Wednesday, 2011-04-13

*** mahadev_ has quit IRC00:06
*** lionel has joined #openstack00:06
*** ccustine has quit IRC00:20
*** sophiap has quit IRC00:21
*** shentonfreude has quit IRC00:24
*** reldan has quit IRC00:25
*** sophiap has joined #openstack00:27
*** joearnold has quit IRC00:30
*** reldan has joined #openstack00:33
*** RJD22 is now known as RJD22|away00:36
*** RJD22|away is now known as RJD2200:40
*** mahadev has joined #openstack00:41
*** cw_ has quit IRC00:42
*** ke4qqq has quit IRC00:42
*** maplebed has quit IRC00:42
*** Ryan_Lane|food is now known as Ryan_Lane00:44
*** cw has joined #openstack00:45
*** ke4qqq has joined #openstack00:45
*** mahadev has quit IRC00:51
*** mahadev has joined #openstack00:58
*** RickB17 has joined #openstack01:02
*** adjohn has joined #openstack01:05
*** benbenhappy has joined #openstack01:05
*** MarkAtwood has quit IRC01:09
*** aliguori has quit IRC01:15
*** jeffjapan has quit IRC01:16
*** ryker has quit IRC01:20
*** clauden_ has quit IRC01:28
*** j05h has quit IRC01:28
*** rchavik has joined #openstack01:29
*** drico has quit IRC01:32
*** rds__ has quit IRC01:32
*** jeffjapan has joined #openstack01:33
*** littleidea has quit IRC01:40
*** RJD22 is now known as RJD22|away01:40
*** RJD22|away is now known as RJD2201:47
*** j05h has joined #openstack01:49
*** reldan has quit IRC01:50
*** shentonfreude has joined #openstack01:55
*** RickB17 has quit IRC01:57
*** RickB17 has joined #openstack02:02
*** reldan has joined #openstack02:05
*** Ryan_Lane has quit IRC02:06
*** eskp has quit IRC02:17
*** eskp has joined #openstack02:17
*** littleidea has joined #openstack02:21
*** littleidea has quit IRC02:25
*** RickB17 has quit IRC02:38
*** chuck_ has joined #openstack02:59
*** zul has quit IRC02:59
winston-dDoes anyone know how to remove 'network' from Nova database?  I mean is there any command to do that rather than altering the database entries manually?03:02
*** adjohn_ has joined #openstack03:05
*** adjohn has quit IRC03:07
*** adjohn_ is now known as adjohn03:07
*** mdomsch has joined #openstack03:09
*** joearnold has joined #openstack03:14
*** littleidea has joined #openstack03:16
*** littleidea has quit IRC03:19
*** hadrian has quit IRC03:20
*** hagarth has quit IRC03:26
*** AimanA is now known as HouseAway03:34
*** dendrobates is now known as dendro-afk03:41
*** lamar has joined #openstack03:42
*** dendro-afk is now known as dendrobates03:44
*** rohitFreeOSS has joined #openstack03:45
rohitFreeOSSAny openstack developer based in India?03:48
*** lamar has quit IRC03:50
HugoKuo_winston-d : 我總是從database empty 掉03:55
*** adjohn has quit IRC03:56
HugoKuo_winston-d : did you install phpmyadmin ?03:56
*** adjohn has joined #openstack03:56
*** mdomsch has quit IRC03:57
*** dendrobates is now known as dendro-afk04:02
*** dendro-afk is now known as dendrobates04:03
*** maplebed has joined #openstack04:07
*** adjohn has quit IRC04:09
*** adiantum has quit IRC04:09
*** adjohn has joined #openstack04:10
*** kashyap has joined #openstack04:23
*** miclorb_ has joined #openstack04:24
*** adiantum has joined #openstack04:26
*** JuanPerez has quit IRC04:30
*** smeier00 has joined #openstack04:34
*** littleidea has joined #openstack04:40
*** maplebed has quit IRC04:47
*** f4m8_ is now known as f4m804:49
*** shentonfreude has quit IRC04:51
*** kashyap has quit IRC04:51
*** miclorb_ has quit IRC04:59
*** mihgen has joined #openstack05:01
*** sophiap has quit IRC05:07
*** miclorb_ has joined #openstack05:07
*** joearnold has quit IRC05:10
*** rohitFreeOSS has quit IRC05:15
*** oneiropolo has joined #openstack05:16
*** mihgen has quit IRC05:16
*** rchavik has quit IRC05:16
oneiropolois there any swift dev?05:17
oneiropoloi want to ask about number of worker threads for proxy and storage05:17
winston-doneiropolo : i asked them similar questions.  maybe I can help05:19
oneiropolohow many worker threads are suitable for proxy server.05:20
winston-dHugoKuo_ : glad to see openstack-china community formed.05:20
oneiropoloi'm doing some kind of performance test05:20
winston-doneiropolo : the suggestion i got was 2 x the number of cores of your proxy server.05:20
oneiropolothe proxy-server process is running on a node05:21
oneiropoloand i made a prototype client test program.05:21
oneiropolothat program, make many threads concurrently and request swift put or get object05:21
oneiropolothe maximum number of threads is 60 at this time.05:22
oneiropolohowever, what i want to ask is05:22
oneiropolothe cpu usage of proxy server is05:22
oneiropololess then 5 %05:22
oneiropoloeven if i made 60 threads and 120 threads.05:22
oneiropolothe cpu usage of proxy server is not going to increase.05:23
winston-doneiropolo : that's possible. in the case that your storage nodes are the bottleneck.05:23
*** smoser has quit IRC05:23
*** mahadev has quit IRC05:23
oneiropoloso i check the cpu usage of storage node too.05:24
winston-doneiropolo : how many storage nodes do you have?05:24
oneiropoloand they are same.05:24
oneiropolothey just use less then 15 % of cpu usage05:24
winston-doneiropolo : bottleneck on storage nodes doesn't have to be on processors, it could also be your harddisks are reaching their peak throughput.05:25
oneiropoloat this time i have 4 storage node and05:25
oneiropolo1 node have many disks.05:25
oneiropolohas ;05:25
winston-dso how many disks do you have for each node? what type of disk? SATA HDD or SAS or SSD?05:25
*** ramkrsna has joined #openstack05:26
*** dendrobates is now known as dendro-afk05:26
winston-dwhat kind of tests your client is doing? upload/download/delete objects?05:26
winston-dwhat performance number do you have?05:27
oneiropoloupload and download05:29
oneiropolo1 node has SAS disks.05:29
oneiropolo1 node has 12 SAS disks05:29
winston-doneiropolo : sorry for asking that many questions but these information is important to determine bottleneck05:29
oneiropolono, it's ok05:30
winston-dso you have altogether 48 SAS disks for storage, right? 10krpm type?  how many zones do you have for swift?05:30
oneiropolo4 zones05:30
oneiropolo1 node is 1 zone and 1 zone has 12 disks05:31
winston-dand how many replication? 3 by default?05:32
winston-dwhat is size of objects you were testing?05:33
oneiropolo512k, 1m, 2m, 5m05:33
oneiropolowhat is05:34
winston-dsorry, that doesn't ring a bell to me. any context?05:34
*** mihgen has joined #openstack05:34
oneiropolowhat is CTCP stand for?05:35
winston-dclient-to-client-protocol, maybe? where do you see it?05:36
oneiropoloHTTP v1.005:36
oneiropolousing java client program i coded05:37
oneiropoloi used apache HTTPClient package05:37
*** littleidea has quit IRC05:37
*** JuanPerez has joined #openstack05:38
winston-dhow many uploads/downloads can u do for ur setup? take 512k for example.  maybe it's also a good idea to try smaller object size to see if disk IOPS is the bottleneck.05:38
*** smoser has joined #openstack05:39
winston-ddid you check 'iostat' data on ur storage nodes? how many IOPS did u have?05:39
oneiropolounder 45 M/sec for 512k objects05:39
oneiropolobut what i want to know is that cpu is not 100% working .05:40
oneiropoloswift proxy doesn't use the cpu fully.05:40
oneiropoloso, i guess that if that use the cpu more efficiently the IOPS will goes up.05:40
winston-dOK. but i won't blame proxy-server so soon.  can you try much smaller object size, say, 512Byte?05:44
oneiropoloi can05:44
oneiropolobut, it will take some time.05:46
*** rsaidan has joined #openstack05:50
*** rchavik has joined #openstack05:52
*** smeier00 has quit IRC05:59
*** omidhdl has joined #openstack05:59
*** rsaidan has quit IRC06:00
JulioBellI have a meeting the 14, in Madrid with the openstack spanish group06:01
*** benbenhappy has left #openstack06:08
*** guigui1 has joined #openstack06:19
*** Gaelfr has joined #openstack06:23
*** CloudChris has joined #openstack06:26
*** zenmatt has quit IRC06:31
*** adiantum has quit IRC06:34
*** kashyap has joined #openstack06:41
*** lionel has quit IRC06:52
*** lionel has joined #openstack06:53
*** rsaidan has joined #openstack07:00
*** daveiw has joined #openstack07:00
*** gregp76 has joined #openstack07:00
*** rcc has joined #openstack07:01
*** Ryan_Lane has joined #openstack07:10
*** antenagora has joined #openstack07:23
*** uksysadmin has joined #openstack07:26
*** martinr has quit IRC07:27
*** Ryan_Lane has quit IRC07:29
*** daveiw has quit IRC07:34
*** DogWater has quit IRC07:41
*** mgoldmann has joined #openstack07:42
*** drico has joined #openstack07:43
*** bkkrw has joined #openstack07:45
*** allsystemsarego has joined #openstack07:50
*** nacx has joined #openstack07:53
*** ibarrera has joined #openstack07:53
*** ibarrera has quit IRC07:53
*** mihgen has quit IRC07:56
*** nagyz has joined #openstack07:58
nagyzhey there07:58
nagyzinternally, how can I figure out the instance id used by libvirt, if I have the instance id used by openstack?07:59
ttxsoren: any chance you could review the two remaining Nova RC BMPs ?08:05
*** martinr1 has joined #openstack08:08
*** martinr1 has left #openstack08:09
sorennagyz: It's that same?08:10
nagyzI must have been too tired yesterday then08:11
nagyzbtw, once someone gets the hang of it, the code is really well structured08:11
nagyzit was a piece of cake to extend the existing compute API, and to add a new API to nova-api08:12
nagyzso congrats on that08:12
sorenttx: I'll try. I've been having trouble jump-starting my brain this morning.08:12
sorenttx: Before you cut the RC, we need to toggle the "final" flag in
nagyzoh, right. so compute_api.get_all() returns a list of instances, and instance["id"] there equals to 3, not instance-00000003, but for libvirt I need to pass the second form08:13
ttxsoren: and the final translations branch.08:19
ttxsoren: I planned to push those tomorrow morning for Glance/Nova. will you be around ?08:20
*** CloudChris has left #openstack08:21
sorenttx: Tomorrow? Sure.08:24
sorennagyz: Right.08:24
nagyzI see that "hostname" is set to that. can I rely on that?08:24
sorennagyz: .aka. instance['name'].08:25
nagyzah, okay08:27
ttxsoren: looking at zero-volumes, we'll need to add "dd" to nova_sudoers08:28
*** daveiw has joined #openstack08:38
openstackjenkinsProject nova build #814: SUCCESS in 2 min 31 sec:
openstackjenkinsTarmac: Updates the documentation on creating and using a cloudpipe image.08:39
rsaidanhey everyone.. what repository should I be running to test nova? Is trunk stable enough for testing purposes or should i be using bexar?08:40
jeffjapanrsaidan: the next release (cactus) is this week, I'd use that instead of bexar08:44
uvirtbotNew bug: #759572 in swift "get_hashes in objrep should use utils.write_pickle" [Undecided,New]
rsaidanjeffjapan: thx, so would you not recommend using trunk as i am now?08:46
jeffjapanthe quality of trunk varies widely depending on when exactly you get it, so you may be great, or you may get something that's very broken08:47
rsaidani see.. if i've installed trunk repos, what is the best way to remove all packages and reinstall bexar?08:48
rsaidani'm getting various errors and think i should move to a stable release for testing08:49
HugoKuo_apt-get autoremove --purge08:49
*** xavicampa has joined #openstack08:49
HugoKuo_and then del the source08:49
jeffjapanand /var/lib/nova08:49
HugoKuo_then add Bexar repos again08:50
rsaidanawesome, thanks guys08:50
rsaidananother quick question... should I remove the database too?08:50
*** mihgen has joined #openstack08:50
jeffjapanbut I'd wait for the cactus RC if at all possible...08:50
jeffjapanI think it's tomorrow08:50
rsaidanany idea when that will be available?08:50
rsaidanoh great08:50
HugoKuo_no need but you have to drop ....nova table or just empty it08:51
jeffjapanbexar is really quite old at this point08:51
rsaidanok, another question regarding db.. if nova is empty I use nova-manage db sync to create tables right? what happens when this is run on a live/worklng system?08:51
jeffjapanI've used it to clean out some cruft in the DB before and it worked fine for that.08:52
jeffjapanremember that glance has its own db though08:52
*** omidhdl has quit IRC08:52
rsaidanso you only need to run it is you're experiencing issues with the db or table names etc?08:53
jeffjapanor if you hit a bug where something that should be removed from the DB wasn't.08:54
rsaidanok, thanks for the help08:55
jeffjapangood luck08:55
*** jeffjapan has quit IRC08:56
HugoKuo_rsaidan : don't forget to apt-get autoclean...09:02
*** dirkx_ has joined #openstack09:08
*** dodeeric1 has joined #openstack09:11
*** dirkx__ has joined #openstack09:12
*** dirkx_ has quit IRC09:14
*** dirkx__ has quit IRC09:16
*** dirkx_ has joined #openstack09:16
*** dirkx_ has quit IRC09:20
*** gregp76 has quit IRC09:21
*** gregp76 has joined #openstack09:22
*** dirkx_ has joined #openstack09:22
HugoKuo_Need I nfs mount /var/lib/nova/images from nova-objectstore to nova-compute host in the latest trunk ??09:22
*** dirkx__ has joined #openstack09:25
*** dirkx_ has quit IRC09:26
*** adjohn has quit IRC09:36
*** zigo-_- has joined #openstack09:41
nagyzif I return an xml thru the API, < gets encoded as &lt, < as &gt, and so on, basically like html09:42
nagyzis there a way to make it stop doing that?09:43
nagyzespecially since Content-Type is application/xml in the reply09:43
*** gregp76_ has joined #openstack09:44
*** gregp76 has quit IRC09:45
*** gregp76 has joined #openstack09:45
*** dirkx__ has quit IRC09:48
*** kashyap has quit IRC09:55
*** kashyap has joined #openstack09:55
*** kashyap has quit IRC09:56
*** kashyap has joined #openstack09:57
*** guynaor has joined #openstack09:58
*** kashyap has joined #openstack09:58
*** rds__ has joined #openstack10:06
sorennagyz: What are you trying to do, exactly?10:21
* soren lunches10:27
*** arun_ has quit IRC10:28
*** dirkx_ has joined #openstack10:35
patri0tany documents about nova-instancemonitor ?10:35
patri0tsoren: do you know if there is any?10:36
nagyzsoren added a new API on a new port, and there I'd like to return an XML as a map value, basically like a string10:36
nagyzthe problem is that it gets encoded into html10:36
nagyzeven if I write it out manually :)10:36
*** smaresca has quit IRC10:38
nagyzlets say I do this:10:41
nagyzreturn dict(servers=[dict(server='<tag>hello</tag>')])10:41
nagyzat index()10:41
nagyzthen with curl, this is what I get:10:41
nagyz <server>10:42
nagyz            &lt;tag&gt;hello&lt;/tag&gt;10:42
nagyz        </server>10:42
nagyzso its an encoding issue, I just don't know where I should specify utf810:42
*** kashyap has quit IRC10:43
ccookeso... what *is* the "CLOUD_SERVERS_USERNAME" openstack-dashboard wants as a configuration value?10:45
ccookeI can't find any references to it outside that10:45
*** lionel has quit IRC10:46
*** lionel has joined #openstack10:46
n1md4ccooke: any user with admin privileges in your nova database (afaik).10:47
*** dirkx__ has joined #openstack10:48
ccooken1md4: so it *is* a nova user? good to know.10:48
ccookeShame that doesn't work :-/10:48
n1md4ccooke:  I managed to get it working, mostly.  Possibly I used 'root'?10:50
ccookeI've tried that10:50
ccookeevery user I try gets me a permission denied from nova10:50
*** smaresca has joined #openstack10:50
ccooke(well, 403 forbidden)10:51
*** dirkx___ has joined #openstack10:51
*** dirkx_ has quit IRC10:52
*** daveiw has quit IRC10:53
n1md4I have a lot of permission errors with nova, throw credentials and keys in the mix, it's a daughting mix for the new user (me).10:53
*** dirkx__ has quit IRC10:53
n1md4ccooke: which user's ACCESS and SECRET key did you use, in
ccookea user I created for the dashboard - called dashboard, in fact :-)10:54
ccookeDouble and triuple checked the local settings - they work11:04
n1md4With 1 user and 1 project, I definitely had a working dashboard interface.11:08
ccookenot here11:08
ccookethe user works from the command-line - I can upload images and add them11:09
ccookehaven't tried to create an instance yet11:09
ccooke... yes, works from the cli11:11
*** ctennis has quit IRC11:11
sorenpatri0t: Not that I know of.11:12
sorennagyz: Well, if you're adding a new API, how should I know what you're doing wrong?11:13
sorennagyz: You avoid escaping by not applying any. :)11:13
nagyzI didn't. :)11:15
nagyzeven if I modify the openstack API, it does the same thing11:15
nagyzif I return an xml as a string somewhere, it gets escaped automatically11:15
sorenWell, naturally.11:15
nagyzthe client provides Accept: application/xml, so I don'11:16
nagyzt see the need to do the escaping11:16
sorenAnd it gets xml.11:16
ccooken1md4: if I have a user called dashboard - which works - with the correct name and keys in the, and a project called "admin" that works and has been able to spawn a VM... then creating the dashboard db with "sudo tools/ dashboard/ syncdb" and specifying my user as "dashboard" should just work, correct?11:16
sorenYou say so yourself.11:16
nagyzbut wrongly.11:16
*** arun_ has joined #openstack11:16
soren<server> <-- That's XML.11:16
nagyzand? the string is escaped11:17
sorenAnd so it should.11:17
nagyzwhy should it be?11:17
nagyzI'm trying to get the output of dumpxml from virsh, btw, that's why its so annoying11:17
ccookenagyz: because the XML schema defines the data in there as non-xml11:17
nagyzthere's no schema anywhere11:18
ccookenagyz: therefore any <> symbols (and a few others) *must* be escaped to preserve them11:18
sorenIf I want to return <sometag>a string</sometag>, I need to escape "a string" to make sure my XML remains valid.11:18
ccookenagyz: there are implicit schemas even if there are no official ones11:18
sorenXML serialisation happens exactly *once*.11:18
nagyzwhat would be the easiest way then to add virsh's dumpxml into an APi call, then?11:19
sorenIt sounds liek you already have it.11:19
sorenHow do you expect to consume this?11:19
nagyzfrom a java client11:19
sorenShow me a java client that won't automatically unserialize that stuff.11:19
*** shentonfreude has joined #openstack11:23
*** dirkx_ has joined #openstack11:24
*** BK_man has quit IRC11:25
*** ctennis has joined #openstack11:25
*** dirkx___ has quit IRC11:28
n1md4ccooke: I'd like to think so.11:29
ccookemyself also :-/11:29
*** dirkx__ has joined #openstack11:29
ccookeHowever, I get: DEBUG:boto:403 Forbidden11:30
ccookeTime to check the logs, I guess11:30
nagyzsoren, it might parse it, but still. looks ugly to my eyes.11:31
*** adjohn has joined #openstack11:31
sorennagyz: I you feel *exceptionally* bored, you could add a CDATA encoding thing, but it sounds like an utter waste of time.11:32
*** mgoldmann has quit IRC11:32
ccookeOperationalError: (OperationalError) database is locked11:32
ccookeThat doesn't look good11:32
*** dirkx_ has quit IRC11:32
ccooke... ah. I think I see11:33
ccookeit's ignoring the project specified in the file11:33
ccookeand asking for a project called "dashboard"11:34
*** adiantum has joined #openstack11:34
ccookeyes, that's it11:35
mihgenI've connected glance to nova, but it seems to be I'm missing some info on how to add new image to it.  I ran: glance -v add name="MyTestTTY" is_public=true container_format='ami' disk_format='ami' < /root/tty.tgz,  then made sure it's in 'nova image-list': | 1  | MyTestTTY | ACTIVE | , ran the instance from this image, but can't connect to it.11:36
mihgenDoes anyone know what I've missed?11:36
*** mgoldmann has joined #openstack11:36
mihgenqemu process is running, but console.log is empty11:37
*** shentonfreude has quit IRC11:37
*** gaveen has joined #openstack11:38
*** gaveen has joined #openstack11:38
ccookemihgen: 64-bit image on 32-bit host?11:38
ccooke(Since that was my issue yesterday :-)11:38
mihgenccooke: well... I don't even know where I specified arch)11:39
mihgenit runs under qemu-system-x86_64 command11:39
ccookemihgen: well, is your system 32-bit?11:39
mihgenhost is 64bit11:40
ccookethen that's not it :-)11:40
mihgenI ran the same image tty.tgz from local storage yesterday, it works fine. Today I tried to add the same image to glance.11:40
mihgenccooke: no problem, will dig into)11:41
ccookemihgen: the kvm output might help11:42
ccookeshould be in /var/log/libvirt/qemu/11:43
*** gaveen has quit IRC11:43
*** flopflip has joined #openstack11:44
ccooke... hmm11:45
mihgenccooke: nothing( It just reports starting up11:45
*** lynxman has joined #openstack11:45
*** dirkx_ has joined #openstack11:45
ccookeStarting VMs now, I'm getting db.NoMoreNetworks raised...11:45
ccookethere *is* a network confuigured11:46
*** drico has quit IRC11:46
*** nagyz has quit IRC11:46
*** flopflip_ has quit IRC11:47
*** dirkx__ has quit IRC11:48
*** mgoldmann has quit IRC11:52
*** gaveen has joined #openstack11:52
*** gaveen has joined #openstack11:52
*** mgoldmann has joined #openstack11:55
*** BK_man has joined #openstack11:55
*** CoreyM-afk is now known as CoreyM11:57
*** s1cz has quit IRC12:02
*** mihgen has quit IRC12:02
*** s1cz has joined #openstack12:02
*** dirkx_ has quit IRC12:03
*** reldan has quit IRC12:05
*** antenagora has quit IRC12:05
*** RickB17 has joined #openstack12:09
*** dendro-afk is now known as dendrobates12:09
*** uksysadmin has quit IRC12:11
*** antenagora has joined #openstack12:12
*** aliguori has joined #openstack12:14
n1md4ccooke: when creating creds do you use root or sudo?12:18
*** dodeeric1 has left #openstack12:21
*** daveiw has joined #openstack12:21
*** RichiH has joined #openstack12:23
*** h0cin has joined #openstack12:25
*** armmer has left #openstack12:27
*** chuck_ is now known as zul12:37
*** zul has joined #openstack12:37
*** dspano has joined #openstack12:41
*** lynxman has quit IRC12:41
*** crago has joined #openstack12:46
*** dprince has joined #openstack12:49
*** antenagora has quit IRC12:50
ttxannegentle: I did a pass on features in, feel free to beautify them like you did for Swift.12:54
uvirtbotNew bug: #759762 in nova "Observing warning in nova-compute logs "WARNING nova.compute.manager [-] Error during instance poll:"" [Undecided,New]
*** antenagora has joined #openstack12:56
*** ppetraki has joined #openstack12:56
*** adjohn has quit IRC12:59
*** Shentonfreude has joined #openstack12:59
annegentlettx: heh, beautification process started...12:59
*** lionel has quit IRC13:02
*** lionel has joined #openstack13:02
ccooken1md4: sudo13:02
*** mray has joined #openstack13:04
ccooken1md4: but if you didn't see, I solved that problem13:05
*** kashyap has joined #openstack13:05
*** icarus901 has joined #openstack13:08
*** Gaelfr has quit IRC13:09
*** rsaidan has quit IRC13:10
n1md4ccooke: I've just installed dashboard again.  I've noticed that there's no project, even though there's a project table row in mysql.13:17
ccookethe documentation for the dashboard really needs some cleanup, though13:18
ccookeI mean, the commands it tells you to run are *correct*, but won't actually work :-)13:18
annegentleccooke: say, go ahead and update the wiki and I'll try to get updates rolled into lp:openstack-manuals as well13:20
ccookeannegentle: I was about to ask if the former would be doable. The latter would be even better13:21
ccookehowever, I do not currently have an account13:21
ccookeand so I'm not sure I'd be able to :-)13:21
*** oneiropolo has quit IRC13:21
ccookeheh. Yeah, thought so. The page is immutable anyway.13:22
annegentleccooke: should be super easy to set up a wiki account, Launchpad and the CLA are for bzr updates13:22
ccookeannegentle: I can't edit the page, but I *can* give someone with permissions a patch for it13:23
annegentleccooke: weird, why does it appear immutable to you?13:23
annegentleccooke: and feel free to email me notes for changes at, I should be able to update.13:24
ccookeLeft hand menu under Page, "Immutable Page"13:24
*** gondoi_ has joined #openstack13:24
ccookethat and the lack of edit functionality gives me a clue :@-)13:24
ccookeokay, will do13:24
annegentleccooke: well yeah all pages are immutable when you're not logged in afaik13:25
annegentleccooke: such big words for early morning, though, immutable. :)13:26
ccookeit's early afternoon for me :-)13:26
annegentleccooke: fun stuff! I haven't installed the Dashboard since last release so I'm behind the times.13:26
annegentleccooke: yep I figured :)13:26
ccookeWell, I can update the page, then - and I can add some natty-specific bits13:27
annegentleccooke: awesomesauce, tahnks13:28
annegentleccooke: thanks even13:28
*** patcoll has joined #openstack13:28
ccookefirst, though, I should submit someone a bug and patch to fix the python2.6 hardcoding13:28
*** guynaor has left #openstack13:28
ccookeannegentle: I take it bugs go to launchpad?13:33
*** adjohn has joined #openstack13:36
*** reldan has joined #openstack13:36
*** dweimer has quit IRC13:37
*** Zangetsue has joined #openstack13:39
zulwhen is the deadline for blueprints?13:40
*** santhosh has joined #openstack13:42
*** mahadev has joined #openstack13:46
*** hadrian has joined #openstack13:47
*** Nick0la has joined #openstack13:47
Nick0laHello.  I am working on setting up compute.  I am trying to set up the database.  I type:   nova-manage db sync   and I it simply says that I need to use "one of the following options.."   Annoyingly, 'db' is one of the options.  What am I missing>13:49
n1md4for some reason, i now keep hitting cert, cred, key, permission problems!  I've not a clear perspect of what key/cred/cert/perm relates to what part of the stack.  I appreciate this is still a developing system, but is there documentation to clearly explain the differences?13:50
ccooken1md4: where are you hitting these problems?13:50
uvirtbotNew bug: #759814 in openstack-dashboard "python2.6 hardcoding breaks installation on Natty/py2.7 platforms (patch included)" [Undecided,New]
*** mahadev has quit IRC13:52
n1md4firstly, it was publishing an image, I got "Unable to run euca-describe-images.  Is euca2ools environment set up?".  I'd got over that, by exporting some access and secret keys.13:52
*** kakella has joined #openstack13:52
*** kakella has left #openstack13:53
*** dirkx_ has joined #openstack13:53
ccooken1md4: use the novarc from the project you're working on13:53
n1md4Then, trying to bundle/upload the kernel I got "Invalid certfailed to upload kernel"13:53
ccookeyeah, you need the whole set of data13:53
n1md4I did use that novarc, but perhaps I've funked up the environment variables somewhere.13:54
ccooken1md4: note that there are certificates in the file13:54
n1md4What's "the whole set of data"?13:54
ccookein the zipfile13:54
*** RickB17 has quit IRC13:56
* n1md4 scratches head13:56
n1md4It must be beer oclock soon!13:56
*** f4m8 is now known as f4m8_13:59
*** littleidea has joined #openstack14:02
*** dirkx_ has quit IRC14:02
Nick0lahmmm...more info for you for the debugging process:14:02
Nick0lathis is the -exact- error message:   "/usr/bin/nova-manage does not match any options:"14:03
*** littleidea has quit IRC14:03
Nick0lait is looking at it's own executable name...and not the commandline arguments.14:04
Nick0laI do not know python well enough to try and fix this...14:04
ccookeannegentle: I've made some changes. Once the bug I reported is fixed, I (or anyone else) can mark this as tested and working on Natty under python2.714:04
*** BK_man has quit IRC14:04
ccooke(Mostly I've just added a bit of help for what seem to be common problems, along with cleaning up the commands so they actually run from the right places :-)14:05
*** zenmatt has joined #openstack14:08
annegentleccooke: thanks for doing that114:09
annegentleccooke: big help14:09
ccookeno problem14:10
*** adjohn has quit IRC14:11
*** dirkx_ has joined #openstack14:11
*** rsaidan has joined #openstack14:16
Nick0lacould this be a python version issue?  what version of python is nova intended to run on?14:17
*** dirkx_ has quit IRC14:17
*** mdomsch has joined #openstack14:17
Nick0laI am completely stuck in my installation.  :-(14:18
ccookeNick0la: it's tested on 2.6 and it's running fine on natty's 2.7 here14:20
ccookewhat's the exact command you're using?14:20
*** dendrobates is now known as dendro-afk14:21
Nick0lathat should have all the info you need.  If not, I can provide more.14:22
*** dirkx_ has joined #openstack14:23
*** ryker has joined #openstack14:24
*** nagyz has joined #openstack14:27
Nick0laoh...there is a log entry.  I will add that to my sec..14:27
rykeri'm going to the openstack conference on the 26th.  Any recommendations for the best way to get from the San Francisco airport to the conference?  Should I just take a taxi or public transportation or is renting a car the best option?14:29
blamarryker: From what I understand the conference is much closer to SJC (San Jose), so maybe a shuttle from SFO to SJC and then public transport from there could work if you're looking to save some $ on car rental14:35
blamarryker: Then again, I'm from the east coast and I've never been to Santa Clara! :D14:36
blamarI'll be renting a car14:36
*** gondoi_ has quit IRC14:36
rykeryeah, I noticed it was closer to san jose as well and was thinking about that.14:36
rykeri didn't know how bad the traffic is around there and if renting a car was just a bad idea. :)14:37
rykerI think I'm just going to rent a car as well.  It's pretty cheap to rent a car anyway.14:37
blamarYeah, lots to see in the area too if you're in to that14:38
rykertrue.  and it's my first time in cali14:38
*** uksysadmin has joined #openstack14:38
*** littleidea has joined #openstack14:39
*** camm_ has quit IRC14:40
*** camm has joined #openstack14:40
*** rnirmal has joined #openstack14:41
markwashryker: I believe you can catch the caltrain at SFO which will take you down to SJC14:41
*** guynaor has joined #openstack14:43
*** guynaor has left #openstack14:43
gholtGoogle maps shows 36-45mins driving and 2.5hrs bussing from SFO to 5101 Great America Parkway, Santa Clara, CA 95054.14:45
gholt10.25hrs walking, if you're into that. :)14:46
*** dragondm has joined #openstack14:46
*** smeier00 has joined #openstack14:51
*** dendro-afk is now known as dendrobates14:51
*** jmckenty has joined #openstack14:51
*** jonkelly has joined #openstack14:53
*** jkoelker has joined #openstack14:56
*** ramkrsna has quit IRC14:59
jmckentyI'm on reviews for nova today, if there's any critical blockers please ping me14:59
jmckentyI'm speaking at a conference as well, of course, so I'll be on and off line15:00
*** dirkx_ has quit IRC15:00
rykerhmm, looks like I should walk if it's only 10.25 hrs. ;)15:01
rykeryeah, seems like a car is definitely the best bet.15:01
ttxjmckenty: ack15:01
glencryker: SJC airport is about 5 minutes from Santa Clara; if you have to come into SFO, you'll need a taxi or shuttle. Caltrain is a real pain, and it doesn't stop anywhere near the Santa Clara convention center/hotel.15:01
*** gondoi_ has joined #openstack15:03
rykerunfortunately, there are only direct flights from chicago to SFO.  All I could find to SJC have at least 1 stop, so I figured I would just drive from SFO.15:03
*** gondoi_ has joined #openstack15:03
glencpretty straight drive down the peninsula from SFO to Santa Clara15:03
rykerthx guys.  definitely renting a car now.  should be a nice drive anyway since it's my first time.15:04
jmckentyIf you're renting and driving, would you mind throwing a listing up on the wiki somewhere for carpool?15:07
jmckentyryker: ==^15:08
rykersure.  i'll post a link back here later.  my company is paying for everything and they haven't purchased the airplane ticket or car rental yet.15:09
*** guigui1 has quit IRC15:09
ttxnotmyname: around ?15:09
rykerprobably won't happen until later this evening.15:09
notmynamettx: yup15:10
nhm_At the lustre conference now.  Some interesting benchmarks about mapreduce on lustre vs hdfs earlier.15:10
ttxnotmyname: I think we should push 1.3-dev -> 1.3.0 in swift/ now15:10
ttxnotmyname: so that I can cut a release candidate tomorrow morning before the swifties get up.15:11
ttxnotmyname: want me to propose a branch merge proposal for it ?15:11
ttxnotmyname: ok, in progress15:11
*** gondoi__ has joined #openstack15:11
*** jmckenty has quit IRC15:13
*** gondoi_ has quit IRC15:15
notmynamettx: ya, got the email, but the link (ie LP) seems busted by it15:15
Nick0laI uninstalled nova and reinstalled.  see here for full details:   Still can't get nova-manage db sync to work.15:16
notmynameah. working now15:16
ttxyes, tere have been a number of transient oopses today15:16
*** gondoi__ has quit IRC15:16
Nick0la  <-- now includes contents of my nova.conf file15:21
*** gondoi_ has joined #openstack15:21
*** ccustine has joined #openstack15:21
Nick0lathere was an error in my config file!  I noticed it when copying.  One of my directives in nova.conf was not proceeded with '--'15:23
Nick0laMay I make a request?  Can we have nova check the syntax of the file?15:23
Nick0laand error if it's wrong?15:23
Nick0lanow nova-manage works fine.15:24
Nick0labut, ideally, it should warn when starting the service that something isn't right.15:24
Nick0laor...not start the service at all.15:24
notmynamettx: since this is a release candidate, it should be marked as 1.3-rc. After that, it can go to 1.3.015:25
*** daveiw has quit IRC15:25
*** nagyz has quit IRC15:26
uvirtbotNew bug: #759908 in swift "Log Processing Needs Atomic "Commit"" [Undecided,New]
ttxnotmyname: in fact, the idea is that the RC tarball can be used directly as the Golden release, that's why I think we should push final versioning at RC time. That's what we do for Glance and Nova.15:26
termiepvo: heya, i think we have a meeting in about 30 mins15:27
termiewondering what room that is15:27
termieor whatnot15:27
ttxnotmyname: note that there will be a RC and a Final tag in case they differ15:27
ttxnotmyname: otherwise the release candidate tarball is not really a release candidate.15:28
pvoare you still in 2059?15:28
*** gaveen_ has joined #openstack15:29
*** jakedahn has joined #openstack15:29
pvotermie: ^^15:30
*** gaveen has quit IRC15:32
notmynamettx: gotta run to a meeting. I want to talk more about it15:32
termiepvo: yup15:32
*** Zangetsue has quit IRC15:32
termieer no15:32
pvowoudl that room work?15:33
pvok, I'll update15:33
*** Zangetsue has joined #openstack15:33
ttxnotmyname: ok, I'll stop for the evening, but should be back for discussion at around 4pm your time15:33
*** antenagora has quit IRC15:35
*** jakedahn has quit IRC15:35
*** jakedahn has joined #openstack15:36
*** zykes- has quit IRC15:37
creihtttx: I think the short story is, if we don't include rc in the version number, how does someone know that they are running the rc or release?15:37
*** gondoi_ has quit IRC15:38
*** smeier00 has left #openstack15:38
ttxcreiht: right, the question is, is it worth the risk of regenerating a tarball at the last minute that will therefore be untested.15:38
* creiht sighs15:39
ttxcreiht: both solutions are suboptimal.15:40
ttxI tend to prefer the solution that simplifies release management :)15:41
ttxcreiht, notmyname: that said, I won't fight over it -- we can try to align after this one, discussing it over a beer.15:41
creihtevery project that I have seen that has release candidates always clearly indicates that the release is a rc (in packaging, etc.)15:42
ttxthat's because they kinda abuse the meaning of RC, imo.15:42
* creiht throws up his hands and gets back to work15:43
ttxwhat they mean is "final beta", not "something that could exactly be the release".15:43
creihtand leaves it in the capable hands of the ptl :)15:43
ttxcreiht: I'll let notmyname decide, in all cases.15:43
ttxcreiht: just explaining my pov :)15:43
*** jmckenty has joined #openstack15:44
* ttx disappears, back in ~5 hours15:45
*** rnirmal has quit IRC15:52
*** BK_man has joined #openstack15:55
BK_man - any thoughts?15:55
*** smeier00 has joined #openstack15:57
*** dprince has quit IRC15:58
*** dirkx_ has joined #openstack15:59
*** MarkAtwood has joined #openstack15:59
*** rcc has quit IRC16:01
jmckentyjustinsb: is still waiting on another fix, yes?16:02
*** mdomsch has quit IRC16:02
jmckentye.g., the "make pythonic" iteration against the checker subclasses16:02
jmckentyI just flipped it to wip, let me know if otherwise16:03
BK_manquestion: is euca-upload-bundle working with Glance? I'm getting error with starting instances both on Glance and local16:03
jaypipesBK_man: mind helping out the poor fellow on
jaypipesBK_man: what are the errors?16:05
jaypipesjmckenty: I would check with ttx before sending any merge proposal to tarmac... we're in RC freeze shortly.16:06
jmckentyjaypipes: ==^16:07
jmckentyWe should update Soren's "review-day alert" to include release status16:07
BK_manjaypipes: it might not be glance-related.16:07
jaypipesBK_man: interesting...16:08
*** jfluhmann has quit IRC16:08
*** Gaelfr has joined #openstack16:08
BK_manjaypipes: I'm starting each time from scratch (empty DB, new user / project / network/ etc)16:08
jaypipesBK_man: I've not seen that before in bug reports. vishy? have you seen a BadStatusLine come up in the bug reports when running euca-upload-bundle?16:08
*** ChanServ sets mode: +o jmckenty16:09
jaypipesBK_man: I assume you are on latest trunk, right?16:09
BK_manjaypipes: installing glance rev111 and trying to repeat all again16:09
BK_manjaypipes: yep16:09
jaypipesBK_man: I figured you were, just was checking :)16:10
*** zykes- has joined #openstack16:10
jaypipessirp_: quick review on please?16:11
BK_manjaypipes: vishy: the same. Still that error during euca-run-instances16:11
vishyBK_Man, jaypipes: i would check the key files in your zip 16:11
BK_manvishy: please explain. What files? Scripts are known to work on early releases and stopped working today's morning16:11
jaypipesvishy: just so I understand better, could you explain why you think the key file is the culprit? trying to better understand how to debug these issues...16:12
* BK_man going to see what stuff was merged last night16:13
vishyBK_man: are you getting fresh creds16:13
BK_manvishy: yep, indeed16:13
BK_manvishy: do u need to review my scripts?16:14
*** maxamillion has quit IRC16:14
*** alekibango has quit IRC16:16
*** maxamillion has joined #openstack16:16
*** bkkrw has quit IRC16:17
*** Gaelfr has quit IRC16:18
vishyBK_man: check pk.pem cert.pem and cacert.pem and make sure that they all have data that looks like keys16:18
*** alekibango has joined #openstack16:18
vishyperhaps you need to set some flags for CA folder?16:19
*** nwl has joined #openstack16:20
BK_manvishy: it's set: /var/lib/nova/CA16:20
vishyso check those files in your zip16:20
vishyand see if anything looks off16:20
BK_manvishy: nothing strange - just keys and certs16:21
uvirtbotNew bug: #759972 in swift "Install trunk debs on Ubuntu 10.10 fails with dup owned file between swift and swift-account" [Undecided,New]
* BK_man discovered that CA folder contents in Nova installation was not reaped by cleanup scripts. Let me fix that16:22
*** maplebed has joined #openstack16:22
*** jmckenty has quit IRC16:22
BK_manis it ok for Nova to start with completely empty CA folder?16:22
*** ryker has quit IRC16:23
BK_mandumb question16:23
*** ryker has joined #openstack16:23
BK_manof course there are should be openssl.cnf.tmpl at least16:23
*** benbenhappy has joined #openstack16:25
vishyBK_man: it does need a few things16:25
vishylook at nova/CA from source16:25
*** max_the_penguin has joined #openstack16:26
sirp_jaypipes: approved fix, had marked needs review, but turned out issue for me is part of trunk so not related to patch-at-hand16:29
uvirtbotLaunchpad bug 759018 in glance "When running tests without virtualenv, test_traceback_not_consumed fails" [Critical,In progress]16:29
sirp_jaypipes: yep16:29
sirp_jaypipes: thx16:29
jaypipessirp_: is unrelated to that patch...16:29
jaypipessirp_: has to do with the raising of the eventlet error in _safe_kill().16:30
jaypipessirp_: if you change _safe_kill() to simply raise the original exception, all tests pass...16:30
uvirtbotNew bug: #759978 in nova "Openstack API Rate limiting isn't flexible enough" [Wishlist,New]
* BK_man started nova-api but does not get CA generated...16:31
*** kbringard has joined #openstack16:32
*** rchavik has quit IRC16:32
kbringardhey guys, quick ?16:33
kbringardI googled around a bit, but couldn't find it... how do I specify multiple security groups when I launch an instance?16:33
kbringardI tried using -g multiple times, but it only took the first one16:33
*** stewart has joined #openstack16:33
*** kashyap has quit IRC16:34
kbringardand I tried adding them comma and space separated, but that doesn't work either16:34
jaypipessirp_: the really strange thing is that the failure only occurs when not running in a virtualenv.16:34
*** tildedave has quit IRC16:35
BK_manit does not even try to create CA...16:36
*** jmckenty has joined #openstack16:37
jaypipessirp_: I'm also a little confused why tarmac let the merge go through with a failing test :(16:37
kbringardoh, from looking at the API logs it looks like specifying -g or --group multiple times does work, it just only displays default in describe-instances16:39
kbringardn/m :-)16:39
devcamcarsoren: can you clarify why you marked the zero volume patch as needs fixing?16:42
devcamcari didn't follow16:42
*** hagarth has joined #openstack16:43
*** tildedave has joined #openstack16:45
vishydevcamcar: added a note to explain16:47
BK_manvishy: which version of httplib2 are you using?16:48
vishyii  python-httplib2                 0.6.0-3build116:49
*** joearnold has joined #openstack16:50
jaypipessirp_: would you mind branching and letting me know if all tests pass for you? they do for me...16:50
devcamcarvishy: still don't follow - where is this hack for 0 -> 100M ?16:50
devcamcarand what does it have to do with this patch16:50
*** kashyap has joined #openstack16:50
vishydevcamcar: it was removed16:50
devcamcarvishy: i'm going to approve it then16:51
BK_manvishy: same on my side. I finished fixing my scripts - now I cleaning CA folder each time. BadStatusLine still here16:51
vishywhich version of eucatools?16:51
BK_man1.3.1 with all Ubuntu patches16:51
* BK_man backported them16:52
vishyand you've had those working?16:53
vishyI've only ever used 1.216:53
BK_manyep. yesterday all was working16:53
*** jakedahn has quit IRC16:54
vishyBK_man: are you using --use_project_ca=True?16:54
BK_manvishy: nope16:55
vishycan you put some logging in euca2ools to figure out which keyfile is giving the BadStatusLine?16:56
vishythe only change to ca stuff recently was from a few days ago16:56
*** bcwaldon has joined #openstack16:56
*** maplebed has quit IRC16:57
*** maplebed has joined #openstack16:57
BK_manvishy: do you think that problem is in keyfile?16:58
*** j05h has quit IRC16:58
*** j05h has joined #openstack16:59
vishyBK_man: wait this is during upload_bundle?16:59
*** smeier00 has left #openstack17:00
BK_manvishy: no. that's from euca-run-instances17:00
vishygah: jaypipes earlier referenced upload-bundle17:00
BK_manvishy: fresh pastie:
vishyhaha ok i've been looking in completely the wrong place17:01
jaypipesvishy: BK_man mentioned upload-bundle, not me :)17:02
*** smeier00 has joined #openstack17:02
vishyjaypipes: yeah I was thinking the problem was with image registration17:02
vishylooks like a problem getting data from glance17:02
*** jmckenty has quit IRC17:02
vishyI do remember seeing something like this before17:02
jaypipes"<BK_man> question: is euca-upload-bundle working with Glance? I'm getting error with starting instances both on Glance and local"17:03
BK_manvishy: I can switch to localimage and there will be another error17:03
*** uksysadmin has quit IRC17:03
vishyoo interesting17:03
BK_manjaypipes: yep, but I tried it with LocalImage service also17:03
vishywhat is the error on local?17:03
BK_manlet me give you an error on local...17:03
patri0tBK_man: I had the exact same error this morning, but I found out republishing images will solve my problem easier17:04
BK_manwhat is correct value for --image_service for local instead of Glance?17:06
BK_manpatri0t: I'm doing everything from scratch - db, image publishing, CA, user, project, network17:07
BK_manpatri0t: each time17:07
vishyoh yeah17:07
vishyold images will not work17:07
*** nacx has quit IRC17:07
BK_manvishy: old images - what does it mean? Do I need to fix my kernel/initrd/rootfs or just republish image using euca-upload-bundle?17:08
BK_manvishy: new pastie for LocalImageService:
*** smeier00 has left #openstack17:10
BK_manvishy: forgot to mention - my rootfs is qcow2 and I have --use_cow_images=true set in /etc/nova/nova.conf17:12
*** dirkx_ has quit IRC17:19
*** deshantm_laptop has joined #openstack17:20
openstackjenkinsProject nova build #815: SUCCESS in 2 min 28 sec:
openstackjenkinsTarmac: Zero out volumes during deletion to prevent data leaking between users17:24
*** smeier00 has joined #openstack17:26
*** stewart has quit IRC17:28
ironcamel_cerberus_: you around?17:29
*** deshantm_laptop_ has joined #openstack17:31
vishyBK_man: looks like it is trying to boot with the wrong image17:32
*** MarkAtwood has quit IRC17:32
nhm_so how's the release going?17:32
Nick0lasilly question.  Is it possible to make a project have a single IP?  /usr/bin/nova-manage network create 1 1   Will that work, or does Nova require a gateway/broadcast?17:32
BK_manvishy: nothing changed on my side. at least by me. let me check it again17:33
*** gregp76 has quit IRC17:33
*** littleidea has left #openstack17:33
alekibangoNick0la: no question is silly...   for example i also do not know :)17:33
Nick0lain the system I am setting up, I am expecting each project to be confined to a single host, so I was planning on giving each project a single IP address from the pool i was given by the networking dept.17:34
*** deshantm_laptop has quit IRC17:34
*** deshantm_laptop_ is now known as deshantm_laptop17:34
Nick0lasorry...confined to a single guest might be a better way of putting it?17:34
Nick0laalekibango:  haha17:34
vishyneed to find out what image id it is passing17:34
vishyBK_man: is the compute host the same as the api host?17:35
BK_manvishy: corrent17:35
BK_manvishy: correct. same host, one host - cc17:35
uvirtbotNew bug: #760037 in nova "os api 1.1 metadata incorrect serialization in json " [Undecided,Invalid]
vishyBK_man: can you check /var/lib/nova/images and make sure that there is an image file?17:37
vishyI wonder if your having a failure decrypting17:37
BK_manvishy: I can see three dirs in /var/lib/nova/images17:37
vishythey should all have an image and an info.json file17:37
BK_manvishy: there are an image file inside each dir + json17:37
*** jfluhmann has joined #openstack17:37
BK_manvishy: # file */image17:38
BK_man036758de/image: ELF 64-bit LSB shared object, x86-64, version 1, statically linked, stripped17:38
BK_man3669705a/image: Qemu Image, Format: Qcow , Version: 217:38
BK_man61dcdb93/image: gzip compressed data, from Unix, last modified: Fri Feb  4 14:20:35 2011, max compression17:38
BK_manvishy: seems OK from my point of view17:38
vishyBK_man: did you reset the image service for the compute host as well?17:38
BK_manvishy: what you mean under 'reset'?17:38
BK_manvishy: I do stop nova services, drop MySQL db, clean all dirs (including images), logs. and starting from scratch17:39
vishyso you need to find out why it can't download the image17:39
vishyone of the opens in line 110-114 in is raising an exception17:40
* BK_man wondering is that merge was problematic?
vishyBk_man: i wouldn't think that would cause an issue17:40
vishyBK_man: that is just for metadata17:40
BK_manvishy: correct. But this is last common string in trace from both Glance and LocalImage17:41
BK_manvishy: metadata = image_service.get(elevated, image_id, image_file)17:41
vishyyeah, so i'm guessing image_id or image_file is wrong, ut we need logging17:42
devcamcarhow's gamma looking? we got the volume zeroing approved, so i think we are looking good?17:48
*** jmckenty has joined #openstack17:48
*** xavicampa has quit IRC17:49
*** imsplitbit has joined #openstack17:50
*** xavicampa has joined #openstack17:51
*** Ryan_Lane has joined #openstack17:52
BK_manvishy: path to json is ok17:54
*** fabiand__ has joined #openstack17:55
BK_manvishy: path to image is also ok17:57
BK_manvishy: /var/lib/nova/images/5f542abd/image17:57
devcamcartermie: you around?17:57
termiedevcamcar: yeah brah, watchooneed17:58
*** drico has joined #openstack18:01
*** MarkAtwood has joined #openstack18:02
*** _vinay has joined #openstack18:03
*** mray has quit IRC18:09
zigo-_-soren: nova-manage should depend on python-amqplib18:10
zigo-_-I'm doing a setup from scratch, and just found out ...18:10
*** zul has quit IRC18:14
*** chuck_ has joined #openstack18:14
*** santhosh_ has joined #openstack18:16
*** santhosh has quit IRC18:16
*** santhosh_ is now known as santhosh18:16
sorenzigo-_-: nova-manage? Why?18:16
*** mihgen has joined #openstack18:18
zigo-_-soren: It did a stack-dump on me when I did apt-get install from my repository.18:20
zigo-_-Not finding this python module ...18:20
*** elambert has joined #openstack18:21
BK_manvishy: NotFound: IO error: [Errno 2] No such file or directory: '/var/lib/nova/images/3670df9c/image'18:23
BK_manvishy: -rw-r--r-- 1 nova nobody 160383488 Apr 13 22:15 /var/lib/nova/images/3670df9c/image18:23
BK_manvishy: don't understand...18:23
zigo-_-soren: Not "nova-manage" but "nova-common" sorry.18:23
zigo-_-I guess it is the 'nova-manage db sync' in the postinst that failed...18:24
uvirtbotNew bug: #760072 in openstack-dashboard "Image detail view needs an x button to resume normal view" [Undecided,New]
sorenzigo-_-: Well, nova-manage does depend on python-nova, which depends on python-carrot which is what we use to transfer messages.18:27
sorenzigo-_-: Typically, this is doen over amqp, and python-carrot does recommend python-amqplib.18:28
sorenI wonder if Nova would work with another of carrot's transport drivers.18:29
zigo-_-recommends != depends18:30
nhm_I enjoy rabbit and carrot as project names btw.18:30
zigo-_-I did willingly, to test dependencies, disabled pulling of recommends ...18:30
sorenzigo-_-: ?!?18:31
*** smeier00 has quit IRC18:31
zigo-_-So you really DO need to explicitely put it in the depends.18:31
sorenzigo-_-: I know depends aren't the same as recommonds.18:31
*** santhosh has quit IRC18:31
sorenzigo-_-: Are you even reading what I'm saying?18:31
zigo-_-soren: So why do you write "python-carrot does recommend python-amqplib" ?18:31
*** santhosh has joined #openstack18:31
sorenBecuase it's true?18:31
alekibangozigo-_-: potentially you do not need amqp18:32
alekibangomaybe :)18:32
zigo-_-I don't care if python-carrot "Recommends:" python-amqplib, we need a stronger dependency here, not a depends !18:32
zigo-_-Not potentially, the postinst of nova-common FAILS!!!18:32
BK_manshutil.copyfileobj(image_file, data) failed in image/ with Not found error. However, there is a file!18:32
alekibangoyou meant prolly  "not a recommend"18:32
sorenzigo-_-: Arguably, a bunch of stuff fails by default because we don't pull in rabbitmq either.18:33
zigo-_-soren: Please don't digress... It would fail puiparts test ...18:34
zigo-_-That's it.18:34
sorenThat's the "uninstalls properly" test, right?18:34
zigo-_-Either remove "nova-manage db sync" call from postinst, either add the dependency.18:35
zigo-_-Nop, that's the "install properly without pulling recommends by default" test ... :)18:35
sorenI don't understand at all why python-amqp is needed for nova-manage db sync, though. That seems odd to me.18:35
alekibangosoren: :)18:35
zigo-_-Maybe simply because there's an import that shouldn't be there?18:35
alekibangothere is so many  odd things with nova...18:35
zigo-_-soren: Maybe you are simply right, and the issue is in the code rather than in the depends...18:36
zigo-_-But right now, it's impossible to install with stricts dependencies without recommends.18:36
mihgenBK_man: may be the process doesn't have access to the directory where file is located? I would try strace anyway..18:37
zigo-_-soren: Also, what is the reason the nova doc asks the admin to create the nova group by hand?18:37
zigo-_-Shouldn't we just put that in the postinst or something?18:38
sorenA nova group?18:38
sorenWhere does it say that?18:38
* soren gave up keeping up with the docs a long time ago.18:38
BK_manmihgen: it fails on 3rd operation, not first! so, it's definitely have access rights to that dir and file:
sorenzigo-_-: No idea.18:39
zigo-_-soren: So, I should NOT use a "nova" group?18:39
sorenzigo-_-: I can't imagine what for.18:39
*** elambert has quit IRC18:39
zigo-_-I can try without, and see how it goes htne, thanks.18:39
*** elambert has joined #openstack18:40
alekibangoannegentle: ^^18:40
sorenLook, these packages work. if there's something they don't do, it can't possibly be needed.18:40
zigo-_-The doc also writes "chown -R root:nova /etc/nova"18:40
sorendon't believe everything you read.18:40
zigo-_-I suppose that I should do chown -R nova /etc/nova ?18:40
*** xavicampa has quit IRC18:40
zigo-_-Eheh! :)18:41
zigo-_-I don't believe anything now!:P18:41
sorenEspecially anything I say.18:41
alekibangosource is the source, sadly18:41
zigo-_-soren: That's exactly WHY I ask you, you know ...18:41
zigo-_-By default, the /etc/nova/nova.conf is left world readable, and it contains passwords...18:42
zigo-_-I'd be for chown nova /etc/nova/nova.conf then chmod to some more restrictive rights.18:43
zigo-_-soren: Don't you think?18:43
sorenYeah, that makes sense.18:43
alekibangosounds good18:43
zigo-_-Maybe we'd better mess-up with unix right AFTER cactus ... :P18:44
*** imsplitbit has quit IRC18:46
zigo-_-Same type of dependency issue with nova-compute, I guess ...18:49
sorenNo, that's a different problem, I think.18:49
sorenIs this with trunk=18:49
*** thor has joined #openstack18:50
zigo-_-This is my own Debian package, bzr925.18:50
zigo-_-It did work before with more python modules.18:50
BK_mansoren: could you please look on ?18:50
zigo-_-soren: Could it be that it needs python-software-properties?18:51
BK_mansoren: I don't really understand what's wrong. strace show same thing - open("/var/lib/nova/images/717f3c22/image", O_RDONLY) = -1 ENOENT (No such file or directory)18:51
*** MarkAtwood has quit IRC18:51
*** santhosh has quit IRC18:52
zigo-_-No, that's not the one...18:52
zigo-_-alekibango: Can you please do a grep in /usr/lib/python2.6 and search for "MessagePlugin" ?18:54
*** santhosh has joined #openstack18:55
*** dysinger` has joined #openstack18:55
sorenBK_man: No clue.18:55
jaypipessirp_: heya, around? wondering about your opinion on bug 759018...18:56
uvirtbotLaunchpad bug 759018 in glance "When running tests without virtualenv, test_traceback_not_consumed fails" [Critical,In progress]
*** imsplitbit has joined #openstack18:56
sirp_jaypipes: yep looking now18:56
jaypipessirp_: thx man.18:56
alekibangozigo-_-: will18:56
*** stewart has joined #openstack18:56
sirp_jaypipes: want me to review the branch you have for it? or have a question?18:57
sorenzigo-_-: That problem was fixed in rev 934.18:57
zigo-_-alekibango: Skip that, it wont be relevant, IMHO.18:57
zigo-_-Oh, thanks.18:57
jaypipessirp_: review would be good. :)18:57
sirp_jaypipes: cool, ill give it a test-drive right now18:58
jaypipessirp_: goes back to that dang eventlet/GreenThread exception thing... the patch fixes the bug, but I want to get your opinion since you've had strong opinions about that particular little block of code in the past! :)18:58
*** mray has joined #openstack18:58
* zigo-_- wonder how it could have one day work then...18:58
sorenzigo-_-: Maybe you had python-suds installed.18:59
jaypipesttx: request exception for It's a blocker/critical IMHO, as it's preventing packaging from building due to a test case failure.18:59
uvirtbotLaunchpad bug 759018 in glance "When running tests without virtualenv, test_traceback_not_consumed fails" [Critical,In progress]18:59
*** stewart has quit IRC19:00
zigo-_-soren: Why doesn't it depends on it then?19:00
*** mramige has joined #openstack19:00
zigo-_-Not needed as well?19:00
sorenzigo-_-: 18:57 <+soren> zigo-_-: That problem was fixed in rev 934.19:00
jaypipessirp_: I *really* am confused why the test passes with virtualenv and doesn't without one, though. :(19:00
zigo-_-soren: So, we never need python-suds anymore or what?19:01
zigo-_-Just asking to know if I need to check its packaging in Debian...19:01
sorenOnly if you want to use VMWare.19:01
zigo-_-Oh ok, thanks.19:01
*** dirkx_ has joined #openstack19:01
* zigo-_- don't care non-free software...19:01
*** dirkx_ has quit IRC19:01
sorenzigo-_-: It would probably be a good idea to forget all about r925. We're in bug fixing mode, and we're at 985. That's ~60 bugs fixed.19:02
sorenWhy waste time experiencing 60 known and fixed bugs?19:02
zigo-_-soren: I'm just trying to reinstall from scratch, and pulled from what I've put in my repo a week ago.19:03
zigo-_-I'll of course upgrade right now.19:03
* soren has to run an errand.19:03
sirp_jaypipes: looks like we're only killing the image in very specific cases; shouldn't we be killing it for any exception that isn't explicitly caught?19:04
sirp_jaypipes: for example mysql gone away errors and such should probably cause the image to be killed19:04
*** mgoldmann has quit IRC19:05
jaypipessirp_: not sure. if you can find a way to fix the bug using the existing code, I'm all for it. I'm *really* confused why running the code in a virtualenv makes a difference. :(19:05
sirp_jaypipes: yeah that is *really* strange19:05
*** Ryan_Lane is now known as Ryan_Lane|food19:06
*** MarkAtwood has joined #openstack19:06
jaypipessirp_: but, as it is, we don't have much time :( I propose either a) sticking with this, since it does fix the test failure or b) adding a except Exception: block at the end of just the _upload function there and calling _safe_kill() (though to be honest, I'd almost rather *leave it out* to give us a heads up when non-Duplicate exceptions occur...19:07
*** lvaughn has quit IRC19:07
*** lvaughn has joined #openstack19:08
jaypipessirp_: I think maybe that except: block is hiding something somewhere (as they have a tendency to do...) unfortunately, I haven't been able to figure out what's being hidden! :(19:08
jaypipessirp_: ^^ refers to pre-patched code...19:08
sirp_jaypipes: im all for re-raising exceptions, so that we get a heads up, but i think we should still kill the image to keep the data consistent19:08
jaypipessirp_: k, so you're suggesting adding an except: block at the end of _upload() that calls safe_kill() then?19:09
jaypipessirp_: the b) option above?19:09
sirp_jaypipes: yeah, as long as it doesn't somehow break the test19:09
sirp_jaypipes: the original code seemed right to me in spirit, even though some evenlet wonkiness seems to be messing that up19:10
jaypipessirp_: hey, I hear you... I'm still puzzled over this one.19:10
*** javidiaz has joined #openstack19:10
sirp_jaypipes: ill go ahead and fiddle with this for a bit and see if anything jumps out but, since we're pressed for time, ill defer to you for the best stop-gap approach19:11
jaypipessirp_: k19:13
*** smeier00 has joined #openstack19:15
vishyBK_Man: permissions issue?   Is the stuff owned by root instead of nova?19:17
*** mramige has quit IRC19:19
*** dendrobates is now known as dendro-afk19:22
jaypipessirp_: k, I added an except: catch-all at the bottom of _upload() and pushed it to that branch.19:23
*** imsplitbit has quit IRC19:23
sirp_jaypipes: k repulling19:23
*** neogenix has joined #openstack19:23
devcamcarttx: FFE request for ?19:25
*** lvaughn_ has joined #openstack19:26
*** koolhead17 has quit IRC19:26
*** zigo-_- has quit IRC19:27
notmynamettx: my thought on the swift release versioning is that we should use 1.3-rc for the release candidate and 1.3.0 for the release. this prevents a possible 2 versions of the code/packages at 1.3.0. If there are no differences from rc to final, bumping the version and making the tarball should be straightforward19:27
*** neogenix_ has joined #openstack19:28
notmynamettx: ultimately, I don't think this whole process is the best, and we should all talk about it at the summit, but for now, I think it should be as close as possible to the previous swift releases (eg 1.2-rc and 1.2.0)19:28
*** neogenix has quit IRC19:29
*** neogenix_ is now known as neogenix19:29
*** lvaughn has quit IRC19:29
*** joearnold has quit IRC19:31
*** dendro-afk is now known as dendrobates19:31
*** joearnold has joined #openstack19:31
*** fabiand__ has quit IRC19:36
*** rds__ has quit IRC19:36
*** HouseAway is now known as AimanA19:38
*** MarkAtwood has quit IRC19:39
*** koolhead17 has joined #openstack19:41
*** Ryan_Lane|food is now known as Ryan_Lane19:43
markwashjaypipes: quick question about your preferred use of the instance_metadata table19:44
*** benbenhappy has left #openstack19:44
markwashjaypipes: would we want to make sure there was some mechanism to distinguish between user-provided metadata and system-provided metadata?19:45
*** adiantum has quit IRC19:45
*** fabiand__ has joined #openstack19:45
kbringardmarkwash and jaypipes: that reminds me... should "type" be a core property in the images table... as it stands it's required for the ec2 api to work19:46
jmckentyvishy: are we thinking two separate NaaS sessions - one for the Cisco proposal, and one for dendrobates rollup? Or just one?19:51
vishyjmckenty: I'm analyzing NaaS now, there are 3+ groups19:51
jmckentyright, was going to suggest that happen AFTER that session19:52
vishyjmckenty: ok19:52
jmckentyand I just wanted to make sure I was subscribed to the right one :)19:52
*** dragondm has quit IRC19:55
*** nelson has quit IRC20:03
*** nelson has joined #openstack20:03
*** jmckenty has quit IRC20:04
*** chuck_ is now known as zul20:07
*** dragondm has joined #openstack20:13
*** deshantm_laptop has quit IRC20:13
*** antenagora has joined #openstack20:16
*** clauden_ has joined #openstack20:16
Ryan_LaneI'm having an issue launching instances since upgrading to trunk from bexar release20:20
Ryan_Laneinstances launch, but don't seem to do anything. nothing shows up in the console log, for instance20:21
Ryan_Laneand I can't ping them20:21
Ryan_LaneI switched to glance, as this seems to be required now20:21
Ryan_Lanecould it be the way I'm adding the images into glance?20:21
kbringardRyan_Lane: can you euca-describe-instances?20:21
Ryan_Laneit shows it running20:22
kbringardsorry, I meant describe-images20:22
Ryan_LaneIMAGE   ami-00000001    None (Ubuntu 10.04 Server LTS (2011-03-22))             available       public                  machine20:22
kbringardand, are the ramdisk and kernel properly referenced (if they're not included in the base image)20:22
Ryan_Lanethat's the thing. no20:22
kbringardah, ok20:22
kbringardone sec20:22
Ryan_LaneI can't seem to use uec-public-tarball either20:22
kbringardyea, that uses the ec2 stuff20:23
Ryan_Laneglance marks them as queued, and I have no way to make them active20:23
kbringardin my experience, you need to use the glance-updloader tool20:23
Ryan_LaneI did use glance-upload for this20:23
kbringardoh, hrmm20:23
Ryan_Lanewait… did i?20:23
Ryan_LaneI remember trying it and still having this problem20:23
Ryan_Lanelet me try using it again20:23
kbringardthat's how my properties look for an image and corresponding kernel20:24
Ryan_Lanehmm. it seems glance is all screwed up now too20:24
kbringardthe kernel id and type are the important bits20:25
Ryan_Laneah. i see20:25
Ryan_Lanehow did you upload this?20:25
kbringardI'll show you20:25
kbringardI wrote this little wrapper20:26
Ryan_Laneoh. sweet. that's really helpful20:26
kbringardit works most of the time20:26
kbringardI'm going to commit it to my github once I tst it throughly enough to feel safe20:26
kbringardit *should* work20:27
Ryan_Lanefor a normal uec image, what's the normal usage?20:27
kbringardput the kernel, ramdisk and image all in the same base dir20:27
kbringardthen do20:27
kbringard./glance-uploader.bash -i ttylinux-uec-amd64-12.1_2.6.35-22_1.img -k ttylinux-uec-amd64-12.1_2.6.35-22_1-vmlinuz -r ttylinux-uec-amd64-12.1_2.6.35-22_1-initrd -v 1.0 -d ttylinux -e 2.6.35 -a amd6420:27
kbringardthe, -d, -d -e, and -a are whatever you want20:28
kbringardbut they're required by the script20:28
kbringardmostly because it's info we want every image to have in our environment20:28
kbringardit also generates some other info, like, it pulls the username and hostname of the machine that did the uploading20:29
kbringardand puts that in there20:29
kbringardif you just run it with no options it'll output a usage menu20:29
Ryan_Lanegreat. let me see if this will fix my instances-not-booting issue :)20:30
kbringardhehe, good luck :-)20:30
kbringardit's kind of a stop gap, I'm working on adding post support to ogle ( so you can just do it straight from ruby20:30
*** joearnold has quit IRC20:31
*** joearnold has joined #openstack20:31
Ryan_Laneheh. ruby :)20:31
Ryan_Laneno python support? :)20:31
kbringardeh, Ruby is what we use where I work20:31
Ryan_LaneI should add php support to my web interface20:31
kbringardso it's what I write out utils in :-)20:32
Ryan_Laneglance support to my php web interface20:32
*** reldan has quit IRC20:32
Ryan_Laneyeah, makes sense20:32
Ryan_LaneI'm using php for a similar reason20:32
*** antenagora has quit IRC20:33
*** reldan has joined #openstack20:34
*** neogenix has quit IRC20:36
*** fabiand__ has quit IRC20:36
kbringardany luck?20:37
Ryan_Laneyou solved my problem20:37
kbringardglad I could help20:37
Ryan_Lanethough I need to rework how I display images to users20:37
Ryan_Lanesince users can select either the kernel or the machine image, which is strange20:37
*** fabiand__ has joined #openstack20:38
kbringardyou should be able to filter by20:38
*** mihgen has quit IRC20:38
kbringardaki is kernel20:38
kbringardari is ramdisk20:38
kbringardand ami is machine20:38
kbringard{"name"=>"maverick-server-uec-amd64-vmlinuz-virtual", "container_format"=>"aki", "disk_format"=>"aki", "checksum"=>"73152adae523a52d481f5150deb30356", "id"=>1, "size"=>4408912}, {"name"=>"maverick-server-uec-amd64.img", "container_format"=>"ami", "disk_format"=>"ami", "checksum"=>"d2d3b3049f897dff2f86295d5a9e879e", "id"=>2, "size"=>1476395008}20:39
kbringardthat's what the glance api returns (for a non detailed image list)20:39
*** fabiand__ has left #openstack20:39
Ryan_Laneyeah. I need to check how nova is returning them via the api :)20:39
kbringardif you read ruby, it's all extracted out in ogle, perhaps you can port it to php and be all pimpin'20:39
Ryan_Lanewell, I'm trying to use the ec2-api solely right now20:40
kbringardah, got it20:40
Ryan_Lanesince it's the only thing that actually supports everything20:40
Ryan_Laneand of course it reports both of them as "machine" type20:40
Ryan_Lanewhich is wrong20:40
Ryan_Laneand should be fixed in nova20:40
*** mihgen has joined #openstack20:45
*** allsystemsarego has quit IRC20:45
*** iammartian has joined #openstack20:50
kbringardyea, that's frustrating20:50
*** iammartian has left #openstack20:50
kbringardwe've written our portal to use the glance API for images20:51
kbringardsince I don't imagine the ec2 api is ever going to support glance meta data20:51
kbringardanyway, I started a new repo for tools I've written here:
kbringardit's got the uploader in there20:52
Ryan_Laneyeah, I'd imagine the ec2 api won't support it either20:52
Ryan_LaneI plan on moving to the openstack api, when it has support for everything20:52
kbringardyea, likewise20:53
*** reldan has quit IRC20:53
jaypipesmarkwash: sorry, stepped away for a bit...20:55
jaypipesmarkwash, kbringard: we used to have "type" as a base image attribute. We split that out into disk_format and container_format in Cactus, to make Glance more flexible for storing non-Amazon-specific stuff.20:56
*** reldan has joined #openstack20:56
kbringardjaypipes: ah, makes sense20:56
jaypipesmarkwash: as for the instance_metadata table, I'm not sure I have a good answer for you on that. I think it's worth discussing, though.20:57
_vinayI am looking at smoketests20:57
_vinayin (class SmokeTestCase)20:58
_vinayin def connection_for_env() we do20:58
*** arun_ has quit IRC20:58
_vinayreturn boto.connect_ec2()20:58
_vinayMy question is this has nothing to do with ec2 right?20:59
_vinayI mean it doesnt connect to EC2 .. am I right?20:59
kbringard_vinay: you mean like, amazon's actual webservice?20:59
kpepple_vinay: no, you should read that as "connect using EC2 API"21:00
kbringardkpepple: beat me to it :-)21:00
_vinayyep thats what I am confused about21:00
_vinayalthough I do see the connection to the api server I started21:00
_vinaynova server I mean?21:00
_vinaynova server I mean21:00
_vinayok cool :-)21:00
_vinayso its named _ec2 because it uses ec2 apis .. cool21:01
uvirtbotNew bug: #760239 in openstack-dashboard "Permission denied page is not styled" [Undecided,New]
*** h0cin has quit IRC21:05
*** bkkrw has joined #openstack21:06
*** ChamSys has joined #openstack21:06
jaypipessirp_: fix pushed to that _upload branch.21:10
sirp_jaypipes: cool21:10
*** Shentonfreude has quit IRC21:11
kpepplejaypipes: have a second for a quick glance question ?21:14
*** ctennis has quit IRC21:15
*** patcoll has quit IRC21:16
ttxdevcamcar: looking21:17
*** mihgen has quit IRC21:18
*** mihgen has joined #openstack21:18
ttxdevcamcar: ok, let's pull it21:19
jaypipeskpepple: shoot.21:19
ttxjaypipes: looking21:19
kpepplejaypipes: does nova-compute need access to glances' image store directly (local or swift) or does it go through glance api to get the actual image ?21:20
ttxjaypipes: I'm a bit unconfortable with it, but not enough to veto it.21:22
ttxjaypipes: if you are confortable with it, I'll take it.21:22
jaypipeskpepple: it uses the glance client to retrieve images. the XenServer plugin, in addition, uses the glance-upload tool to do snapshotting.21:22
jaypipesttx: pls check with sirp_ too.21:23
*** bcwaldon has quit IRC21:23
*** patcoll has joined #openstack21:23
ttxjaypipes: I'll mark it RC. If you can get it in by the RCFreeze deadline, it's ok.21:24
jaypipesttx: it's in final review from sirp_ now.21:24
kpepplejaypipes: thx21:25
sirp_jaypipes: ttx just marked it as approved, lgtm21:25
* ttx looks into what blocks refresh of
ttxnotmyname: agreed. That means I'll need you around early on Friday.21:26
notmynamettx: np21:26
ttxor any swift-core dude to push the versioning branch.21:26
*** j05h has quit IRC21:27
*** j05h has joined #openstack21:28
notmynamettx: by "early friday", when did you have in mind? like 5 or 6 am? or like 8 or 9 am?21:29
notmynameeither are good for me, just want to know :-)21:29
*** ctennis has joined #openstack21:29
ttxI promised the release to Lauren by 10am. So 8am should be alright.21:29
notmynameok, np21:29
ttxi'll have the branch merge proposal ready for you.21:29
notmynameok, I'll reject the one you sent earlier21:30
*** gaveen_ is now known as gaveen21:30
ttxnotmyname: ideally we'd have a jenkins-powered job that I can trigger that would bypass swift-core review and limit the risk of fumbling21:30
ttxlike the one I use for tagging21:30
*** clauden_ has quit IRC21:30
*** clauden_ has joined #openstack21:31
*** ppetraki has quit IRC21:31
* ttx makes mental note21:31
ttxarh, someone put unicode in their branch name ?21:32
*** dweimer has joined #openstack21:32
ttxjaypipes, devcamcar, soren; any of you can accept jk0's ? I just granted rc exception21:35
ttxarh, Jason Kölker21:36
ChamSysanyone use zend framework with the openstack api, does it work fairly well using the simple cloud api yet?21:36
ttxsomeone put non-ascii in their name :)21:36
ttxbreaking my crappy reviewslist code :)21:36
Ryan_LaneI'm tracking down a bug in either the glance client, or in the ec2 api, and am having trouble finding the root cause21:40
Ryan_Laneall images are being returned as 'machine' type, even if they are not21:40
Ryan_Lanethe header being returned by glance, for a kernel image is: X-Image-Meta-Container_format: aki21:40
*** bcwaldon has joined #openstack21:41
Ryan_Laneglance utils. get_image_meta_from_headers() should be returning result['container_format'] = aki21:41
Ryan_Lanebut it seems it is returning None21:41
*** dspano has quit IRC21:42
ttxjaypipes: so you have 1-2 branches left to land in Glance, and nova-core has one more branch to land in Nova. Feeling confident ?21:45
*** smeier00 has quit IRC21:45
jaypipesttx: on one of the branches, not sure about the doc one...21:45
Ryan_Lanehmm, in fact, it seems it isn't even being returned by the glance client21:45
ttxjaypipes: i don't really care about that one. If it's ready it can go in, if not, too bad.21:46
jaypipesttx: k21:46
jaypipesRyan_Lane: can you reproduce this with just glance, no nova?21:46
ttxjaypipes: going to bed soon, any question for me ?21:47
jaypipesttx: nope21:47
Ryan_Laneglance shows the images properly21:47
ttxjaypipes: cool.21:47
Ryan_LaneContainer format: aki <- from glance show21:47
*** dendrobates is now known as dendro-afk21:47
*** dendro-afk is now known as dendrobates21:47
ttxjaypipes: if any issue is raised while I sleep, I'll defer to your guidance.21:48
jaypipesttx: k21:48
* ttx disappears21:48
jaypipesRyan_Lane: k, so must be in the ImageService adapter layer in Nova.21:49
Ryan_Lanethinking so21:49
Ryan_Lanetrying to track it down :)21:49
sirp_Ryan_Lane: yeah, just tried glance.client, seemed to work21:49
Ryan_Laneah. I think I found the problem21:49
Ryan_Lanecontainer_format is returned in image['properties'], not in image21:50
tr3buchetcame across an oddity in network/ most times db is referred to as self.db.function() but there are occurrences of db.fucntion() as well. shouldn't it one way or the other?21:50
Ryan_Lanethat's it21:50
Ryan_LaneI don't know how that will affect non-glance services though21:51
Ryan_Lanelet me open a bug21:51
*** smeier00 has joined #openstack21:53
*** rds__ has joined #openstack21:55
*** pothos_ has joined #openstack21:56
vishytr3buchet: should be using self.db everywhere21:57
*** keekz_ has joined #openstack21:57
*** openstackjenkins has quit IRC21:58
*** laurensell has quit IRC21:58
*** pothos has quit IRC21:58
*** keekz has quit IRC21:58
*** Ryan_Lane1 has joined #openstack21:58
*** Ryan_Lane1 has joined #openstack21:58
*** pothos_ is now known as pothos21:58
*** Ryan_Lane has quit IRC21:58
*** Ryan_Lane1 is now known as Ryan_Lane21:58
*** laurensell has joined #openstack21:59
tr3buchetalright i'll fix it in multi-nic21:59
*** clauden_ has quit IRC21:59
*** clauden__ has joined #openstack21:59
tr3buchetunless it's big enough to have it's own bug report22:00
*** openstackjenkins has joined #openstack22:00
uvirtbotNew bug: #760289 in nova "All image types return as 'machine' when using glance client in ec2 api" [Undecided,New]
*** patcoll has quit IRC22:02
*** bcwaldon has quit IRC22:02
*** smeier00 has quit IRC22:03
jaypipesRyan_Lane: from that bug "The glance client returns the container_format in image['properties']". It does?22:05
Ryan_Lanejaypipes: correct22:07
Ryan_LaneI traced it down to that22:07
jaypipesRyan_Lane: the Glance client or the GlanceImageService?22:07
Ryan_Laneso, should I fix this by changing glance client, or how the ec2 api checks it?22:07
Ryan_Laneglance image service22:07
Ryan_Lanelet me be sure22:07
*** mihgen has quit IRC22:08
*** mihgen has joined #openstack22:09
Ryan_Lanejaypipes: glance image service22:09
openstackjenkinsProject nova build #816: SUCCESS in 2 min 30 sec:
openstackjenkinsTarmac: Try to be nicer to the DB when destroying a libvirt instance.22:09
jaypipesRyan_Lane: ok. (that's not quite the same as the glance client.) There is a glance.client.Client class, which is why I was confused.22:09
Ryan_Lanein /usr/share/pyshared/glance/ def get_image_meta_from_headers(response)22:09
Ryan_Lanesorry :)22:10
Ryan_LaneI was going to fix this by changing the ec2 api22:10
Ryan_Laneto check in image, and in image['properties']22:10
Ryan_Lanethe local image service returns this in image, I believe22:11
jaypipesRyan_Lane: I'm looking at nova/image/, line 42-43. container_format is shown as a base image attribute, not a part of the properties collectio.22:11
_vinayin the smoketests I am trying to use xenapi_conn (instead of ec2 conn)22:12
*** smeier00 has joined #openstack22:12
Ryan_LaneI printed the keys of image, and image['properties'] in the ec2 api22:12
_vinaywhat should the url for xenapi_conn look like?22:12
Ryan_Lanecontainer_format shows up in image['properties']22:12
_vinayI am guessing its somewhere in some config of my nova installation22:13
Ryan_Lanejamiec: so I'd imagine that it's incorrect there as well22:13
Ryan_Lanewrong page :)22:13
Ryan_Lanejaypipes: so I'd imagine that it's incorrect there as well22:14
jaypipesRyan_Lane: if container_format is showing up in properties, that means someone is manually adding a container_format property I believe...22:15
Ryan_Lanehmm. lemme see22:15
*** adiantum has joined #openstack22:15
jaypipesRyan_Lane: doh.22:16
Ryan_Lanejaypipes: looking at glance show container format isn't showing up in the "Property" list22:18
jaypipesRyan_Lane: I think I see the issue. This is exactly why I don't like using @classmethods...22:18
jaypipesRyan_Lane: assign the bug to me.22:20
Ryan_Lanejaypipes: I don't think I have permissions to do so22:20
jaypipesRyan_Lane: ah, ok, sorry, I'll do it.22:20
*** mihgen has quit IRC22:21
Ryan_LaneI should have used "driver" instead of "client" in that bug report :)22:21
*** mihgen_ has joined #openstack22:22
*** mihgen has joined #openstack22:23
vishylooks like it should happen with local image service and not with glance22:23
*** jonkelly has quit IRC22:24
*** dysinger` has quit IRC22:25
*** rsaidan has quit IRC22:26
vishysomething is definitely wrong with the glance service translate to base stuff22:27
*** JuanPerez has quit IRC22:28
* tr3buchet pulls out hair22:29
tr3buchet(nova): TRACE: RemoteError: AttributeError 'dict' object has no attribute '__dict__'22:29
alekibangotr3buchet: so its not a dictionary22:29
alekibangomaybe wrong config option, or a bug :)22:29
vishyno it is not an object22:29
tr3buchetthat's probably my favorite error so far22:29
vishy__dict__ belongs to instance22:30
alekibangobetter was ERROR: operation ended successfully22:30
tr3buchetyeah that's pretty good i suppose22:30
*** joearnold has quit IRC22:31
*** joearnold has joined #openstack22:31
*** clauden has joined #openstack22:33
jaypipesRyan_Lane: try merging lp:~jaypipes/nova/bug760289 and lemme know if that fixes it for ya.22:33
jaypipesvishy: see my revision in lp:~jaypipes/nova/bug760289.22:34
*** joearnold has quit IRC22:34
jaypipesvishy: this is partly why I despise using @classmethods with inheritance.22:34
*** joearnold has joined #openstack22:35
*** dendrobates is now known as dendro-afk22:35
*** clauden__ has quit IRC22:36
*** stewart has joined #openstack22:36
Ryan_Lanejaypipes: works for me22:36
vishywhere does glance log?22:37
vishyby default22:37
vishyi was using glance combined so it was stdout. with glance-control i need logs22:38
Ryan_Laneusing the packages, it logs to /var/log/glance22:39
Ryan_Laneotherwise it logs however you have it set in glance.conf22:39
*** ivan has quit IRC22:40
jaypipesvishy, Ryan_Lane: k, I will propose for merging... not sure whether this should go into RC, but it seems pretty critical...22:40
*** hadrian has quit IRC22:40
*** ivan has joined #openstack22:40
Ryan_Lanewell, this is pretty problematic for web clients22:40
Ryan_LaneI filter based on machine type22:40
Ryan_Laneif this isn't working I'll have users launching instances with kernel images22:40
vishyjaypipes...i'd say it is critical22:41
jaypipesvishy: I would agree.22:41
*** zul has quit IRC22:42
jaypipesvishy: I'm going to target it to cactus-rc. ok with you>?22:42
Ryan_LaneI'm glad I started testing before release :D22:43
*** clauden__ has joined #openstack22:43
*** clauden has quit IRC22:43
*** clauden__ is now known as clauden22:43
jaypipeskinda concerned that testing didn't pick this up, but from what I can tell, everything is so stubbed out in the unit tests, I'm not surprised.22:44
mirrorboxmihgen: hey there22:45
vishystill having no luck with logging22:47
vishyapparently i'm missing some important setup22:48
kbringardaiight peoples22:48
kbringardit's time for beer22:48
kbringardhave a good night22:48
*** kbringard has quit IRC22:48
*** ChamSys has quit IRC22:50
vishyok i can run the workers separately, but glance-control does nothin22:51
Ryan_Laneis glance going to ship with a daemon in the packages at some point?22:52
jaypipesvishy: what are you trying to do? you can put log_file in [app:glance-api] and [app:glance-registry] sections in glance.conf.22:52
jaypipesRyan_Lane: you mean besides glance-control?22:52
Ryan_Lanes/daemon/init or upstart script/22:52
jaypipesRyan_Lane: unlikely.22:52
jaypipesRyan_Lane: maybe upstart script if soren's not already worked on one.22:53
jaypipesvishy: can I help with logging? are you not seeing any log output at all?22:55
vishyjaypipes: that is a nasty one22:58
*** Gaelfr has joined #openstack22:58
vishyjaypipes: not sure if it is logging related, but workers just immediately crash with glance-control22:58
*** Gaelfr has quit IRC22:58
jaypipesvishy: error messages? any output at all?22:59
vishyjaypipes: no output22:59
vishyjaypipes: this is from a trunk checkout22:59
jaypipesvishy: can you show me exactly what you are doing?22:59
vishyand i can run the workers fine individually22:59
vishy./bin/glance-control all start22:59
jaypipesand what is in your glance.conf?23:00
vishyjaypipes: it tells me that they start23:00
vishybut they aren't running23:00
jaypipesvishy: could you paste your glance.conf?23:00
vishyonly change is image dir23:00
vishyotherwise it is default23:01
jaypipesvishy: so no log_file, then?23:01
*** dysinger` has joined #openstack23:01
vishynot unless it is in the default config23:01
*** elambert has quit IRC23:01
*** Gaelfr has joined #openstack23:01
*** elambert has joined #openstack23:02
vishyjaypipes: curiosity question, is there a way to call a base class method but pass the current class?23:02
jaypipesvishy: not with @classmethods. That's what normal instance methods in an object hierarchy are for.23:03
vishyjaypipes: you can't use super?23:03
*** nelson has quit IRC23:04
jaypipesvishy: hmm, I see what you're getting at... not sure, haven't tried it. seems hacky to me.23:04
jaypipesvishy: as does the existing code...23:04
vishyjaypipes: i think it was sirp23:04
vishybut i think super will work, lemme try it23:04
jaypipeswtf.. launchpad seems to be down. :(23:04
*** stewart has quit IRC23:06
*** mihgen has quit IRC23:08
*** Zangetsue has quit IRC23:11
jaypipesvishy: gah, not sure what's going on with glance-control all start... it's only starting up glance-api for me for some reason. I'll file a bug.23:11
*** santhosh has quit IRC23:11
vishyjaypipes: cool, I just tried super, no dice so i guess your patch is good for now23:11
*** elambert has quit IRC23:13
jaypipesvishy: ah, so glance-control is indeed working... it's just that by default, it does what swift does and tries to write the pidfiles to /var/run/. This obviously won't work if you don't run it as sudo. You can run glance-control api start and it works fine.23:14
jaypipesvishy: ack on the super thing.23:14
jaypipesvishy: I'll still file a bug about the /var/run/ bug. Really, if not run as root, it should probably try to write the pidfiles to the current dir, don't ya think? ;)23:15
medberryis registration closed for the Diablo Design Summit?23:16
jaypipesmedberry: check with spectorclan whenever he pops up on IRC :)23:18
medberryjaypipes, nod, tx.23:19
vishyjaypipes: i was sudoed and pids created in var run23:19
*** Ryan_Lane has quit IRC23:22
*** elambert has joined #openstack23:25
*** elambert_ has joined #openstack23:32
*** elambert has quit IRC23:32
*** elambert_ is now known as elambert23:32
*** dovetaildan has quit IRC23:36
*** JuanPerez has joined #openstack23:36
*** dovetaildan has joined #openstack23:37
*** zedas has quit IRC23:42
uvirtbotNew bug: #760342 in nova "API returning Error after running a number of builds" [Undecided,New]
*** joearnold has quit IRC23:42
*** santhosh has joined #openstack23:45
*** joearnold has joined #openstack23:45
*** smeier00 has quit IRC23:45
*** clauden has quit IRC23:47
*** nelson has joined #openstack23:48
*** santhosh has quit IRC23:51
*** elambert has quit IRC23:52
*** elambert has joined #openstack23:52
*** RobertLaptop has quit IRC23:52
*** lvaughn_ has quit IRC23:53
*** icarus901 has quit IRC23:54
*** bkkrw has quit IRC23:55
*** jeffjapan has joined #openstack23:57

Generated by 2.14.0 by Marius Gedminas - find it at!