Tuesday, 2010-11-23

*** joearnold has quit IRC00:00
*** ptremblett has joined #openstack00:00
*** joearnold has joined #openstack00:01
sorenkirkland: Found the problem. It'll take at least one publisher run to fix it. Tomorrow morning it is.00:05
kirklandsoren: okay, thanks00:05
soren(python-sphinx version skew, as it turns out)00:05
* soren heads bedwards00:05
soreng'night all00:05
kirklanddendro-afk: annegentle: okay, soren triaged the problem with the packages in the PPA;  ISO is awaiting fresh packages in the PPA;  will be tomorrow before I have another candidate ISO for you to see00:06
*** rsampaio has joined #openstack00:11
*** jimbaker has quit IRC00:14
*** dubsquared has quit IRC00:17
*** tunettnt has joined #openstack00:18
exltgithub mirrors are both functional again - history is rewritten, but you can clone and fork again ;)00:20
*** pvo_ has joined #openstack00:20
*** ChanServ sets mode: +v pvo_00:20
tunettntWhat's for infiniband support ?00:23
*** joearnold has quit IRC00:27
notmynameexlt: \o/00:27
*** dizz|away has quit IRC00:30
*** ccustine has quit IRC00:32
*** tunettnt has quit IRC00:35
*** burris has quit IRC00:43
*** johnbergoon has quit IRC00:52
*** johnbergoon has joined #openstack00:53
*** daleolds has quit IRC00:53
*** johnbergoon has quit IRC00:57
*** joearnold has joined #openstack01:02
*** burris has joined #openstack01:15
*** joearnold has quit IRC01:17
*** pvo_ has quit IRC01:20
*** iammartian__ has joined #openstack01:20
*** burris has quit IRC01:21
*** joearnold has joined #openstack01:24
*** iammartian has quit IRC01:25
*** iammartian__ is now known as iammartian01:25
*** joearnol_ has joined #openstack01:25
*** joearnold has quit IRC01:25
*** burris has joined #openstack01:26
*** joearnol_ has quit IRC01:48
*** joearnold has joined #openstack01:48
*** maplebed has quit IRC01:54
*** burris has quit IRC02:07
*** burris has joined #openstack02:12
*** gaveen has joined #openstack02:23
*** HugoKuo has joined #openstack02:37
HugoKuoHello this is hugo02:39
HugoKuoI have a question about branch02:42
notmynamewhat about?02:42
HugoKuowhile I bzr branch lp:XXXX02:43
HugoKuothen  I copy the  /XXXX  into /nova  then run nova.sh branch02:44
HugoKuois this the correct step to branch it ?02:44
notmynameI'm not a bzr expert, but that doesn't sound right to me02:44
notmynamewhen I do work on swift, I do bzr clone lp:XXXX inside of a local directory that I previously set up with bzr init-repo02:45
notmynamethat creates the local folder XXXX02:45
notmynamebut I don't know about nova.sh. perhaps one of the nova devs will be on soon02:46
xtoddxI think with nova.sh there is a way to specify a branch name other than trunk02:46
notmynamextoddx to the rescue02:46
HugoKuothanks dude02:46
xtoddx./nova.sh branch lp:XXXX02:46
openstackHugoKuo: Error: "!!!!!!!!" is not a valid command.02:47
xtoddxhaha, silly openstack02:47
HugoKuo! I remember it02:47
openstackHugoKuo: Error: "I" is not a valid command.02:47
notmynameseriously, how many bots do we have in here?02:47
xtoddxover 900002:47
xtoddxat least02:47
notmynameit would be nice if they would only speak when spoken to :-)02:48
openstackxtoddx: Error: "true" is not a valid command.02:48
openstackxtoddx: Error: "verbose=0" is not a valid command.02:48
openstackxtoddx: (help [<plugin>] [<command>]) -- This command gives a useful description of what <command> does. <plugin> is only necessary if the command is in more than one plugin.02:48
notmynamethat's not help02:48
openstacknotmyname: Admin, Channel, ChannelLogger, Config, MeetBot, Misc, Owner, Services, and User02:49
openstackHugoKuo: Admin, Channel, ChannelLogger, Config, MeetBot, Misc, Owner, Services, and User02:49
xtoddx!help Misc02:49
openstackxtoddx: Error: There is no command "misc".02:49
* notmyname slaps openstack with a trout02:49
HugoKuoactually I'm very new in OPENSTACK02:50
notmynameHugoKuo: well then, welcome02:50
xtoddxhugokuo: most everyone is02:50
xtoddxdon't be intimidated02:50
HugoKuobtw thanks guys ..... there are not many user play it in Taiwan02:50
*** pvo_ has joined #openstack02:51
*** ChanServ sets mode: +v pvo_02:51
notmyname(not that it's the same at all, but) we have some partners in japan02:51
HugoKuoI feel poor in information02:51
notmynameHugoKuo: what sort of info are you looking for? how to contribute? how to install?02:52
HugoKuoI know .....that's quite nice  nice to meet you , and I have to keep on my research :>02:52
HugoKuoboth .... my work is to deploy an IaaS enviroment02:52
HugoKuothere is a Eucalyptus on air02:53
notmynameare you more interested in a particular component like computer or storage or both parts together?02:53
HugoKuoboth parts...02:53
HugoKuoI hope the use opensatck nova & SWIFT as well as EC2 & S302:54
*** ksteward has quit IRC02:55
HugoKuoat first I success to install and run instance  in NOVA02:56
notmynamewe're still (always) working on docs. a starting point for swift could be http://programmerthoughts.com/programming/swift-openstack-object-storage-overview/. from there looks at specific components of it (documented at http://swift.openstack.org)02:56
*** laurensell has joined #openstack02:56
HugoKuowow  take a look02:57
notmynamebest place to start with nova is http://nova.openstack.org02:57
*** dragondm has quit IRC02:57
notmynameI personally can't help a lot with nova, but I'm happy to answer any questions about swift02:57
notmynamewe have some docs describing how to set up swift in a VM so you can play around with it without any hardware deployment. in fact, it's what the swift devs use as a dev target02:59
* notmyname is a big fan of developing against a VM02:59
HugoKuoI installed SWIFT by All-in-one instruction in  a single machine03:01
HugoKuoshould I ask a question about openstack03:01
notmynamedon't ask to ask, just ask ;-)03:02
HugoKuois that openstack already can be use on production ?03:03
notmynameswift is used in production at Rackspace03:03
HugoKuoabout 36 nodes right ?03:03
notmynameI believe nova is used to some extent in production at nasa, but I don't know to what degree03:03
notmynamextoddx: ^03:03
notmynameHugoKuo: Rackspace's swift install base is vastly larger than 36 nodes03:04
HugoKuosoga .....03:04
xtoddxHugoKuo: I don't know the exact number of users, but we are in two locations and have lots of people using it in production03:04
HugoKuoI'm more clearly about nova & swift ......03:05
*** laurensell has quit IRC03:06
HugoKuoIn honest , I have to face another problem is API & RESTful03:06
*** johnbergoon has joined #openstack03:06
notmynameyes, both swift and nova have RESTful APIs03:07
notmynamextoddx: does nova also have xmlrpc?03:07
HugoKuoIn past while I test eucalyptus , I use web-console........03:07
xtoddxIt mimics the EC2 API that is not SOAP, which I guess is xmlrpc03:08
xtoddxthere is a web console for Nova coming soon03:08
xtoddxin the next few weeks, it just got approved for release03:08
HugoKuoawait ......03:10
*** jimbaker has joined #openstack03:14
*** jimbaker has quit IRC03:17
*** pvo_ has quit IRC03:18
*** jimbaker has joined #openstack03:18
*** johnbergoon has quit IRC03:29
HugoKuoa foolish question ..... how colud I store something into SWIFT by cmd on my single machine ?03:31
notmynameby cmd? you could use the st tool provided with swift or use curl03:31
*** iammartian has quit IRC03:32
*** iammartian_ is now known as iammartian03:32
HugoKuolet me think03:33
*** schisamo has quit IRC03:35
*** rsampaio has quit IRC03:39
*** rsampaio has joined #openstack03:40
*** rsampaio has quit IRC03:43
*** nelson__ has joined #openstack03:48
nelson__Silly openstack question: is there a single point of failure (e.g. an MRC)?03:49
notmynamenelson__: not for swift03:50
nelson__kewl, thanks.03:50
notmynamea quick search for MRC yields Military Revolutionary Committee (which I kinda like)03:51
nelson__Does swift support access to files via HTTP?03:51
nelson__(I'm reading the intro docs and not finding answers to these questions)03:51
notmynameswift supports whatever the proxy server will answer to03:52
notmynameusing SSL or not should be based on your particular auth scheme (a token based one as in the devauth server is a bad idea to leave non-SSL)03:52
notmynamebut, for example, I use non-SSL for my SAIO in a VM or on a VPS with no problems03:53
notmynamenelson__: does that make sense?03:55
nelson__ahhhh, I see, yes, it does. The files are exposed to the filesystem, so they can just be served by any http server.03:55
notmynamenot quite sure I follow that logic. but the entire system in exposed via the proxy server. and ssl is completely separate from swift03:57
notmyname("filesystem" is where I'm struggling with you comment)03:57
nelson__Ah! I should be reading this: http://swift.openstack.org/03:58
notmynamea great starting place :-)03:58
notmynamealthough it is currently lacking a good "what is this and why do I want it" overview, it has a lot of docs on the individual parts and how they work together03:59
nelson__yeah, that's what I'm needing. I have a checklist and need to fill in the OpenStack column.04:02
notmynamemy attempt at what the start of that doc would look like04:02
nelson__ah! I'll share a URL with you, too: http://whatthecommit.com/04:03
*** joearnold has joined #openstack04:07
*** mdomsch has joined #openstack04:10
*** mdomsch has quit IRC04:14
*** kashyapc has joined #openstack04:35
*** daleolds has joined #openstack04:35
*** rds__ has quit IRC04:41
*** omidhdl has joined #openstack04:49
*** rds__ has joined #openstack04:53
*** daleolds has quit IRC04:53
*** ArdRigh has quit IRC04:57
*** rds__ has quit IRC05:00
*** suchitp has joined #openstack05:05
*** suchitp has left #openstack05:05
*** AimanA is now known as HouseAway05:11
*** omidhdl has quit IRC05:20
*** omidhdl has joined #openstack05:21
*** f4m8_ is now known as f4m805:42
openstackHugoKuo: (help [<plugin>] [<command>]) -- This command gives a useful description of what <command> does. <plugin> is only necessary if the command is in more than one plugin.05:44
HugoKuo!help who05:44
openstackHugoKuo: Error: There is no command "who".05:44
openstackHugoKuo: Error: "who" is not a valid command.05:45
notmynameHugoKuo: why not poke the bots via /msg05:45
HugoKuothanks I just play around it's ok . first time to use IRC05:46
*** sirp1 has quit IRC05:48
*** aimon_ has joined #openstack05:52
openstackHugoKuo: Error: "hugo" is not a valid command.05:55
openstackHugoKuo: Error: "swift" is not a valid command.05:55
*** aimon has quit IRC05:56
*** aimon_ is now known as aimon05:56
notmynameHugoKuo: so if you precede the command with "/msg <nick>" you will talk to that person directly05:58
*** rds__ has joined #openstack05:59
*** joearnold has quit IRC06:04
*** stewart_ has joined #openstack06:11
*** stewart has quit IRC06:12
*** ArdRigh has joined #openstack06:18
*** aimon has quit IRC06:18
*** aimon has joined #openstack06:19
*** ramkrsna has joined #openstack06:22
*** ramkrsna has joined #openstack06:22
odyiAnyone have experience with VMs launched with nova using the nova.sh script aren't getting dhcp?06:52
*** guigui has joined #openstack06:57
HugoKuois that means you don't want to use dhcp for instances ?07:00
odyiNo meaning. The lucid ami shipped with testing just prints Discovering after it has booted.07:01
odyiNever gets at ipaddress assigned07:01
*** burris has quit IRC07:11
*** burris has joined #openstack07:12
*** ibarrera has joined #openstack07:16
*** miclorb_ has quit IRC07:19
*** gaveen has quit IRC07:19
HugoKuohave any error msgs in networker screen or others ?07:24
odyiNothing besides the sqlite locking issues which eventually go away.07:26
*** gaveen has joined #openstack07:30
HugoKuowhat's your ip address of  VMs host07:30
*** stewart_ is now known as stewart07:30
odyiIts NIC address is, the interface that nova-network creates is for listen to dhcp requests and "assigns" to the vm, but dhcp inside the VM's actual OS never gets the IP.07:32
*** aimon_ has joined #openstack07:33
*** aimon has quit IRC07:37
*** aimon_ is now known as aimon07:37
*** anotherjesse has joined #openstack07:47
*** befreax has joined #openstack07:47
HugoKuoI have a experience like this07:51
HugoKuobut i know the problem is that image07:51
HugoKuowhere is your image come from07:52
*** zaitcev has quit IRC07:53
HugoKuohttp://wiki.openstack.org/GettingImages    <~~~images over here are worked for me07:53
HugoKuohttps://answers.launchpad.net/nova/+question/134239     my experience....07:56
*** anotherjesse has quit IRC07:58
*** jkakar has joined #openstack08:06
*** almaisan-away is now known as al-maisan08:12
*** miclorb has joined #openstack08:26
*** ramkrsna has quit IRC08:38
*** calavera has joined #openstack08:44
*** kashyapc has quit IRC08:48
*** kashyapc has joined #openstack08:49
*** winston-d has joined #openstack08:49
winston-dHello? Anyone could help with my Bzr issue?  I have trouble branching NOVA source code08:50
*** allsystemsarego has joined #openstack08:52
*** allsystemsarego has joined #openstack08:52
*** stewart_ has joined #openstack08:52
*** stewart has quit IRC08:54
HugoKuomay I ?08:56
*** stewart_ is now known as stewart08:58
*** zykes- has quit IRC09:09
sorenwinston-d: What's the problem09:15
*** zykes- has joined #openstack09:15
sorenttx: Do you remember when the PPA publishers run?09:24
sorenttx: I thought it was every 20 minutes with a slight offset (something like: minutes % 20 == 2)09:25
*** tmarble has quit IRC09:27
*** tmarble has joined #openstack09:28
sorenttx: Never mind, I asked bigjools.09:28
*** KnightHacker has quit IRC09:30
*** KnightHacker has joined #openstack09:30
*** ramkrsna has joined #openstack09:34
*** ramkrsna has joined #openstack09:34
*** irahgel1 has joined #openstack09:35
*** rsampaio has joined #openstack09:36
*** meatz has quit IRC09:55
*** ptremblett has quit IRC09:57
ttxsoren: ok :)09:59
*** kashyapc has quit IRC10:00
*** kashyapc has joined #openstack10:00
*** aimon has quit IRC10:02
*** aimon has joined #openstack10:02
sorenkirkland: Maverick builds work again: https://launchpad.net/~nova-core/+archive/trunk/+build/205894910:03
sorenkirkland: Once someone gets a new branch merged, the Lucid builds should also get sorted out.10:03
*** miclorb has quit IRC10:04
*** arthurc has joined #openstack10:06
*** miclorb_ has joined #openstack10:12
*** suchitp has joined #openstack10:41
*** suchitp has left #openstack10:41
*** suchitp has joined #openstack10:53
ttx_0x44: around ?10:56
*** ArdRigh has quit IRC10:57
*** gaveen has quit IRC10:57
*** suchitp has left #openstack10:57
ttx_0x44: had a question for you about glance's teller-s3-backend and backend-imports blueprints. Is backend-imports a prerequisite of teller-s3-backend ?11:00
sorenttx: Do you have an opinion on when stuff should be "fix released"?11:03
*** ptremblett has joined #openstack11:03
sorenttx: When it's been merged or when it's been in a release?11:03
ttxsoren: well, ideally, we would use fix committed for the comit and fix released for the release... However the way LP keeps fix committed in buglists makes it slightly painful... and I don't think there is a nice way to move all fixcommitted -> fixreleased at release time.11:10
sorenttx: It's pretty simple with launchpadlib, actually.11:11
ttxsoren: if all bugs have a linked branch, probably11:11
sorenttx: bugs are the most thoroughly exposed object type in the lp api. You can do almost anything.11:11
*** omidhdl has quit IRC11:12
ttxsoren: would you say there is value for someone in knowing a bug has been fixcommitted but not released in a formal release ?11:12
sorenttx: Some.11:12
sorenttx: It tells them that it's fixed, but they just have to wait for the next release before they can get it.11:13
sorenttx: Unless they're cool with dev snapshots.11:13
ttxsoren: the drawback, for me, is that the buglists tend to be a bit crowded with fixcommitted bugs11:13
ttx...until you reach release.11:14
ttxsoren: i don't really have strong opinions either way.11:14
ttxlet's raise it at the meeting tonight.11:14
sorenttx: Me neither. I was hoping you had, so I could just adopt yours :)11:14
ttxI need to write the BugsLifecycle document to put some meaning into the bugs fields values11:15
ttx(the same way I did BlueprintsLifecycle)11:15
ttxsoren: I'm pretty sure some people will have a stronger opinion than we have.11:15
sorenSure. I just hope they're right.11:16
*** miclorb_ has quit IRC11:20
* soren lucnhes11:28
sorenlunches, even.11:28
*** MarkAtwood has quit IRC11:47
*** al-maisan is now known as almaisan-away11:56
*** kashyapc has quit IRC12:03
*** schisamo has joined #openstack12:07
*** kashyapc has joined #openstack12:11
*** kashyapc has quit IRC12:23
*** ctennis has joined #openstack12:27
*** GasbaKid_ has joined #openstack12:29
*** GasbaKid_ has quit IRC12:30
*** GasbaKid has joined #openstack12:31
*** that1dude has joined #openstack12:34
* soren <3 bzr lp-propose-merge12:46
ttxAliases: lp-propose12:47
* soren has it aliased to "p" :)12:49
*** omidhdl has joined #openstack13:19
*** omidhdl has left #openstack13:39
*** abecc has joined #openstack13:39
*** abecc has joined #openstack13:40
*** aimon_ has joined #openstack13:44
zulsoren: do you want me to upload a new version to the archive today?13:45
sorenzul: That would be great, but not right now. I have one more thing I want to change.13:45
zulsoren: sure just let me know when...13:46
zulsoren Daviey: from here on...i think tuesday should be the upload day13:46
*** aimon has quit IRC13:46
*** aimon_ is now known as aimon13:46
*** pvo_ has joined #openstack13:47
*** ChanServ sets mode: +v pvo_13:47
sorenzul, Daviey: Why have an "upload day"?13:48
sorenSomething to do with QA?13:48
*** jdarcy has joined #openstack13:48
zulsoren: no but i would just like to nail down a regular occurance of one13:49
Davieysoren: frequent is more important than regular IMO... but if a day per week is laid down, then it makes it easier to track and build QA processes around - perhaps?13:50
Davieysoren: What do you think?13:50
sorenDaviey: I'm just wondering why only once a week.13:50
zulsoren: gives more time to land things and test them before we upload13:50
Davieysoren: Is it moving enough to warrant more?13:50
sorenDaviey: I'm not sure.13:51
sorenDaviey: Perhaps not right now.13:51
sorenDaviey: Slow part of the cycle and all.13:51
zulsoren: also i would like to start porting our qa rig that we have as well13:51
zulporting might not be the right word in this case :)13:52
sorenIdeally, the upgrade process and all this stuff should be seamless enough that we could upload on each commit.13:52
Davieyzul: Are you talking with Carlos about setting up the preseeds for the QA lab?13:52
soren...although people might get annoyed by the constant updates.13:52
zulDaviey: yeah i was helping him in it13:52
Davieysoren: I agree, that might be overkill :)13:52
zulyeah i would get annoyed at the constant updates13:52
sorenDaviey: ...and if people want that, they can just enable the ppa.13:53
Davieysoren: /me imagines a commit -m "Fixed typo in some file nobody cares about" :)13:53
sorenDaviey: that's exactly what it does.13:53
*** kashyapc has joined #openstack13:53
*** kashyapc has quit IRC13:54
zulmaybe the one in the ubuntu archive might need a little less bleeding edge than the ones in the ppa13:55
zulsoren: the way i see it...long time down the road if a user finds a bug in the openstack version in the archive we can tell them to install the ppa version to see if its fixed and if it is we can do an SRU or something13:58
*** almaisan-away is now known as al-maisan14:01
sorenzul: Right, _down the road_ we may keep maintenance branches of releases open for a while. Like, for the version we will ship in Ubuntu 12.04 or whatnot.14:07
zulsoren: right14:07
soren(since it's an LTS)14:08
sorenI know you know, just providing context for everyone else :)14:08
* soren is very, very confused.14:09
*** ppetraki has joined #openstack14:12
*** dendro-afk is now known as dendrobates14:16
sorenHm. New error.14:19
zykes-from one to the next ?14:22
ttxjaypipes, _0x44: ping me when you're around. I'd like to clean up the glance/bexar BP list.14:22
sorenzykes-: Story of my life.14:28
zykes-alekibango: Is Clanavi out yet ? like downloadable ?14:29
*** rchavik has quit IRC14:33
*** gondoi has joined #openstack14:37
alekibangozykes-: afaik no14:37
alekibangozykes-: i would think that the web console from nasa will be interesting14:38
*** Ryan_Lane|away is now known as Ryan_Lane14:44
zykes-alekibango: that's internal for NTT only or ?14:45
*** f4m8 is now known as f4m8_14:47
zykes-alekibango: what about groundcontrol ?14:48
*** Podilarius has joined #openstack14:50
zykes-when is it reckoned that openstack will be productionable ?14:50
alekibangozykes-: i am not all knowing old magician :)   ask gundlach14:51
alekibango(about groundcontrol)14:51
alekibangozykes-: swift is in production14:52
alekibangonova --  when themselves will start using it... which is i think soon after bexar release14:53
creihtNasa is currently using nova in production14:53
zykes-you mean NASA then or ?14:53
alekibangozykes-: i mean companies who are making it, yes... creiht didnt know this14:53
zykes-creiht: all the features they have is in openstack nova allready ?14:53
creihtThe next releases are for RS to get the features needed to use it14:54
alekibangocreiht: live migration would be very nice to have14:54
zykes-RS ?14:54
creihtsorry :)14:55
creihtThat is about all I know on the Nova side of things :)14:55
zykes-k, doesn't nasa allready have most features if they have it in prod ?14:55
alekibangocreiht: imho they are using somewhat different code....14:55
creihtvishy, xtoddx -^ ???14:56
uvirtbotNew bug: #680504 in nova "Admindocs fail to build." [Undecided,New] https://launchpad.net/bugs/68050414:56
*** damon__ has joined #openstack14:57
alekibangocreiht: btw last line from me should be 'creiht: I didnt know this' sorry for missing ": I"14:58
creihtalekibango: I interpreted it as suck :)14:58
creihterm such14:58
alekibangojust making sure14:58
creihttoo early in the morning14:58
* annegentle giggles14:59
*** jonkelly has joined #openstack15:01
jaypipesttx: just grabbing some coffee/breakfast.  back in 30 min15:02
*** kashyapc has joined #openstack15:04
pvo_alekibango: they're using code that has been adapted to their environment, sure. But the bulk of the code is nova.15:06
*** pvo_ has quit IRC15:08
*** pvo_ has joined #openstack15:08
*** dubsquared has joined #openstack15:10
*** perestrelka has quit IRC15:14
openstackhudsonProject nova build #255: SUCCESS in 1 min 13 sec:
*** perestrelka has joined #openstack15:15
*** kashyapc has quit IRC15:18
*** sirp1 has joined #openstack15:19
*** f4m8__ has joined #openstack15:21
*** atobey_ has joined #openstack15:21
*** jpipes has joined #openstack15:21
*** jeremyb_ has joined #openstack15:21
*** patri0t_ has joined #openstack15:22
*** rbergero1 has joined #openstack15:23
*** ramkrsna has quit IRC15:24
*** redbo has quit IRC15:24
*** rossij has quit IRC15:24
*** jeremyb has quit IRC15:24
*** patri0t has quit IRC15:24
*** jaypipes has quit IRC15:24
*** atobey has quit IRC15:24
*** rbergeron has quit IRC15:24
*** f4m8_ has quit IRC15:24
*** rossij has joined #openstack15:24
*** redbo has joined #openstack15:25
*** ChanServ sets mode: +v redbo15:25
*** mdomsch has joined #openstack15:27
*** holoway has quit IRC15:30
*** ivan has quit IRC15:30
*** spectorclan has joined #openstack15:33
sorenvishy: ping15:37
sorenvishy: I have a few questions about lp:~vishvananda/nova/fix-flatdhcp15:37
sandywalshwith bzr, should you still do a merge with trunk before commit/pushing? (as you would with git)15:38
*** kashyapc has joined #openstack15:38
annegentlesandywalsh: the merge happens after commit/push with bzr I believe15:39
sandywalshannegentle, ah, so my push will only have my changes in it ... nothing that has happened in trunk since I started, correct?15:40
creihtsandywalsh: depends on your context15:40
sorensandywalsh: You don't *have* to merge with trunk, but it's generally a good idea.15:40
annegentleok, thanks creiht15:40
sorensandywalsh: Just to make sure noone pushed anything that would conflict with your changes.15:40
creihtin the launchpad workflow, your merge proposal will get auto-merged after it is accepted15:40
sandywalshok, it just looks like my branch is now cluttered with other changes15:41
creihtYou only *have* to merge your branch with trunk if there is a conflict15:41
sandywalshno conflicts15:41
creihtthen I wouldn't worry about it, and let hudson do its magic :)15:42
sandywalshgroovy ... thanks all15:42
sorensandywalsh: How does it look like it's cluttered with other changes?15:43
sandywalshsoren, the merge from trunk brought in all the changes since I started. I'm suspecting my commit/push will show, not only my changes, but a repeat of these other changes too. Unless I'm mistaken.15:46
sorenNono. Just your changes.15:46
sorenSince everything else is already in trunk.15:46
sandywalshah, good ... no worries then. thx15:46
sandywalshand when I push, it's in my own branch, so I can do integration testing against my branch and not affect other work in progress, correct?15:47
sandywalshI'm not pushing to trunk directly15:48
sandywalshfor example: bzr push lp:~sandy-walsh/nova/admin-only-api15:49
sorensandywalsh: Right.15:50
*** westmaas has quit IRC15:50
zulsoren: vishy's branch fixed the build failure for me15:53
*** rbergero1 is now known as rbergeron15:53
sorenzul: Which branch?15:54
sorenzul: Which build failure did it fix?15:54
zulsoren: the testsuite failure15:54
sorenzul: Err... I seriously doubt it affected it.15:55
sorenzul: it doesn't even go near the docs.15:55
zulsoren: right the doc failure is because of the sphinx version i was using the test failure is something else i saw15:57
sorenErr... So which failure did it fix again?15:57
sorenHave I heard of it?15:57
zulsoren: the testsuite was failing because it was trying to un allocate an ip address i dont think you have heard about it yet15:58
zulill open a bug about it just a sec15:58
jpipesttx: back now if you want to chat...15:59
ttxjpipes: just a sec16:00
*** joearnold has joined #openstack16:00
creihtkarmabot: list16:03
creihtkarmabot: take me to your leader16:03
creihttoo many bots in this channel :)16:03
zulsoren: bug 68054316:04
uvirtbotLaunchpad bug 680543 in nova "Testsuite failure" [Undecided,New] https://launchpad.net/bugs/68054316:04
creihtsoren: do you know if there is a way to get emailed when blueprints get created for a project?16:05
sandywalshAnyone care to review: http://bazaar.launchpad.net/~sandy-walsh/nova/admin-only-api/revision/40916:06
zulcreiht: subscribe to the blueprints?16:06
*** jdarcy has quit IRC16:07
jk0you can't subscribe to a blueprint if it doesn't exist yet :)16:07
*** joearnold has joined #openstack16:07
gholtSo how are we to know when new blueprints are created?16:07
gholtOr should we just not care? Heh16:08
sorencreiht: I don't think so, no.16:09
vishysoren: pong16:09
sorenvishy: Yay. One question in particular is on my mind:16:09
sorenvishy: Looking at https://code.launchpad.net/~vishvananda/nova/fix-flatdhcp/+merge/4111816:09
sorenvishy: When calling _on_set_network_host..16:09
vishyzul: that bug looks like the annoying race condition that was showing up16:10
sorenvishy: We update the given network, setting dhcp_start to FLAGS.flat_network_dhcp_start.16:10
ttxarh, fixing breakage takes longer than expected16:10
sorenvishy: Does that mean all networks for which this host is the designated network host, will use the same dhcp start IP?16:10
vishysoren: yes16:10
sorenvishy: That seems... wrong.16:10
vishythere is only one dhcp server for all addresses16:11
vishymeaning there should only be one network16:11
vishyper controller16:11
sorenYes, I suppose that makes sense.16:11
*** ivan has joined #openstack16:11
sorenvishy: Rocking. In that case, It's great.16:11
uvirtbotNew bug: #680543 in nova "Testsuite failure" [Undecided,New] https://launchpad.net/bugs/68054316:11
sorenDaviey, zul: One more tiny fix to the packaging, and I'm happy to have it uploaded.16:11
zulsoren: cool16:12
vishysoren: i really don't know how useful it is at scale, it is outside our use case so I'm hoping someone who actually be using it in production will take it over and figure out how it "should work"16:12
vishysoren: any thoughts on the question about what should be the default networking mode?16:13
sorenvishy: I don't really expect it to be used on a large scale.16:13
*** joearnold has quit IRC16:13
sorenvishy: I'm leaning towards VlanManager.16:13
vishysoren: not sure it really plays well with public ips by default16:14
sorenvishy: VlanManager?16:14
vishysoren: yeah16:14
sorenvishy: I thought it was what you guys used?16:14
vishysoren: or was that a response to my default mode question16:14
*** patri0t_ is now known as patri0t16:14
sorenHeh... Was what a response to your default mode question? :)16:15
sorenLet's start over :)16:15
vishysoren: so the default issue is that VlanManager needs a bridge_dev16:15
vishywhereas FlatManager will splode if you set a bridge_dev that already has an ip16:16
sorenIs that a bridge or a physical interface that gets hooked into the bridge?16:16
vishybridge_dev is the physical device16:16
vishyeth0 was the default16:16
sorenOk. that's rather confusing :)16:16
vishyi suppose it is :)16:16
sorenPerhaps "bridged_dev"?16:17
vishysure, but regardless, should we have two flags?16:17
vishyvlan_bridged_dev and flat_bridged_dev?16:17
*** mdomsch has quit IRC16:18
vishyor just hope people don't blow it up16:18
*** zaitcev has joined #openstack16:19
vishyI'm going to add some long docstrings to the managers explaining what they do regardless16:19
*** holoway has joined #openstack16:21
*** jkakar has quit IRC16:22
*** al-maisan is now known as almaisan-away16:22
sorenI think two flags may be a good diea.16:24
sorenidea, even.16:24
sandywalshis there a tool for issuing openstack api calls from the cmdline? (other than curl)16:24
sorenFrom python-cloudservers.16:25
soren(jacobian's cloudservers, not mike mayo's)16:25
sandywalshgreat ... thanks16:25
sorensandywalsh: Sure.16:26
sorensandywalsh: It's in Ubuntu.16:26
sandywalshsoren, even better16:26
*** joearnold has joined #openstack16:27
sorenvishy: Do you want to make that change to the fix-flatdhcp branch right away?16:28
sorenvishy: If so, I'll hold off on approving it. Otherwise, it's good (as expected).16:29
ttxsirp1: around ?16:29
* soren dinners16:30
* soren verbs some more nouns16:30
*** hazmat has joined #openstack16:30
ttxsirp1: I had a question for you about the bexar-teller-caching blueprint16:30
ttxsirp1: is it work you think you can do for the bexar release ?16:30
uvirtbotNew bug: #680551 in nova "Stop referring to FAKE_subdomain in docs" [Undecided,New] https://launchpad.net/bugs/68055116:31
sirp1ttx: at this point im not sure, but i'm beginning to think that caching may have to slip into the next release (since it's really just a nice-to-have)16:31
*** calavera has quit IRC16:32
sirp1ttx: there are some other blueprints that i'm working on that def are taking priority16:32
ttxsirp1: ok, we'll untarget it from bexar for the moment, and we'll bring it back into scope later if needed16:32
sirp1ttx: sounds good16:32
ttx(trying to keep the bexar list to match what we intend to do)16:32
ttxsirp1: thanks !16:32
ttxjpipes: so for the moment you can unaccept for bexar https://blueprints.launchpad.net/glance/+spec/bexar-teller-caching16:33
vishysoren: yeah I'll do that change and docstrings16:33
jpipesttx: done! :)16:34
ttxjpipes: but you didn't "accept for bexar" https://blueprints.launchpad.net/glance/+spec/clients yet.16:34
vishyomg its an imposter jaypipes!16:34
ttxvishy: he seems to have the smae LP powers, so suits me fine :)16:35
ttxsame, even.16:35
*** almaisan-away is now known as al-maisan16:37
jpipesttx: I was waiting for you to do that ;)16:37
jpipesttx: done :)16:37
ttx_0x44: around ?16:39
*** jkakar has joined #openstack16:39
*** joearnold has quit IRC16:40
* ttx breaks for dinner16:44
*** grizzletooth has joined #openstack16:48
*** grizzletooth has left #openstack16:53
odyiAnyone else hav issues with VlanManager not giving out dhcp to instances? Using the novascript for setup on a single machine.16:54
kirklandsoren: thanks16:59
kirklandsoren: i'll rebuild16:59
*** rsampaio has quit IRC17:02
*** daleolds has joined #openstack17:09
*** guigui has quit IRC17:13
*** maplebed has joined #openstack17:14
*** arthurc has quit IRC17:15
spectorclanalekibango: Looking at http://wiki.openstack.org/Overview and it seems that there are 2 options for the solution? Has one or the other been chosen?17:16
*** jkakar has quit IRC17:21
*** maplebed has quit IRC17:21
jk0should aws_access_key_id and aws_secret_access_key be consistent throughout the entire code base? I'm seeing several different variations17:30
*** ibarrera has quit IRC17:32
kirklandsoren: hmm, my ISO build is still picking up the bzr397 packages from https://launchpad.net/~nova-core/+archive/ppa/+packages17:34
kirklandsoren: i suspect because i386 is FTBFS?17:34
*** dizz has joined #openstack17:35
*** jeremyb_ has quit IRC17:35
*** jeremyb_ has joined #openstack17:35
*** jeremyb_ is now known as jeremyb17:35
*** kashyapc has quit IRC17:41
*** maplebed has joined #openstack17:43
kirklandsoren: hmm, should i be building against trunk?17:43
kirklandsoren: i've been building against "ppa"17:44
*** joearnold has joined #openstack17:48
jk0is the control_exchange flag the default region?17:52
*** al-maisan is now known as almaisan-away17:54
*** MarkAtwood has joined #openstack17:56
*** dendrobates is now known as dendro-afk17:56
*** dendro-afk is now known as dendrobates17:57
* creiht sighs17:58
creihtThe built in python elementtree libs didn't include SimpleXMLWriter17:59
*** nelson__ has quit IRC18:01
*** nelson__ has joined #openstack18:01
zulkirkland: can you pull the packages out of the archive and use that when you build the image18:03
vishyanyone remember what was causign the double defined flags?18:10
alekibangospectorclan: i cant really say, vishy could you coment on what spectorclan is asking?18:18
vishyalekibango, spectorlan: option 1 is closer to what is being worked on18:19
alekibangoi would love having that documented better :)18:20
alekibangofor example using diagrams18:20
alekibangolike those i showed to yuo18:20
alekibango... even if this openoffice drawing is pretty nice and explanatory18:21
alekibangowe need to get the big - and the detailed big picture right to be able to grok it...18:21
alekibangoand to effectively help achieving the goal18:22
vishythis is the current plan for changes, maybe eday / pvo can hack up their diagrams a bit to make them more accurate18:23
annegentlethanks Vish18:23
alekibangothanks, looks like good work18:23
*** dizz has quit IRC18:24
dendrobates eday: are you planning on working on this https://blueprints.edge.launchpad.net/nova/+spec/bexar-openstack-api-floating-ips18:26
dendrobatesor did you put it in for ozone?18:26
pvovishy: have we figured out what an aggregation level would look like/18:29
pvore: distributed scheduler wiki18:29
vishypvo: i think we pushed it to implementation specific18:30
pvoreference implementation would be central datastore, large implementation could be local datastore to compute node?18:31
kirklandsoren: okay, rebuilt from the trunk ppa;  get's closer, looks like openssl issues are solved18:31
kirklandsoren: still failing at downloading the zipfile18:31
pvoI know there will always be differences between large installations18:31
jpipessirp1: made all those changes you suggested in your review of the glance client coding.  feel free to check out the updated diff.18:41
sirp1jpipes: cool ill take a look shortly18:41
jpipessirp1: cheers18:42
*** ptremblett_ has joined #openstack18:43
*** ptremblett_ has quit IRC18:43
*** ptremblett has quit IRC18:44
*** dendrobates is now known as dendro-afk18:45
vishypvo: well probably not compute node writing for security concerns.  So compute node passes message up and it is written by the next layer18:47
vishypvo: or aggregated as necessary18:47
vishypvo: there will be some local state on compute, whether it is a local db or file system or hypervisor18:48
jpipessandywalsh: heya, I see your admin-api blueprint is Needs Code Review, but I don't see a merge proposal... am I missing something?18:48
*** dendro-afk is now known as dendrobates18:51
dendrobatesjpipes: I think he means almost needs code review18:52
sandywalshjpipes, sorry still kinda unclear on the process. Thought code review came before merge proposal18:53
jpipessandywalsh: no worries, mate. :) yeah, Needs Code Review means that there is a merge proposal for a branch for others to review.18:54
sandywalshjpipes, k, thx. I'm nearly done integration testing so I'll flip the switch shortly.18:55
jpipessandywalsh: sounds good!18:55
jpipesjk0: you're on a bug fixing roll today...18:55
jk0that's just how I roll ;)18:55
dendrobatesxtoddx: how likely are you to get to this in bexar: https://blueprints.edge.launchpad.net/nova/+spec/nova-paste-deploy18:55
xtoddxdendrobates: didn't made it to this sprint, I'll talk to our PM about it18:56
dendrobatesxtoddx: it's a really good idea, better sooner than later, imho18:57
*** jinesh_ has joined #openstack18:59
*** hggdh has quit IRC19:00
*** hggdh has joined #openstack19:00
*** jinesh_ has quit IRC19:01
Ryan_Lanejpipes: good thing is that the two schemas should cover all popular ldap servers, except for maybe active directory19:02
*** 50UAABCVA has joined #openstack19:02
jpipesRyan_Lane: :) ya, well done on that.  you're a better man than I.19:02
* jpipes looks at LDAP schemas and wants to throw up.19:02
Ryan_Laneheh. ldap support is easier than database server support19:02
jpipesRyan_Lane: hmm, I don't know about that... :)19:03
jpipesRyan_Lane: but maybe I'm just set in my old ways, eh? ;)19:03
Ryan_Lanepretty much there is sun, openldap, and ad, and openldap :)19:03
Ryan_Lane(I don't like openldap ;))19:03
jpipessoren: pep8 strikes again!19:05
jpipesRyan_Lane: what do you use at work?19:05
Ryan_Laneopendj (was previously opends)19:05
*** adrian_otto has quit IRC19:06
*** westmaas has joined #openstack19:06
*** jdarcy has joined #openstack19:08
openstackhudsonProject nova build #256: SUCCESS in 1 min 12 sec:
jk0does the last comment here: https://bugs.launchpad.net/nova/+bug/655171 seem like a viable option?19:16
uvirtbotLaunchpad bug 655171 in nova "SQL Errors on a fresh install" [Medium,Confirmed]19:16
jk0it would be easy enough to add, I don't see a reason not to unless I'm overlooking something?19:16
sandywalshso, I'm trying to get python-cloudservers to auth against nova api19:24
sandywalshhas anyone done this before?19:25
sandywalshI had to change the URL in their code from19:25
sandywalshAUTH_URL = 'https://auth.api.rackspacecloud.com/v1.0'19:25
sandywalshto  AUTH_URL = ''19:25
sandywalsh(note the http vs https)19:25
sandywalshbut the auth request gets no route19:26
sandywalshDEBUG:routes.middleware:No route matched for GET /v1.019:26
sandywalsh192.168.2.109 - - [23/Nov/2010 11:11:13] "GET /v1.0 HTTP/1.1" 404 175 0.00073419:26
sandywalshis there a special configuration I need to do in api to enable the OS api?19:26
nelson__Sorry if this is a stupid question, but I've  been poring through http://swift.openstack.org and it doesn't tell me how the files stored in swift get accessed.19:27
annegentlenelson__: Not stupid at all. You access your data from a web client using the REST API19:28
jpipessandywalsh: see flags in /nova/api/__init__.py19:30
jpipessandywalsh: FLAGS.osapi_subdomain.19:30
sandywalshjpipes, looking at it right now. But that only defines the subdomain, correct? os.mysite.com vs. ec2.mysite.com ?19:30
sandywalshhow about if I'm using IP address19:31
jpipessandywalsh: hmm, not sure! :) one sec...19:31
annegentlenelson__: there's a short walkthrough at the bottom of http://swift.openstack.org/howto_installmultinode.html19:32
nelson__annegentle: sweet. So our squids will be happy just issuing a GET for the files being accessed.19:32
jpipessandywalsh: try setting FLAGS.osapi_subdomain to ''19:33
sandywalshjpipes, yup, thx. Messing with it now. That's gotta be it.19:34
*** befreax has quit IRC19:34
nelson__I'm likin' it! It's as if somebody looked at our requirements and pre-coded a solution!19:35
jpipessandywalsh: looks like a new flag "default_api" needs to be created and used as the default for nova.api.API's constructor...19:36
sandywalshjpipes, ok, I'll have a look at it and add a bug if needed. :)19:37
*** rsampaio has joined #openstack19:38
*** Ryan_Lane is now known as Ryan_Lane|food19:40
annegentleI've got a couple of doc review/merge requests in https://code.launchpad.net/nova/+activereviews if anyone's up for easy reading (not easy listening, ha)19:40
annegentlethey're in lp:~annegentle/nova/novanetdoc and lp:~annegentle/nova/apihowto19:41
*** GasbaKid has quit IRC19:42
*** ptremblett has joined #openstack19:43
jk0anyone see a problem with this re sqlite locking? http://paste.openstack.org/show/183/19:44
jk0it seems to work as far as I can tell19:44
*** burris has quit IRC19:49
edayjk0: You fixed a bug gundlach already had a branch for, but I'm not sure if we really want to change it (the fault name)19:52
*** jc_smith has joined #openstack19:52
jk0eday: I'll get rid of it19:55
uvirtbotNew bug: #680618 in swift "log-stats-collector logs as "swift"" [Low,New] https://launchpad.net/bugs/68061819:56
sandywalshjk0, that would be awesome if it works ... I had to switch to mysql19:57
jk0sandywalsh: https://code.launchpad.net/~jk0/nova/lp655171 give it a shot19:57
sandywalshhmm, but it won't be the same thread right19:57
*** Ryan_Lane|food is now known as Ryan_Lane19:57
sandywalshdifferent processes hitting the same file19:57
sandywalshnot different threads hitting the same library19:58
*** GasbaKid has joined #openstack20:01
edayjk0: do you have a test case that reproduces the locking problem?20:01
nelson__another possibly dumb question: my sysadmins are wanting to know if they can access files on the Object Server by the same name they're known as publicly, or if they need to go through a program.20:02
jk0eday: I've only actually seen it once, but I've heard several complaints about it. I figured it might not hurt for some people to try that branch and see if it helps20:02
jk0eday: but I'm not able to reproduce it at the moment20:02
alekibangonelson__: this is becoming a faq but its not dumb...    note that swift is object store. its not filesystem...20:03
edayjk0: I worry if checks are disabled it might cause corruption20:03
dendrobatessandywalsh: This explains all the fields in a blueprint >> http://wiki.openstack.org/BlueprintsLifecycle20:03
jk0eday: ew, we don't want that20:03
sandywalshdendrobates, great ... instapaper!20:03
ttxTeam meeting in one hour in #openstack-meeting !20:04
nelson__alekibango: I'm guessing "no" then. So it looks like they'll have to have to look at files using swift.obj.server.DiskFile20:04
*** westmaas has quit IRC20:04
alekibangonelson__: i am not really sure if there is no mapping system :)  iam more into nova20:05
alekibangobut generally, it stores objects, not files in filesystems...20:05
edayjk0: I'm wondering if there is a legitimate query that is just taking a while to commit, causing the locking issue20:05
edaythe default timeout is 5 seconds20:05
nelson__I'm not sure why they asked that question. I'll just say "no" and them them whinge at me. :)20:05
uvirtbotnelson__: Error: "them^let" is not a valid command.20:06
sandywalsheday, it could be ... when I saw it the vm was under load20:06
alekibangothey think its scalabe filesystem20:06
alekibangolook for lustre or other systems like that20:06
jk0eday: I think sandywalsh might still be able to reproduce it20:06
nelson__alekibango: I'll evaluate lustre also, but it looks like swift is meeting our requirements better.20:06
alekibangonelson__: :)20:06
sandywalshjk0, eday yup ... I'll try again after I get my merge in20:06
nelson__"Object stored as files" is just a want, not a need.20:07
*** ctennis has quit IRC20:07
creihtnelson__: there is a command like tool to get the info you want20:07
sorenkirkland: Sorry, there's a new PPA in town.20:07
nelson__I'll tell them: You're sysadmins. This is Unix. Stop acting so helpless.  :)20:07
alekibangonelson__: glusterfs is nice, but slow...20:07
sorenkirkland: ppa:nova-core/trunk20:07
alekibangoceph could be interesting for me, but still experimental20:07
creihtnelson__: that will tell you what nodes and what exact location on the nodes that the files are at20:08
nelson__alekibango: have looked at glusterfs. It will be one of the two, and yes, ceph isn't production yet. creiht: ahhhh, THAT is what they're wanting, I'm sure.20:08
alekibangoty creiht, i already exhausted my swift knowledge20:09
*** dragondm has joined #openstack20:09
*** pvo has quit IRC20:09
*** pvo has joined #openstack20:09
*** ChanServ sets mode: +v pvo20:09
alekibangonelson__: also, swift is 'cold' storage20:10
creihtnelson__: I am going to need to run some more errands shortly, do you have any other swift questions?20:11
zaitcevApropos Nelson's Q, do we have a client for Swift that presents a FuSE interface? if not, do we want one?20:12
creihtzaitcev: redbo has a start of one called cloudfuse20:12
nelson__creiht: no, thanks, I had another question, but got it answered last night. swift looks pretty good!20:12
zaitcevShit they already implemented all the obvious ideas.20:13
creihtnelson__: awesome... thanks :)20:13
creihtzaitcev: lol20:13
creihtzaitcev: I'm sure redbo would welcome some love to cloudfuse :)20:13
zaitcevcreiht: Thanks, saved.20:14
*** greenisus has joined #openstack20:15
vishydelayed imports + gflags = pain20:19
sorenvishy: Indeed.20:19
sorenkirkland: Can you describe how you reproduce this openssl error? I can't trigger it.20:19
vishysoren: DuplicateFlagError...I'm running into it because of a late import_object chain.  Have you seen that before?20:20
sorenvishy: Not that I recall.20:21
vishyquite annoying20:21
sorenvishy: Would you mind sharing the branch so that I can see it?20:21
sorenvishy: I'm curious how that can happen.20:21
*** ctennis has joined #openstack20:21
vishysoren: it is happening when I'm trying to unpack context that has gone through the queue20:22
sorenvishy: unpack as in unpickle?20:22
sandywalshis EC2_ACCESS_KEY = CLOUD_SERVERS_USERNAME or CLOUD_SERVERS_ACCESS_KEY ... or something else?20:22
vishysoren: context.from_dict()20:23
vishywhich is calling AuthManager.__init__20:23
*** rlucio has joined #openstack20:24
*** dendrobates is now known as dendro-afk20:25
vishywow my internet is biting it20:25
zykes-hmm, bexar-django-nova-ui - what's that ?20:27
sorenvishy: i still don't get it. All flags are defined at module level. Importing a module that is already importet should reuse the already imported module, not import it again.20:29
redboI dunno, adding anything else to cloudfuse would have to make some choices about how it's used.20:30
vishyyeah apparently it doesn't though20:30
redbowhich is okay, but I don't know which direction it should go20:30
sorenvishy: Hence my desire to inspect what's going on :)20:31
vishydubya tee eff20:31
vishyso apparently the issue disappears if get_my_ip works properly20:32
vishythat is so freaking random20:32
sorenget_my_ip doesn't work properly? You've got no routing?20:33
vishythis is the old version of get my ip20:33
vishybefore you fixed it20:34
vishybut if it throws an exception in messes up flag loading20:34
vishyand later dynamic imports will bomb20:34
edayvishy: probably because it causes the import throws halfway through, you might already have some flags added20:36
edayvishy: are the duplicate flags the ones defined before the flag that call get_my_ip() on module load?20:36
vishyeday: yeah it makes sense now that I found the underlying issue20:36
vishytime to port soren's patch to deploy20:37
*** rsampaio has quit IRC20:38
*** dendro-afk is now known as dendrobates20:42
ttxTeam Meeting in 15 minutes in #openstack-meeting20:47
jk0thanks for the heads up ttx20:47
*** daleolds has quit IRC20:48
*** adjohn has joined #openstack20:51
*** burris has joined #openstack20:52
*** sirp1 has quit IRC20:54
sorenttx: Wow, time flies.20:55
ttxsoren: yep20:58
ttxMeeting starting *NOW* in #openstack-meeting, please join20:59
zullike right now?20:59
ttxlike now.21:00
openstackhudsonProject nova build #257: SUCCESS in 1 min 13 sec:
sorenvishy: It's baaaaack: https://code.launchpad.net/~annegentle/nova/apihowto/+merge/4164421:02
vishysoren: yeah it seems to still happen 1/10 times or so21:05
vishyyay for eventlet21:05
* soren attempts to contain his excitement.... and succeeds21:05
*** mdomsch has joined #openstack21:07
*** adrian_otto has joined #openstack21:08
vishysoren: I guess we need to figure out how to wait for a call to finish in eventlet instead of doing the silly sleep in thread21:08
*** adrian_otto has quit IRC21:08
openstackhudsonProject nova build #258: FAILURE in 9 sec:
*** adrian_otto has joined #openstack21:09
sorenvishy: Yes. Yes we do.21:10
*** sirp1 has joined #openstack21:17
vishyecay: any ideas?21:23
vishyeday: ^^21:23
edayvishy: it already handles it, sleep is monkey patched21:24
vishyeday: sorry, let me be clearer21:24
vishyeday: in cloud_unittest we are polling the reactor21:25
vishyeday: some of the methods need to wait until return before moving on21:25
vishyeday: for example, self.cloud.release_address needs to finish before db.floating_ip_destroy is called21:26
vishyeday: in the twisted model you could just put a yield and it would wait for the call to finish before moving on21:27
edayok, one sec. looking at meeting window :)21:27
*** hornbeck has joined #openstack21:30
*** daleolds has joined #openstack21:39
zykes-j openstack-meeting21:41
*** abecc has quit IRC21:43
*** laurensell has joined #openstack21:45
*** jonkelly has quit IRC21:47
*** iammartian has quit IRC21:47
*** daleolds has quit IRC21:50
edayvishy: looking...21:50
*** daleolds has joined #openstack21:50
*** joearnold has quit IRC21:50
edayvishy: eek, yeah those sleeps in there are nasty. the call should block until it's complete21:51
vishyeday: threw greenthread.sleep() to fix the build but it still races sometimes, I'm assuming there is some way to wait for a thread to finish before continuuing21:51
edayvishy: other green threads may run at the same time though21:52
*** joearnold has joined #openstack21:52
*** irahgel1 has left #openstack21:52
*** rlucio has quit IRC21:52
vishyeday: so how do you wait for the calls to finish in eventlet world?21:52
*** joearnold has quit IRC21:53
zykes-anyone know what distribution that nasa runs nebula on ?21:53
edayvishy: you just wait, they don't return until they are done21:54
vishyzykes-: lucid21:54
zykes-just standard ubuntu ?21:54
vishyeday: apparently not21:54
vishyeday: without the sleep in there destroy gets called to fast21:55
edayvishy: well, everything in self.cloud.release_address has completed when it returns21:55
edayif it's doing a cast() or something in there there is no guarantee it was processed by the worker, but that a queue sync issue, not eventlet21:55
vishyeday: also: I'm assuming that the yields in test_console_output are useless21:55
vishyeday: it is a cast and it doesn't get processed until the polling in attach_to_eventlet hits21:56
ttxas a sidenote, it's IMHO normal to have 20% of people from the valley when you have a summit in the US. If you run a summit in Europe, you're likely to have 50% of attendees coming from Europe... In the end what matters is which local community you want to tap into21:56
soren\o/ We have per-commit builds working again for Lucid, Maverick, and Natty now. Yay.21:56
edayvishy: ok, so there really needs to be a call() ro make sure the queue got processed or something21:57
vishyeday: so we need a way to make sure that the polling happens21:57
vishymaybe a helper wait_for_clean_queue() or something?21:57
edayvishy: well, more geenrically, we probably want some way to know when a specific task completes (optionally)21:58
edayie, flag in the cast() message that says cast back to me when you're done (eventually supporting webhooks too)21:58
*** laurensell has quit IRC21:58
openstackhudsonProject nova build #259: STILL FAILING in 9.1 sec:
vishyeday: sure, for the moment, I'd just like the test to be reliable21:59
edaythe quick hack is to increase sleep() :/21:59
*** clayg has quit IRC21:59
vishyor i could turn the cast()s into call()s22:00
vishyfor release/associate/disassociate/22:00
*** clayg_ has joined #openstack22:00
sorenopenstackhudson: erk.22:00
vishywe could make cast always use call in test mode22:01
edayassuming the fake queue mechanism does something that will let eventlet run another threads )socket io or sleep)22:01
*** masumtok has joined #openstack22:01
edayvishy: eww22:01
edayvishy: we should be testing what actually happens :)22:01
edayas a general rule, if we're adding features just to make tests run, we're doing something wrong22:02
* ttx heads bedwards22:04
sorenCan someone please review/approve this? https://code.launchpad.net/~soren/nova/doc-build-dir/+merge/4166222:05
sorenIt's blocking everything else:
*** damon__ has quit IRC22:09
sorenPretty please?22:09
*** miclorb has joined #openstack22:09
annegentlesoren: doh! Did I do that (remove doc/build) - I was trying to avoid adding in all the doctree dir. Sorry.22:10
annegentlesoren: and you're so nice not to point fingers in the merge request :)22:10
sorenannegentle: Possibly. Don't worry about it. It's an easy fix.22:10
sorenannegentle: :)22:10
sorenWell, assuming someone would go ahead and review the darn thing :)22:11
* jk0 would be more than happy to if he could22:13
sandywalsh:) Has anyone ever tried python-cloudserver with the openstack api?22:14
sandywalshI'm deep in the bowels of cloudserver and openstack auth right now ... there doesn't seem to be a bottom.22:14
sorenI haven't actually, no.22:15
sandywalshI'm almost there ... will post findings22:16
*** sirp2 has joined #openstack22:20
*** sirp1 has quit IRC22:21
*** jdarcy has quit IRC22:24
uvirtbotNew bug: #680693 in swift "st download -o - still creates the directories" [Medium,New] https://launchpad.net/bugs/68069322:26
*** spectorclan has quit IRC22:28
sorenjk0: We should fix that. Soon.22:36
* soren heads bedwards22:36
jk0gnight dude22:36
sandywalshaha ... after authentication, nova returns 'x-Server-Management-Url' which differs from the desired url. We need a way to unify the definitions of our public facing endpoints22:43
*** ArdRigh has joined #openstack22:45
*** miclorb has quit IRC22:46
*** ppetraki has quit IRC23:02
*** clayg_ is now known as clayg23:07
*** greenisus has quit IRC23:10
*** greenisus has joined #openstack23:10
dubsquaredDEAD MAN WALKING23:11
openstackhudsonYippie, build fixed!23:15
openstackhudsonProject nova build #260: FIXED in 1 min 12 sec:
*** mdomsch has quit IRC23:17
*** gondoi has quit IRC23:20
uvirtbotNew bug: #680709 in nova "Auth should allow for fully customized URL's when defining x-Server-Management-Url" [Undecided,New] https://launchpad.net/bugs/68070923:21
*** rlucio has joined #openstack23:26
*** rlucio has quit IRC23:32
*** rlucio has joined #openstack23:39
*** adjohn has quit IRC23:46
*** adrian_otto has quit IRC23:55

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!