Monday, 2011-02-21

*** maplebed has joined #openstack00:17
*** MarkusT has quit IRC00:18
*** maplebed has quit IRC00:18
*** jbarratt has quit IRC00:23
*** jbarratt has joined #openstack00:24
*** zenmatt has joined #openstack00:27
*** adjohn has joined #openstack00:38
*** olivier_ has quit IRC00:50
*** olivier_ has joined #openstack00:51
*** RobertLaptop has quit IRC00:53
*** maplebed has joined #openstack01:09
*** maplebed has quit IRC01:11
*** RobertLaptop has joined #openstack01:28
*** matiu_ has joined #openstack01:31
*** matiu_ has joined #openstack01:31
*** matiu has quit IRC01:34
*** matiu_ is now known as matiu01:34
*** Ryan_Lane has quit IRC01:38
*** littleidea has joined #openstack01:46
*** matiu has quit IRC02:06
*** rds__ has quit IRC02:06
*** jsgotangco has joined #openstack02:08
*** miclorb_ has quit IRC02:10
*** hadrian has joined #openstack02:15
*** burris has quit IRC02:40
*** hadrian has quit IRC02:56
*** miclorb has joined #openstack02:57
*** vvuksan has quit IRC02:58
*** jsgotangco has quit IRC03:06
*** anotherjesse has quit IRC03:07
*** jakedahn has quit IRC03:19
*** littleidea has quit IRC03:30
*** miclorb has quit IRC03:37
*** MarkAtwood has quit IRC04:01
RobertLaptopHi I am running the a nova install script but not understanding the question "Total amount of usable IPs for ALL projects" a quick google search show numbers from 8 to 64?  I would have thought that number would have been in the 1000's04:03
*** miclorb has joined #openstack04:11
kopheycupi believe it's asking for the /cidr04:21
dsockwellRobertLaptop: what exactly are you trying?04:22
dsockwellwhen I set up the networks, I had to specify the IP range for all VMs, (CIDR notation), the number of vlans to create, and the size of each vlan04:23
*** kashyapc has joined #openstack04:27
*** bird has joined #openstack04:30
RobertLaptopIt was part of an install script04:34
dsockwellwhich install script?04:35
RobertLaptopI have moved on to tiring this one
dsockwellI found the best write-up at
*** kopheycup has quit IRC05:31
*** localhost has quit IRC05:39
*** localhost has joined #openstack05:40
*** HugoKuo has joined #openstack05:57
*** viirya has joined #openstack06:02
*** localhost has quit IRC06:15
*** localhost has joined #openstack06:15
*** deepheat has quit IRC06:51
*** burris has joined #openstack06:54
*** guigui1 has joined #openstack07:06
*** Cybo has joined #openstack07:13
*** acmurthy has joined #openstack07:16
*** adiantum has quit IRC07:18
*** deepheat has joined #openstack07:20
berendtat the moment quotas can only be set through flags ( what's the status about implementing real quotas?07:20
*** gasbakid has joined #openstack07:22
*** photron has joined #openstack07:23
*** MarkAtwood has joined #openstack07:24
uvirtbotNew bug: #722477 in nova "iptables-restored failed " [Undecided,New]
*** ccustine has quit IRC07:30
*** adiantum has joined #openstack07:31
sorenttx: o/07:32
ttxsoren: p/07:39
*** stewart has joined #openstack07:40
*** CloudChris has joined #openstack07:40
*** naehring has joined #openstack07:41
*** acmurthy has quit IRC07:45
*** deepheat has quit IRC07:51
*** rcc has joined #openstack07:57
*** MarkAtwood has quit IRC07:57
*** MikeZ has joined #openstack08:01
*** MikeZ1 has quit IRC08:03
ttxsoren: I tried to build a bexar update candidate in PPA for testing, but it ftbfs:
ttxsoren: I just copied the packaging from the release PPA. Any clue ?08:07
sorenttx: Have you specified the build-depends PPA as a dep?08:11
ttxsoren: hm, no, I only specified the release PPA as a dep.08:11
* ttx adds08:13
*** Nacx has joined #openstack08:20
*** photron has quit IRC08:26
*** deepheat has joined #openstack08:28
*** fabiand_ has joined #openstack08:29
*** mgoldmann has joined #openstack08:33
*** calavera has joined #openstack08:38
*** skiold has joined #openstack08:39
*** MarcMorata has joined #openstack08:39
ttxsoren: success !08:41
*** _0x44_ is now known as _0x4408:53
*** ChanServ sets mode: +v _0x4408:53
*** ramkrsna has joined #openstack09:01
*** Cybo has quit IRC09:01
*** eikke has joined #openstack09:01
sorenttx: \o/09:03
*** DigitalFlux has joined #openstack09:05
*** gasbakid_ has joined #openstack09:12
DigitalFluxMorning ..09:14
*** gasbakid has quit IRC09:15
*** MarcMorata has quit IRC09:16
*** MarcMorata has joined #openstack09:18
*** MarcMorata has quit IRC09:19
*** MarcMorata has joined #openstack09:19
*** miclorb has quit IRC09:28
*** eikke has quit IRC09:47
DigitalFluxttx: around ?09:51
ttxDigitalFlux: yes09:51
DigitalFluxI had a successful KVM+Glance+novatools installation :)10:02
DigitalFluxttx: some stuff are broken of course, but it was worth it10:03
DigitalFluxttx: one weird issue, is that the compute node is trying to fetch the image from a glance server that is located on the same node10:03
DigitalFluxttx: not from the nova-api node10:03
DigitalFluxttx: should i put some config parameter to point to the glance server IP or something ?10:04
vishyberendt: huh? you can set quotas per project10:04
ttxDigitalFlux: there is a --glance_host parameter10:04
DigitalFluxttx: Cool, will try that right away10:05
DigitalFluxttx: not in Glance's docs yet, correct ? :)10:05
ttxShould be in nova docs. Once we add all flags :)10:05
DigitalFluxttx: Gr810:05
vishyberendt: nova-manage project quota10:06
DigitalFluxI will document the process10:06
ttxDigitalFlux: cool, thanks !10:06
berendtvishy: yes I found it.. forgot to answer myself here.. ;)10:10
DigitalFluxttx: about that libvirt_xml_template flag, it should go in the compute node's nova.conf, correct ?10:11
ttxyes, but won't hurt if it's in the others.10:12
DigitalFluxttx: Yeah, i am kinda confused of where should the parameters go :), having a single config file is pretty cool but confuses me sometimes when lots of components is distributed across different nodes10:13
uvirtbotNew bug: #722524 in nova "euca-terminate-instances fails with PostgreSQL backend" [Undecided,New]
*** abr has joined #openstack10:23
*** BK_man has joined #openstack10:24
*** BK_man has quit IRC10:26
*** adiantum has quit IRC10:31
*** miclorb_ has joined #openstack10:35
berendtwhat's wrong with
berendtI wrote a message in #launchpad.. seems that it's not possible to access +merge/xxxxx at the moment10:45
ttx<wgrant> Fix in progress.10:48
*** metoikos has joined #openstack10:54
berendtseems to be up again...10:55
*** uksysadmin has joined #openstack10:56
*** uksysadmin has joined #openstack10:58
uksysadminHi - quick q (aren't they all?!): What's the state of development of passing user-data to an instance?10:59
sorenuksysadmin: That10:59
sorenhas worked for ages.10:59
sorenI use it constantly.10:59
sorenuksysadmin: eth0 is the default for public_interface now, by the way.10:59
uksysadminthat's good news - just been getting conflicting info on this11:00
uksysadminyeah - noticed that default now11:00
*** adjohn has quit IRC11:01
uksysadminin fact just had a conversation with someone on this here on the user-data stuff.  I'll have a play and no doubt (lucky you!) I'll possibly have more questions - its possible some wires have been crossed at some point11:02
ttxuksysadmin: user-data relies on metadata service, which can be tricky to set up in some networking modes11:04
ttxIIRC VlanManager is the only mode that sets the forwarding up for you.11:05
*** miclorb_ has quit IRC11:06
uksysadminthat could explain an email from Rightscale - they told our CTO that they're waiting on user-data injection and "Unfortunate as we are reliant on the Open Stack Community we have no ETA at this point, but we will keep you posted."11:06
uksysadminmust say - I had a go - didn't get it working but it was during a time where I had other networking issues occuring.  now I've a repeatable build and set up procedure I was going to demo this, mimicing what Rightscale were doing with their EC2 instances.11:07
ttxuksysadmin: userdata injection has been working forever... it's just atht metadata service requests need to be properly routed to the api server, which requires manual networking setup in some network modes11:07
ttxsomething we may need to simplify somehow.11:08
sorenWhat we really need to do is to stop forwarding that stuff to the API server and keep it in the compute worker.11:08
uksysadmin*nods* - I can understand that and see that simplifying things if that could be done11:10
ttxsoren: sounds like a good project to nail while traveling to Berlin.11:10
ttx(just sayin)11:10
sorenI'll be driving.11:10
sorenOf course I could tell my wife what to type on my laptop.11:11
ttxbah, you can do it single-handed.11:11
berendtsoren: what are you doing in berlin?11:14
sorenberendt: holiday11:14
uvirtbotNew bug: #722543 in nova "nova.pot out of sync with source code in trunk" [Undecided,New]
*** uksysadmin_ has joined #openstack11:22
*** uksysadmin has quit IRC11:23
*** gasbakid_ has quit IRC11:23
*** h0cin has joined #openstack11:25
*** uksysadmin_ has quit IRC11:25
*** uksysadmin has joined #openstack11:26
*** rds__ has joined #openstack11:27
uvirtbotNew bug: #722554 in nova "bin/nova-manage hides disabled services" [Undecided,New]
* soren dances around11:56
sorenTyping while dancing.... Not very easy.11:56
sorenThat's a delightful amount of green.11:56
*** olivier_ has quit IRC12:01
*** olivier_ has joined #openstack12:01
berendtis there a way to reload modified python files automatically in the running services or do i have to restart the services using the modified files every time?12:02
Failbaitrberendt: normal python usage dictates that you do not want to reload python modules on the fly while there are objects in the program using the old files12:05
sorenberendt: Yeah, what Failbaitr said. Restart is the way to go.12:05
berendtI forgot it in 50% after doing some modifications and I'm always surprised why my changes are not working correctly :(12:06
*** GasbaKid has joined #openstack12:13
DigitalFluxAny hints on where can i find an ubuntu desktop image that i can upload to glance and run it with novatools ?12:13
*** rcc has quit IRC12:14
*** matclayton has joined #openstack12:14
*** rcc has joined #openstack12:15
*** rcc has joined #openstack12:16
*** ovidwu has quit IRC12:16
*** metoikos has quit IRC12:20
*** uksysadmin has quit IRC12:22
*** ctennis has quit IRC12:28
*** kashyapc has quit IRC12:29
*** rcc has quit IRC12:36
*** drico_ has quit IRC12:42
*** ctennis has joined #openstack12:43
*** ctennis has joined #openstack12:43
*** kashyapc has joined #openstack12:46
*** GasbaKid has quit IRC12:56
*** zul has joined #openstack12:56
*** DigitalFlux has quit IRC13:00
*** DigitalFlux has joined #openstack13:00
*** DigitalFlux has quit IRC13:01
*** adiantum has joined #openstack13:10
*** chmouelb has joined #openstack13:12
*** rds__ has quit IRC13:12
*** rds__ has joined #openstack13:16
sorenMan... Moving this channel to window 5 in my irssi is going to save me a lot of typing.13:28
soren(It used to be at /win 61)13:28
chmouelbalt-5 FTW13:29
sorenOh, right.13:29
sorenI was just thinking about the time I'd save by typing "/win 5" instead of "/win 61" :)13:30
sorenOr, actully, the time I'd save when typing Alt-A and not having to cycle through windows 6-60.13:30
sorenI could watch this all day:
* ttx wonders what yes/no means13:33
ttxah, userdata13:33
ttxnova-core: please give some review love to
*** mustfeed has joined #openstack13:34
zulsoren: is there a "/win 666" :)13:34
ttxThis is blocking progress on 2011.1.1 ^13:34
sorenzul: Not yet.13:36
sorenzul: There's a /win 221.13:37
*** mustfeed has quit IRC13:41
*** eikke has joined #openstack13:48
*** ovidwu has joined #openstack13:58
*** sebastianstadil has joined #openstack13:59
*** Ryan_Lane has joined #openstack14:06
*** dillon-w has joined #openstack14:07
*** JoeGana has joined #openstack14:08
*** hadrian has joined #openstack14:10
dillon-wHi, guys.  I've some questoins about CloudFile Java binding14:10
*** GasbaKid has joined #openstack14:10
*** gondoi has joined #openstack14:10
jaypipesdillon-w: may want to ping creiht, redbo, notmyname. they all work on Swift.14:11
jaypipesdillon-w: and gholt :)14:11
dillon-wthanks jay!14:11
dillon-wcreiht: hello14:12
*** Ryan_Lane has quit IRC14:13
chmouelbdillon-w: what kind of question about the Java binding ?14:21
*** imsplitbit has joined #openstack14:29
*** JoeGana has quit IRC14:31
*** aliguori has joined #openstack14:34
imsplitbitwhat a glorious morning to be alive!14:49
iRTermiteimsplitbit: indeed!14:49
*** herki has joined #openstack14:51
*** littleidea has joined #openstack14:52
dillon-wchmouelb: hi, i'd like to know what exactly createPath/createFullPath of FileClient class is doing.14:53
dillon-wchmouelb: it's said swift/cloudfile has no directory/folder concept, it is flat. then what is the difference between FileClient.createFullPath("foo/bar") + FileClient.storeObjectAs("Container", "a.jpg", "image/jpeg") and FileClient.storeObjectAs("Container", "a.jpg", "image/jpeg", "foo/bar/a.jpg")14:57
chmoueldillon-w:  it does what notmyname explained here
dillon-wchmouel: thanks. let me read that first14:59
dillon-wchmouel: i don't quite understand why "As a final step, I need to upload ¡°directory marker¡± files. These are empty (zero-sized) files with a content-type of ¡°application/directory¡±"15:05
*** uksysadmin has joined #openstack15:05
*** hggdh has quit IRC15:06
*** hggdh has joined #openstack15:07
dillon-wchmouel: since the object name itself has already contained extra characters for pseduo directory15:08
notmynamedillon-w: the directory markers are no longer needed in swift15:08
chmouelit's probably best to talk about this in #cloudfiles since it will be better belong there but basically this is marker for listing15:08
chmouelor that :)15:09
notmynamedirectory markers are for the path query. using the delimiter (a more generic form of the path query), you don't need directory marker objects15:09
*** uksysadmin has quit IRC15:10
dillon-wnotmyname: does that mean if i don't use directory marker object, then I won't be able to do path query?15:11
*** dendrobates is now known as dendro-afk15:12
notmynamedillon-w: yes, but I wouldn't recommend using the path at all :-)15:12
*** CloudChris has left #openstack15:15
*** uksysadmin has joined #openstack15:15
dillon-wnotmyname: why not?  What I need is to put thousands of files like "People/000N/000N/a.jpg" into Swift.  I'd like to keep the directory information.15:16
uksysadmin(apologies if repeating - irc session cocked up): in a nice and simple, was working, vlan network mode, vlan100 is created automagically on eth0. How do I change this to use another interface? I've no mention of eth0 in /etc/nova/nova.conf15:17
notmynamebut with the delimiter query ("GET /v1/<your acct>/<your contianer>/?delimiter=/") you will be able to traverse the pseudo-dirs15:17
notmynamecombine the delimiter with prefix and you can walk as deep as you want into the hierarchy15:17
notmynameapologies, but I need to go. creiht, redbo, gholt, or other swift devs should be able to answer other questions you have about path vs delimiter15:18
dillon-wnotmyname: thank you for the hints.  see you15:19
*** ramkrsna has quit IRC15:21
*** dragondm has joined #openstack15:25
redboI don't completely agree with that.  prefix/delim queries don't completely negate the need for directory markers.  There's no other way to indicate empty directories.15:26
*** dragondm has quit IRC15:27
dillon-wredbo: can i use prefix/delim style object for _none_empty directory and only use directory markers for empty directories?15:28
redboprobably, but I don't really wanna try explaining that to people :)15:29
dillon-wredbo: so hybrid/mixed style is not perferred? at least in the same container?15:30
ttxdevcamcar, eday, jk0, mtaylor, _0x44, termie, xtoddx, vishy: if one (or more) of you could review and approve it if ok, that would be great. It's blocking progress on 2011.1.1 (we want the fix to land in Cactus first)15:31
*** jfluhmann has quit IRC15:32
_0x44ttx: Do I need to know Czech/etc?15:32
jaypipes_0x44: no, but you do need to know Jive.15:32
jk0I will look it over15:33
*** that__guy has joined #openstack15:33
_0x44jaypipes: Mama didn't raise no dummies! I dug her rap!15:33
ttxsounds a bit like a leap of faith anyway, since only jaypiipes and a few others understand how it's supposed to work once merged15:33
*** jfluhmann has joined #openstack15:33
gholtdillon-w: That is what the 'st' utility included with Swift does actually. dir markers just for empty dirs.15:33
redbodillon-w: if your app can deal with squashing together a delim rolled up subdir and directory marker, it'll work fine.  But I don't know what the preferred method of handling that should be.15:33
redboWe'd thought about making "path" queries magical so they'd fill in any missing directory markers from object name prefixes, but that never went anywhere.15:35
ttxsoren: is our packaging supposed to pick up the translations automatically ?15:35
jaypipes_0x44: well played.15:36
jaypipesttx: no.15:37
ttx_0x44: feel free to switch the "approved" bit as well, so that it actually gets merged15:37
_0x44ttx: Done15:37
ttxjaypipes: so we also have to do an extra pass in packaging ?15:37
jaypipesttx: so, what happens is that when someone commits to trunk, Launchpad builds a new set of .po files from the base nova.pot template. These .po files are the language-specific non-compiled catalogs. The thing we have to do manually before final release is pull from lp:~nova-core/nova/translations, call python build_i18n, and make sure all catalogs are built properly...15:39
dillon-wredbo: actually i'm curious how to _just_ use path to simulate directories.  Say a file name x.jpg, i'd like to put it into "foo/boo" directory, how do i do that with only path?15:39
*** rnirmal has joined #openstack15:40
*** msassak has joined #openstack15:40
ttxjaypipes: hrm, sounds a bit like something you really don't want to do at the very last moment15:41
sorenttx: Not a clue.15:41
* ttx ponders a stringfreeze at gamma or rc15:41
jaypipesttx: we *have* to do it at the last moment, because people add i18n'd strings all the time..15:41
redboThe idea is you'd make an empty file "foo" and an empty file "foo/boo" and then upload your file named "foo/boo/x.jpg".  Then you can do a query with "path=foo" and it'll give you the logical equivalent of a directory listing of foo, so you'll just see boo.15:42
ttxjaypipes: that leaves very little time to catch problems in that toolchain15:42
redboso you can do a query with "path=" and you'd see just foo, then with "path=foo" and just see boo, then "path=foo/boo" and just see x.jpg15:43
jaypipesttx: There's nothing I can do about that.15:43
ttxjaypipes: we can do it at a reasonably-late moment.15:43
ttxjaypipes: i.e. sufficently late to get lots of translations, but not too late to catch errors15:44
dillon-wredbo: and that is not achievable without path?15:44
jaypipesttx: we can do it on every commit if we automate it, but I'm not sure of the effects of doing that... building all the message catalogs, etc...15:44
jaypipesttx: the reason is because then people will constantly be downloading large compiled catalogs.15:44
jaypipesttx: when really, those catalogs only need to go into the packaging, not the source trees.15:45
jaypipesttx: compiled catalogs are the .mo files. Binary files.15:45
redbodillon-w: It is possible, all you have to do is upload "foo/boo/x.jpg" and you can use prefix/delim to walk the tree to x.jpg, and it sort of automagically figures out that there are "foo" and "foo/boo" directories.15:46
ttxjaypipes: i'm a bit lost. Why would we want to ship .mo files in the release tarball ?15:47
ttxjaypipes: looks like this "python build_i18n" should be done in packaging instead ?15:48
jaypipesttx: in other words, the process, in a "flowchart" is this:   commit code -> create pot template (LP) -> translators translate the POT -> construct the PO files from the translator's strings (LP) -> export PO files into lp:~nova-core/nova/translations (LP) -> pull lp:~nova-core/nova/translations and run python build_i18n, which produces MO files -> put MO files into release distribution...15:48
dillon-wredbo: how to 'use prefix/delim to walk the tree'?15:49
jaypipesttx: exactly. running python build_i18n should be done for packaging.15:49
jk0ttx: you want me to flip that to approved?15:49
ttxjk0: I think it's done now15:49
jaypipesttx: and that's what we need to do for Bexar, but you wanted to try this out on Cactus first...15:49
jk0ttx: ah, looks like unit tests failed, so it went back15:49
jk0per hudson15:49
jk0authors file is out of date15:50
ttxjaypipes: so the end of your process would be:15:50
ttx-> pull lp:~nova-core/nova/translations -> release tarball with those POs -> take tarball in packaging and run python build_i18n, which produces MO files -> generate binary packages including the MO files15:51
jaypipessoren: why, since it's *my* branch (branched originally from ~nova-core/nova/translations) is Hudson saying that the Author is "nova-core"? Grr.15:51
redbodillon-w: you start with "prefix=&delim=/" and you'll get a subdir type record named "foo/", then you do a query with "prefix=foo/&delim=/" and you get a subdir record named "foo/boo/", then do "prefix=foo/boo/&delim=/" and you'll get an object record for "foo/boo/x.jpg"15:51
jaypipesttx: correct. but someone has to remember that the POs from translations don't magically get into trunk. :) They magically get created by Launchpad from the translations on Rosetta, but those PO files don't magically get committed and merged intro trunk. Someone still has to do that.15:52
ttxjaypipes: sure. That can be part of the milestone release process, or autocommitted daily or at every commit15:53
ttxat this point it will be "part of the milestone release process" and I'll ensure it's done.15:54
*** hub_cap has joined #openstack15:54
*** Ephur has joined #openstack15:55
dillon-wredbo: i see. very helpful, thanks!15:56
ttxjaypipes: actually LP is saying one of the authors is nova-core... which is the LPid under which LP translations commits po changes15:58
jaypipesttx: ok.16:00
ttxlooks like we'll need to add it or specialcase it in the test.16:00
jaypipesI'll do it.16:00
*** guigui1 has quit IRC16:03
*** vvuksan has joined #openstack16:04
*** msassak has quit IRC16:05
*** aliguori has quit IRC16:06
*** msassak has joined #openstack16:07
*** naehring has quit IRC16:09
*** msassak_ has joined #openstack16:10
*** mdomsch has quit IRC16:11
*** msassak has quit IRC16:12
*** mgoldmann has quit IRC16:13
*** bird has quit IRC16:17
*** j05h has joined #openstack16:19
*** chmouelb has quit IRC16:22
*** dendro-afk is now known as dendrobates16:25
*** dillon-w has quit IRC16:30
*** kashyapc has quit IRC16:33
*** msassak_ has quit IRC16:33
*** annegentle has left #openstack16:34
*** annegentle has joined #openstack16:34
*** mustfsk has joined #openstack16:34
mustfskhi, can someone help me configuring swift on a single system ? got some probs with configuring the rings16:35
that__guyAnyone up for answering some questions regarding multiple nova-volume servers?16:39
creihtmustfsk, that__guy: ask away, and someone will hopefully answer when time allows :)16:40
*** dendrobates is now known as dendro-afk16:41
*** dendro-afk is now known as dendrobates16:42
*** grapex has joined #openstack16:43
*** johnpur has joined #openstack16:43
*** ChanServ sets mode: +v johnpur16:43
*** skiold has quit IRC16:47
*** uksysadmin has quit IRC16:48
*** fabiand_ has quit IRC16:50
*** kashyapc has joined #openstack16:52
*** GasbaKid has quit IRC16:53
*** MarkAtwood has joined #openstack16:59
*** alekibango has quit IRC17:01
that__guycreiht: cool.  so... we want to split off storage/retrieval of images, and have an installation on a storage node.  A 'nova-manage service list' shows the node running nova-volume service, but we're not sure how to get the publishing tools to use this new node17:10
*** MarkAtwood has quit IRC17:10
that__guyI saw in the nova-volume code there's a 'storage-availability-zone' flag... will this have to be set in the storage node's nova.conf?17:11
*** calavera has quit IRC17:12
annegentlethat__guy: nova-objectstore is the simple VM image store/retrieve system for nova... nova-volume is for extra block level storage using LVM (
*** reldan has joined #openstack17:15
annegentlethat__guy: you can use the Glance project ( for storage/retrieval of images17:15
that__guyoh... well that explains it then :)17:15
that__guyis glance a better option for our use case?17:16
annegentlethat__guy: almost certainly17:17
*** MarkAtwood has joined #openstack17:17
*** jheiss has joined #openstack17:18
annegentlethat__guy: jaypipes can help better than I with connecting Glance to Nova17:19
*** imsplitbit has quit IRC17:19
*** imsplitbit has joined #openstack17:19
jaypipesthat__guy: what are you looking to do?17:20
that__guyI'll look in to that then.  Is there a preferred version of glance for interfacing with other services running trunk version bzr655?17:20
that__guyjaypipes: I want to get all image storage/retrieval off of the head node17:20
that__guyjaypipes: we have some dedicated storage nodes with more physical storage and network bandwidth17:21
jaypipesthat__guy: easy enough to set up Glance and direct image requests to it... are you wanting to run Glance separately from Nova?17:24
jaypipesthat__guy: that said, there are a number of severe limitations in Glance right's still a very new project.17:25
that__guyjaypipes: not seperately; just on a node other than the one running nova-api, nova-scheduler, etc.17:25
jaypipesthat__guy: k. you using the OS or the EC2 api?17:26
that__guyjaypipes: both17:26
jaypipesthat__guy: hmm, ok then :)17:26
*** justinsb has joined #openstack17:27
jaypipesthat__guy: alright, so you need to download the latest trunks of both Glance and Nova, first.17:27
jaypipesthat__guy: there are some bugs that have been fixed recently re: integrating between Nova and Glance.17:27
that__guyjaypipes: can't upgrade nova... 655 is the version we're runnign with17:27
that__guyjaypipes: we had to patch it a bit to get it working17:27
jaypipesthat__guy: hmm... not sure if I'll be able to help you then. You would need to patch in the fixes to the image service layer that are in Nova's trunk..17:28
that__guyjaypipes: would running nova-objectstore on the storage node work for the time being?17:29
jaypipesthat__guy: yep. that is what I would recommend.17:29
*** abecc has joined #openstack17:29
jaypipesthat__guy: you running XenServer on the OpenStack side?17:29
that__guyjaypipes: that will work then.  I'll give glance a look see when our dev cluster is up17:30
that__guyjaypipes: thanks!17:30
jaypipesthat__guy: sounds good.17:30
jaypipesthat__guy: thank YOU!17:30
*** herki has quit IRC17:36
*** GasbaKid has joined #openstack17:39
*** littleidea has quit IRC17:44
*** mdomsch has joined #openstack17:50
justinsbI'd like to love Bazaar, but I'm struggling .... is there a fast way to branch? (it's downloading 20MB now, even when I already have a similar branch locally)17:51
*** fabiand_ has joined #openstack17:53
*** littleidea has joined #openstack17:57
*** littleidea has quit IRC18:00
*** littleidea has joined #openstack18:01
*** littleidea has quit IRC18:03
*** eikke has quit IRC18:04
*** dirakx has quit IRC18:10
jaypipesmtaylor: around?18:11
jaypipesjustinsb: what are you executing?18:11
*** nelson has quit IRC18:12
justinsbjaypipes: e.g. bzr branch lp:~justin-fathomdb/nova/justinsb-openstack-api-keys18:12
jaypipesjustinsb: why are you branching from launchpad? don't you have that branch locally?18:12
jaypipessoren: around?18:13
justinsbjaypipes: I have several computers.  One branches & runs unit tests while I code on another.  But now I'm trying to see if I can move off git-bzr...18:13
justinsbjaypipes: Is the answer to pull once locally and then essentially keep track of everything by hand to avoid pulling over the net?18:14
*** hub_cap has quit IRC18:14
jaypipesjustinsb: takes about 14s to branch from Launchpad for me:18:15
jaypipesjpipes@serialcoder:~/repos/nova$ time bzr branch lp:~justin-fathomdb/nova/justinsb-openstack-api-keys18:15
jaypipesBranched 703 revision(s).18:15
jaypipesjustinsb: if this is on the machine running unit tests, is that 14seconds going to be an issue?18:15
jaypipesjustinsb: meaning, if this is work that an automated testing script is doing, why care all that much?18:15
jaypipesjustinsb: and FYI, about 80% of that time is because you're pulling the code from somewhere in the UK versus a Rackspace data center in Texas housing GitHub...18:16
justinsbjaypipes: Well, for me: justinsb@jsb-rackspace-desktop:~/bazaar/nova/t2$  time bzr branch lp:~justin-fathomdb/nova/justinsb-openstack-api-keys18:17
justinsbBranched 703 revision(s).18:17
jaypipesjustinsb: I can't help your Internet speeds....18:17
justinsbjaypipes: Do you have to do this every time you want to do a branch though?  Or do you keep a copy of lp:nova locally and then branch that?18:18
justinsbjaypipes: Is there a 'pro bazaar' guide anywhere?18:18
jaypipesjustinsb: I keep a trunk locally that I do a bzr pull on every day.18:18
creihtjustinsb: I keep an updated trunk local that I branch from18:18
jaypipesjustinsb: and when I want to branch, I branch from there, not lp. But, that's not going to fix your issue on a separate box, since you aren't developing on that box, you are just grabbing code from Launchpad and running tests...18:19
mtaylorjaypipes: whazzup?18:19
jaypipesjustinsb: what you could do I suppose is have that other box bzr branch from your development box...18:19
mtaylorjaypipes: I think you might want to also communicate about shared repos to justinsb18:20
jaypipesmtaylor: still getting stuck on this: I have NO idea wtf is going on. I can't reproduce this on my box, either in a virtualenv or running -N. I guess I need to log into the Hudson box to see what the heck is up.18:20
jaypipesmtaylor: he was using git-bzr-ng so I didn't bother at first.18:20
justinsbjaypipes, creiht: OK, I'll try the local branch trick.  I just wanted to check I wasn't doing anything completely stupid!18:20
mtaylorjaypipes: oh, nevermind then18:20
jaypipesjustinsb: nope, has to do with what bzr branch is doing behind the scenes. If you're branching a URL in the UK, going to be much slower than branching a local URL...18:21
justinsbmtaylor, jaypipes: I'd rather use bazaar if I can :-)18:21
jaypipesjustinsb: there's also the issue of making sure you have done a bzr init-repository locally... not sure if you have or not.18:21
mtaylorjustinsb: then I believe the answer to your earlier question of "Is the answer to pull once locally and then essentially keep track of everything by hand to avoid pulling over the net?"18:22
mtaylorjustinsb: is that you need to be doing your branching work inside of a dir that you have run bzr init-repo on18:22
*** acmurthy has joined #openstack18:22
jaypipesmtaylor: but he's using >1 machine, one for dev, one that simply bzr branches and executes the tests...18:22
justinsbjaypipes: I thought I had... I create a directory bazaar, do a "bzr init ." in it, and then do bzr branch lp:....  But each bzr branch seems to pull everything every time still :-(18:22
mtaylorjustinsb: which will cause that dir to store and keep the revision history, so that you aren't pulling from scratch each time18:23
jaypipesjustinsb: bzr init != bzr init-repository :)18:23
justinsbjaypipes: Ah!18:23
mtaylorjustinsb: bzr init and bzr init-repo are not the same18:23
justinsbmtaylor, jaypipes: Off to try bzr init-repo and learn to love bazaar :-)18:23
jaypipesjustinsb: :) one initializes a repository, the other initializes source control in a directory18:23
*** fabiand_ has quit IRC18:24
jaypipesjustinsb: not about learning to love it. Heck, I don't like certain things about bzr that I love about git. Just about making sure you're working with bzr in the most efficient way you can, given the obvious limitations and efficiency issues with Launchpad not being nearly as quick as GitHub for a variety of reasons (the biggest, of course, being network latency)18:24
justinsbjaypipes: Well, I think I'll love it a lot more if it doesn't take two minutes to branch, so thank you!18:25
jaypipesjustinsb: heh, indeed. :)18:25
jaypipesmtaylor: so, any idea on that issue of glance-control apparently not "finding" glance.version in the tests? I'm a bit stumped :(18:26
jaypipesmtaylor: lemme rephrase that. I'm a LOT stumped.18:26
mtaylorjaypipes: no - it's weird- but let me look at it again18:27
justinsbjaypipes, mtaylor: time bzr branch lp:~justin-fathomdb/nova/justinsb-openstack-api-volumes18:27
justinsbBranched 702 revision(s).18:27
justinsb.  Thank you so much!!18:27
jaypipesjustinsb: very welcome :)18:27
mtaylorjustinsb: w00t18:27
*** reldan has quit IRC18:28
jaypipesmtaylor: I think it MUST be a path issue, but I'm unsure how to check it...18:28
mtaylorjustinsb: now you get to have fun learning all the places where git and bzr use similar words to mean completely different things. it's eversomuch fun18:28
jaypipesmtaylor: and I'm wondering why if it finds glance-control, that it wouldn't find glance.version...etc.18:28
jaypipesjustinsb: indeed, and all the loveliness in the fact that bzr uses the convention: bzr action SOURCE DEST, which git uses the opposite: git action DEST SOURCE.18:29
justinsbmtaylor, jaypipes: Maybe I'll install a GUI to shield me from the command line differences :-)18:29
* jaypipes wonders how many times that has tripped him up..18:29
jaypipesjustinsb: I've heard the bzr gui is cool, but I've never used it... same for git. :(18:30
*** Ryan_Lane has joined #openstack18:30
*** rds__ has quit IRC18:34
*** reldan has joined #openstack18:38
*** pvo has joined #openstack18:43
*** vvuksan has quit IRC18:47
*** vvuksan has joined #openstack18:50
*** fabiand_ has joined #openstack18:50
jaypipes_cerberus_: thx :)18:51
jaypipesOMG, I am getting so friggin sick of this authors check.18:53
jaypipeshow the heck does this:18:54
jaypipes                    if contributor == 'nova-core':18:54
jaypipes                        pass18:54
jaypipesnot ignore nova-core in the contributor list? :(18:54
jaypipesgah, I'm an idiot. continue, not pass.18:55
* jaypipes still wondering why locally did not fail... grr.18:55
jaypipescreiht: no shit, right?18:56
jaypipeshow long have I been programming? meh.18:56
creihtwe all do it :)18:56
jaypipescreiht: heh, perhaps. me more than others lately.18:56
jaypipesmtaylor: any ideas on that glance.version thing? sorry to bother you, it's just a blocker for more work in Glance..18:57
jarrodok i am going the xen route this morning18:57
mtaylorjaypipes: sorry man - haven't looked at it yet this morning, been heads down in something else. promise I will look at it though18:57
*** vvuksan has quit IRC18:58
jaypipesmtaylor: cool, thx brother.18:58
jaypipesdabo: wanna re-review sirp_ 's branch here?
*** acmurthy has quit IRC18:59
jk0anyone recall what the fix was for the recent evenlet problem?19:00
jk0was it just upgrading?19:01
kpepplejk0: i thought it was a patched eventlet ... soren did it19:01
jk0ah, hmm19:01
jk0kpepple: you recall what bug it was?19:02
*** rnirmal has quit IRC19:02
dendrobatesjk0: I can find it19:02
kpepplejk0: patch here --
kpepplejk0: can't remember the bug id19:03
jk0cool, thanks19:03
*** fabiand_ has quit IRC19:03
*** dendrobates is now known as dendro-afk19:05
jaypipesjk0: yeah, install eventlet from the PPA:
*** MarcMorata has quit IRC19:05
jk0jaypipes: excellent, I was just about to ask :)19:05
*** acmurthy has joined #openstack19:06
*** alekibango has joined #openstack19:07
RobertLaptopQuestion Nova Compute Only nodes do they talk directly to the public IP addresses or do they get routed though the the Nova Network Component?19:07
RobertLaptopI am tiring to write up internal doc in on possibly using Nova19:08
jaypipesjk0: np :)19:08
jaypipesRobertLaptop: nova compute nodes talk with other components in the Nova fabric via AMQP (message queue). The IP addresses that *instances* are assigned is determined by the network manager, if that's what you are asking?19:10
jaypipesmtaylor: python-distutils-extra now needs to be installed on Hudson:  Thanks in advance!19:12
RobertLaptopjaypipes, kinda.  Once they are assigned an IP by the network manger do the node handle requests in an out for that IP or does it go though the network controller.  I have to admit it is the fact that I am using eucalyptus might be showing though.19:12
*** dendro-afk is now known as dendrobates19:13
*** Nacx has quit IRC19:14
jaypipesRobertLaptop: ah, I see. AFAIK (and vishy could correct me if I'm wrong), once an instance is assigned an IP, anyone with the appropriate keypairs can access the instance via that public IP, but communication with the instance for administrative commands may still occur via the Computer worker, which receives its commands via AMQP.19:14
jaypipesvishy, devcamcar, soren: feel free to correct me above ^^19:14
vishyrouted through nova network19:15
vishynova-network is the gateway for all of the vms19:15
dendrobatesjk0: bug 71642719:16
uvirtbotLaunchpad bug 716427 in nova/bexar "RPC concurrency problem" [High,Fix committed]
openstackhudsonProject swift build #204: SUCCESS in 30 sec:
openstackhudsonTarmac: Reverted the wal+index code.19:16
jk0dendrobates: thanks for finding that19:17
jk0couldn't remember the subject name for the life of me19:17
RobertLaptopso jaypipes if I am understanding you correctly if I setup an instance and pointed a public IP to it and opened up say port 80 even if the network controller died the public IP would be accessible?19:17
jaypipesvishy: ? ^^19:18
RobertLaptopThat is one of the things I really hate about eucalyptus if the fact that all network traffic goes though it cluster controller.19:19
vishyRobertLaptop: no19:19
RobertLaptopvishy, ok :(19:20
vishyRobertLaptop: if the network worker dies it will still be accessible, but if the network host goes down it won't19:20
jaypipesvishy: sorry, I don't understand the difference...19:20
*** kbringard has joined #openstack19:21
jaypipeswhat is the "network host"?19:21
RobertLaptopjaypipes, bet me to the question19:21
jaypipesvishy: the machine that has iptables?19:21
*** gregp76 has joined #openstack19:21
dendrobatesjk0: fyi, soren fixed eventlet  in natty, but has not heard back from upstream19:22
vishyjaypipes nova-network is a process19:22
vishyif that process dies/crashes/gets turned off, everything works19:22
*** joearnold has joined #openstack19:22
vishybecause nova-network just controls dnsmasq and iptables19:23
jaypipesvishy: ok. but you are talking about the "network host". do you mean the machine that has had the routing tables altered?19:23
vishybut if the host machine where nova-network is running fails...19:23
jaypipesvishy: sorry if the questions sound retarded. none of this is in the docs..19:23
jaypipesvishy: OK, got it. So, as long as the machine that has had its routing table altered remains up, then the project's network remains up... and when you said above that all network traffic went through nova-network, you were referring to the *host*, not the worker process. Is that correct?19:24
jaypipesvishy: thx. appreciate the insight.19:25
RobertLaptopand to confirm since I am been running a nova node for about 5 hours.  The IP tabes gets changed on the system running nova-compute not the actually system running the instance.19:26
kbringardand along those lines, each network controller can control a project, but you can't run 2 with the intent of being redundant, yea?19:26
* jaypipes notes Gmail's interface just changed... I like the new look of the top navbar..19:26
kbringardsince you can't have 2 default routes, correct?19:27
jaypipesvishy: ^^19:27
*** mahadev has joined #openstack19:29
vishykbringard: correct, some investigation needs to be done into redundancy19:29
vishykbringard: failover is probably the best solution19:29
RobertLaptopI think it would make more since to assign the IP to the instance host and alter the iptables there keeping the pass though down but if the setup is to do it at the network controller level then I need to understand that for planing.19:29
*** brd_from_italy has joined #openstack19:30
vishyRobertLaptop: that doesn't really solve the gateway and dhcp19:30
kbringardthat's one of the main reasons we're looking at Flat/FlatDHCP if we can still VLAN tag (and is one of the main reasons I'm super excited about the Open vSwitch stuff in cactus)19:30
kbringardlet the netops team deal with HSRP or whatever way they want to failover the routers19:31
kbringardand we'll just deal with dataflows19:31
jaypipeskbringard: in Cactus, eh? I didn't know OpenVSwitch was going into Cactus...19:31
kbringardoh, I just saw a blueprint about it...19:31
vishyRobertLaptop: it requires writing code to let an external system control ips19:32
*** retr0h has joined #openstack19:32
*** retr0h has joined #openstack19:32
jaypipeskbringard: hmm. good to know! :)19:32
retr0hephur! :)19:32
RobertLaptopvishy, true that would be a security issue.19:32
kbringardalthough, depending on how much work needs to be done, Diablo may be a better candidate for full integration :-)19:33
EphurHey what's the good word retr0h  :P19:33
kbringardoh jeez, not this guy19:33
retr0hEphur: good to see you in here :)19:33
vishyRobertLaptop: i think we will get there once the plans for network-service get underway19:34
kbringardretr0h: agreed19:34
RobertLaptopvishy, That is good to know.  Biggest issue is I need to get off esxi now so I testing stuff today.  I have no doubt in a Year Open Stack will have everything I want. :)19:35
kbringardvishy and RobertLaptop: if open vsweitch integration becomes a reality, you could skip the iptables stuff altogether and just alter your data path on the fly19:35
*** matclayton has quit IRC19:35
kbringardopenflow ftw19:35
RobertLaptopNow on to my next question.  Is there any documentation on the file structure of compute and/or recommended file system layout for it?19:38
*** GasbaKid has quit IRC19:38
RobertLaptopI can't find any but that doesn't mean someone hasn't written it.19:38
annegentleRobertLaptop: I can't think of a page to point you to either, and I'm practically the librarian around here :)19:41
jaypipeswhy running databases on cloud block storage is not the best idea:
annegentleRobertLaptop: your questions are good ones, we should try to capture them (springing a trap now via a todo)19:42
*** mustfsk has quit IRC19:42
*** ctennis has quit IRC19:47
*** pvo has quit IRC19:49
*** ctennis has joined #openstack19:50
*** ctennis has joined #openstack19:50
*** Jbain has quit IRC19:52
creihtjaypipes: well the first problem is that he is running mysql :P19:52
jaypipescreiht: ah, we're going to start a flamewar, eh? ;)19:53
* creiht nah just making a quick jab and then ducking away :)19:53
jaypipescreiht: :)19:53
RobertLaptopannegentle, were does one insert those todo?19:53
RobertLaptopAs lanchpad bugs?19:54
*** gregp76 has quit IRC19:54
jaypipescreiht: just fyi, the author there is a PG contributor :)19:54
*** gregp76 has joined #openstack19:55
*** gregp76 has quit IRC19:58
*** h0cin has quit IRC20:02
*** dubsquared has joined #openstack20:03
dubsquaredGreetings, openstack!20:03
dubsquaredTrying to create a network, and ran into this:  any ideas?20:03
*** maplebed has joined #openstack20:04
annegentleRobertLaptop: you can create a bug in Launchpad and say it affects the openstack-manuals project20:05
RobertLaptopannegentle, ok.20:05
annegentleRobertLaptop: I've also started a quick wiki page at
RobertLaptopIs one of those to combine all 3 wiki's into one :)20:06
openstackhudsonProject nova build #545: SUCCESS in 1 min 37 sec:
openstackhudsonTarmac: introducing a new flag timeout_nbd for manually setting the time in seconds for waiting for an upcoming NBD device20:07
annegentleRobertLaptop: heh. I'm about to put in a merge request to pare down nova.openstack to avoid such confusion20:10
annegentleRobertLaptop: is that a wiki to you too? Just curious.20:13
*** maplebed has quit IRC20:15
RobertLaptopWell I have spent most of my time in wiki.openstack but yea looks like a wiki to me.20:15
RobertLaptopannegentle, taking a closer look I guess it isn't a wiki.  It looks like some kind of generated system.20:21
annegentleRobertLaptop: yeah, if you want to write for it, just hack at the .rst files in the /doc/source directory of the source code. RST is similar to wikitext, and the build creates the HTML sites.20:24
annegentleRobertLaptop: it's wiki-like in that the RST text is super simple, and it's collaboratively authored within the source code. Sphinx is what builds it.20:24
*** eikke has joined #openstack20:29
*** reldan has quit IRC20:32
jaypipesvishy: those are great articles on the Anso site. I encourage y'all to add your article feed to the OpenStack planet so others can read them. :)20:35
jaypipesdubsquared: just the man I was looking for...20:35
dubsquaredut oh20:35
dubsquaredi didn't do it20:35
uvirtbotLaunchpad bug 710882 in glance "cloudservers boot fails on ubuntu/xenserver" [Undecided,New]20:35
jaypipesdubsquared: wondering any progress on this and whether it has to do with Nova, not Glance?20:36
jaypipesdubsquared: you are Wayne, right? :)20:36
dubsquaredjaypipes: who?20:36
dubsquaredyeah :D20:37
dubsquaredi got that fixed....20:37
jaypipesdubsquared: damn you. I was almost *really* embarrassed...20:37
dubsquaredforgot that was open20:37
jaypipesdubsquared: ah, cool, feel free to mark that closed then...20:37
*** adiantum has quit IRC20:37
dubsquaredthanks for follow-up20:39
dubsquaredill look for the solution and post the close update20:40
dubsquaredforgot what i did off-hand :/20:40
*** herki has joined #openstack20:40
*** msassak has joined #openstack20:43
*** joearnold has quit IRC20:51
*** joearnold has joined #openstack20:52
sorenjaypipes: It's not that "nova-core" is *the* author. It's *an* author. Every single commit is considered in that unit test, not just the top one.20:58
*** mgoldmann has joined #openstack20:58
sorenjaypipes: I'm around now. For a while, at least.20:59
*** hggdh has quit IRC21:00
*** hggdh has joined #openstack21:00
sorendendrobates: I heard back from upstream now. They included the patch (but haven't made a release yet).21:02
*** ctennis has quit IRC21:02
jaypipessoren: no worries, that's been fixed. we just now need python-distutils-extra installed on the hudson box...21:02
jaypipesmtaylor: hey, around? imsplitbit may have some info that is related to our glance.version bug...21:03
sorenjaypipes: Really? Why is that?21:03
jaypipesmtaylor: he's running into a similar issue on his machine:
sorenjaypipes: I'm not questioning it, I'm just wondering if it's something I'll need to do in a few other places, so I'd like to understand the motivation.21:04
jaypipesmtaylor: HOWEVER, glance.common.config does contain load_paste_app: :(21:04
jaypipessoren: no longer using babel for i18n, since it's halfbaked. distutils.extra is now used to build the message catalogs. (python build_i18n)21:05
sorenjaypipes: Try "python -v" You can see exactly which module it loads as "config".21:05
sorenjaypipes: Cool.21:05
sorenjaypipes: Thanks.21:05
jaypipesimsplitbit: see soren's note above21:05
jaypipesimsplitbit: try python -v21:05
sorenjaypipes: It's already installed on the hudson box.21:05
jaypipessoren: See here: "To build nova you need"21:06
sorenberendt: your bug 722477 is exactly why I've been working on iptables these last couple of days.21:06
uvirtbotLaunchpad bug 722477 in nova "iptables-restore failed" [High,Triaged]
jaypipessoren: perhaps mtaylor installed it just earlier?21:06
sorenjaypipes: I can check.21:06
sorenjaypipes: 2011-02-21 19:23:51 status installed python-distutils-extra 2.22-221:06
sorenjaypipes: Looks like it.21:06
sorenberendt: LOL. And I forgot that I replied on the bug.21:07
jaypipessoren: lol, ok, thx :)21:07
berendtsoren: :)21:08
jaypipessoren: hmm, well python -v was less than helpful... see pastie above. however, note line 112's UserWarning. Think you can explain what'd going on there? I'm not sure. :(21:10
sorenimsplitbit: Oh, I meant something like: python bin/glance-api /etc/nova/glance.conf21:11
imsplitbitone sec21:11
*** berendt has quit IRC21:16
openstackhudsonProject nova build #546: SUCCESS in 1 min 39 sec:
openstackhudsonTarmac: Adds translation catalogs and distutils.extra glue code that automates the process of compiling message catalogs into .mo files.21:17
*** reldan has joined #openstack21:18
imsplitbitjaypipes: I am sorting through the output but here it is.
*** nelson has joined #openstack21:19
sorenimsplitbit: Looks fine.21:19
sorenimsplitbit: Try: python -c 'import glance.common.config ; print dir(glance.common.config)'21:21
imsplitbitjaypipes: soren:
jaypipesimsplitbit: looks like load_paste_app is in there...21:25
jaypipesvery frustrating...21:25
*** miclorb has joined #openstack21:25
* soren makes Twilight zone noises21:26
jaypipessoren: I think the exact same thing is happening here:
sorenjaypipes: No, that's different.21:28
sorenjaypipes: That one I understand.21:28
sorenjaypipes: python-glance is installed on the Hudson box (to satisfy Nova's test suite).21:28
sorenjaypipes: um...21:29
sorenjaypipes: Err..21:29
sorenjaypipes: <something about sys.path>21:29
jaypipessoren: OK, so can you explain how to run glance's tests without a virtualenv when Nova on the same box needs some prior version of Glance installed?21:30
sorenjaypipes: Well, normally it should Just Work[tm].21:31
jaypipessoren: but it doesn't :)21:31
sorenjaypipes: $CWD should take precedence.21:31
*** eikke has quit IRC21:32
*** gregp76 has joined #openstack21:32
dubsquaredsoren:  are the ppa/release packages up-to-date with the latest point release?21:34
*** Ryan_Lane has quit IRC21:34
sorendubsquared: As far as I now.21:34
sorendubsquared: ...because as far as I know, we haven't made the point release yet.21:34
dubsquaredso, no.21:35
sorenSo yes.21:35
sorenIt's up-to-date.21:35
sorenThere is nothing newer than what is in the ppa => it's up-to-date.21:35
sorenjaypipes: hm... It does for me.21:35
termiesoren: any idea how hard it is to get tarmac to include the names of reviewers?21:36
sorentermie: Shouldn't be much work at all.21:36
*** gregp76 has quit IRC21:36
jaypipessoren: it works for me locally too, but Hudson throws the glance.version error every time it runs.21:36
termiesoren: cool, i'll check it out21:36
sorentermie: It seems to not even require code.21:38
sorentermie: It's a configuration thing.21:38
sorenjaypipes: Gimme a sec.21:39
*** ctennis has joined #openstack21:39
*** ctennis has joined #openstack21:39
sorentermie: It allows you to set a commit_message_template variable.21:41
sorentermie: It can include such cool things as <bugs_fixed> <author> <approved_by>..21:42
termiesoren: aye, just running across that now21:42
termiesoren: so i guess next step is getting the one we want in?21:42
sorentermie: Whip up a template, I'll apply it.21:43
sorentermie: Yeah.21:43
termiesoren: do we have any template right now? i'm not sure where these files live21:43
sorenNo template, so it's just the default.21:43
*** Ge0RgeR^ has joined #openstack21:44
sorentermie: ...which is whatever is set on the merge proposal.21:45
*** pothos_ has joined #openstack21:45
*** eikke has joined #openstack21:46
*** pothos has quit IRC21:47
*** pothos_ is now known as pothos21:47
termietarmac/lp has the opposite idea for review vs approve than what we've been discussing21:48
sorenjaypipes: Oh, got it.21:49
termielooks like default is just <commit_message>21:49
sorentermie: Yeah.21:49
sorenjaypipes: Ok, so in nova's bin/*, we manipulate the sys.path to make sure we always grab the stuff in $PWD.21:51
jaypipeswow, sqlalchemy-migrate's documentation is shit.21:51
sorenjaypipes: You need to do the same thing.21:51
dubsquaredsoren:  "There is nothing newer than what is in the ppa => it's up-to-date." —> Just wanted to know if the items left out of the release tarball were updated, etc21:51
termiejaypipes: less than stellar, agreed, many circular references21:51
jaypipessoren: we do this in all bin/* files:21:52
jaypipesROOT_DIR = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))21:52
jaypipessoren: that does not do it?21:52
sorenjaypipes: Nope.21:52
jaypipeshehe, janope.21:52
sorenjaypipes: That adds bin/ to the sys.path.21:52
sorenjaypipes: Not bin/../21:52
jaypipessoren: aha.21:52
soren...which is what you want.21:52
Ge0RgeR^hello a question there21:52
jaypipessoren: gotcha. ok, lemme push a change with that.21:53
jaypipessoren: thx mate.21:53
sorenjaypipes: That will almost certainly make it happy.21:53
sorenjaypipes: Sure thing.21:53
jaypipestermie: you know of a way to include an "UPDATE blah SET blah" in a migration file?21:53
Ge0RgeR^can openstack sit on Citrix XenServer?21:54
termiejaypipes: just put the relevant sqlalchemy commands in the upgrade unction21:54
termiejaypipes: you can do selects and all that, you'll probably need to import the models or make a relevant duplicate of one21:55
dubsquaredGe0RgeR^:  technically, it could, but right now the most used implementation is to install nova on domU and use XenAPI to make calls back to the hypervisor21:55
termiejaypipes: easiest way is probably just make a table with the same name and the appropriate columns and then check the sqlalchemy docs on how to run an update command21:56
termiejaypipes: (doesn't need to be a complete schema for the table)21:56
jaypipestermie: k, thx man21:57
*** littleidea has joined #openstack22:00
termieis sandywalsh the main person dealing with novatools ?22:00
*** littleidea has quit IRC22:03
*** ctennis has quit IRC22:03
*** littleidea has joined #openstack22:03
*** j05h has quit IRC22:04
mtaylorsoren, jaypipes: what's the idea you've been discussing about review vs. approve22:05
termiemtaylor: i was just mentioning that when we talk about "review" it is the process of reviewing, where to tarmac/launchpad the "reviewer" is the one who flips the approved state22:06
mtaylortermie: ah22:06
termiemtaylor: and "approvers" are the people who have done the review and marked themselves as being in approval22:06
*** Ephur has quit IRC22:07
jaypipestermie: ya, sandywalsh is the novatools master.22:12
*** mgoldmann has quit IRC22:12
sorenjaypipes: I've added a bit of docstring love to the iptables thing.22:12
jaypipessoren: cool. thx man22:12
sorenjaypipes: I'm not sure it's wha tyou're looking for.22:12
sorenjaypipes: If iptables it Greek to you, I'm not sure it'll help a whole lot.22:13
*** j05h has joined #openstack22:14
*** ctennis has joined #openstack22:16
*** ctennis has joined #openstack22:16
*** eikke has quit IRC22:16
czajkowskisoren: how's you22:20
*** dragondm has joined #openstack22:22
sorenIrish grammar. Can't beat it.22:22
* czajkowski stabs soren 22:23
czajkowskiIrish stabbing too :)22:23
kbringardif anyone is interested, I wrote a script to get OpenStack setup and running on CentOS 522:25
kbringardyou'll probably have to make some modifications to your environment, but it should mostly work22:25
sorenczajkowski: :)22:25
kbringardmake some modifications to make it work in your environment I mean22:26
kbringardnot *to* your environment22:26
*** jc_smith has joined #openstack22:27
sorenczajkowski: I is good, thank you.22:27
* soren is learning22:28
imsplitbitsoren, jaypipes: hub_cap helped me figure out what was going on22:28
sorenimsplitbit: pray tell.22:28
czajkowskisoren: :)  So when is the next sprint  for ye guys ?22:28
sorenczajkowski: Two months from now. Santa Clara.22:29
czajkowskisoren: I'm sure you'll all make that one :)22:29
sorenczajkowski: You coming'?22:29
imsplitbitI had an old python-glance deb installed and was trying to run glance from trunk.  I apt-get removed python-glance and everything then started running from the trunk root folder22:29
*** Dumfries has joined #openstack22:29
*** kbringard has quit IRC22:29
imsplitbitpython was prefering the glance.common.config from the package rather than CWD22:29
czajkowskisoren: tempting, running an open source and education conference over here in April.22:30
sorenimsplitbit: jaypipes just fixed that.22:30
czajkowskiI'm hoping to head to Canada in July if possible22:30
justinsbjaypipes, devcamcar: Any idea why this hasn't merged?  You've both approved it, I think you're both nova-core, and it just needs 2 nova-core Approves, right?
Dumfriesgetting  error with nova-network, czajkowski you should come to Vancouver if you heading this way ;)22:31
sorenjustinsb: Someone needs to set it to "Approved".22:31
sorenDumfries: Your eventlet is broken.22:31
justinsbsoren: Ah - didn't realize it was manual!  Thanks!22:32
sorenjustinsb: I'll do it.22:32
justinsbsoren: Cheers!22:32
* soren tips his hat22:32
Dumfriessoren: thanks man22:32
sorenDumfries: Sure thing.22:32
czajkowskiDumfries: I'll head to Guelph more than likely.22:32
Dumfriesczajkowski: nice, bloody cold that way. I am shaking here in Vancouver with plus 4 :)22:33
imsplitbitsoren: sweet!22:34
*** drico has joined #openstack22:34
*** mahadev has quit IRC22:34
czajkowskiDumfries: mate is over in manitoba  not visiting her, getting her to come to Guelph22:34
czajkowskisoren: so what is the criteria for picking a sprint location?22:35
Dumfriesczajkowski: good call, you are a lot safer bring her your way ;)22:35
czajkowskiDumfries: yup went over last year for my 30th had a blast22:35
Dumfriesczajkowski: awesome, where you coming from?22:36
*** GasbaKid has joined #openstack22:36
*** eikke has joined #openstack22:36
sorenczajkowski: Darned if I know.22:36
czajkowskiDumfries: Ireland22:36
czajkowskiDumfries: living in the UK now22:37
sorenczajkowski: "Far away from Soren" seems like a common criterion, at least.22:37
Dumfriesczajkowski: Very nice, coming for the warm warm sunshine I take it?22:37
czajkowskiDumfries: best mate moved over there 3 years ago. so an excuse to see her and have some sunshine!22:38
czajkowskisoren: poor you!22:38
czajkowskisoren: make them come to EU they'll have the long haul then22:38
Dumfriesczajkowski: great, stay away from molson canadian and you should be fine22:39
czajkowskino idea what that is.22:39
*** mdomsch has quit IRC22:39
annegentlewow the page is the fastest moving wiki page up the charts, now at number ten based on Pageviews22:40
sorenczajkowski: Not holding my breath on that one :-/22:40
openstackhudsonProject nova build #547: SUCCESS in 1 min 38 sec:
openstackhudsonTarmac: The OpenStack API was using the 'secret' as the 'access key'.  There is an 'access key' and there is a 'secret key'.  Access key ~= username.  Secret key ~= password.  This fix is necessary for the OpenStack Python API bindings to log in.22:42
sorenmtaylor: Do you have a simple way to test tarmac things?22:45
sorenmtaylor: We want to set this commit_message_template:
*** imsplitbit has quit IRC22:53
*** brd_from_italy has quit IRC22:59
*** gregp76 has joined #openstack22:59
*** gondoi has quit IRC23:02
*** eikke has quit IRC23:03
*** brd_from_italy has joined #openstack23:04
mtaylorsoren: why?23:05
mtaylorsoren: all of that information is already included in the commit as revision metadata23:05
*** that__guy has quit IRC23:06
sorenmtaylor: orly?23:06
mtaylorsoren: yup23:07
sorenmtaylor: How do I get my hands on the list of approvers?23:08
mtaylorsoren: one sec - asking in #bzr23:09
*** brd_from_italy has quit IRC23:09
sorenmtaylor: Got it.23:10
sorenmtaylor: bzr cat-revision -r -1, for instance.23:11
*** MarkAtwood has quit IRC23:12
sorenmtaylor: I had no idea that was included.23:14
mtaylorsoren: see! I anticipate your every need23:15
sorenmtaylor: It was really termie's request.23:15
sorenmtaylor: Perhaps he can fill you in on the use case.23:15
*** aimon has joined #openstack23:16
termiemtaylor: this is related to the thread on "developer review days"23:20
*** gregp76 has quit IRC23:20
termiemtaylor: the goal being to better call out the reviewers of the code23:20
*** pvo has joined #openstack23:29
*** Ryan_Lane has joined #openstack23:37
*** johnpur has quit IRC23:51
*** pvo has quit IRC23:57

Generated by 2.14.0 by Marius Gedminas - find it at!