:52
*** adiantum has quit IRC23:53
vishysoren: migration scripts use version control!23:53
vishysoren: good night.23:55
*** jmckenty has joined #openstack23:56
kpepplejk0: did you want to wait for that fix to land before we push ?23:57
jk0I'm comfortable moving ahead for now23:57
*** adiantum has joined #openstack23:58
* vishy waits for the push23:59

Generated by 2.14.0 by Marius Gedminas - find it at!