*** rlucio has left #openstack | 00:02 | |
*** miclorb has joined #openstack | 00:04 | |
*** pvo has joined #openstack | 00:08 | |
*** ChanServ sets mode: +v pvo | 00:08 | |
*** daleolds has joined #openstack | 00:10 | |
*** Cromulent has quit IRC | 00:13 | |
*** schisamo has quit IRC | 00:14 | |
vishy | fixed flat dhcp | 00:14 |
---|---|---|
vishy | lp:~vishvananda/nova/fix-flatdhcp if anyone wants to try it | 00:14 |
vishy | found out you have to be very careful attaching bridges to interfaces that already have an ip | 00:15 |
vishy | it also includes a new nova.sh which allows you to specify the switch easily | 00:19 |
*** maplebed has joined #openstack | 00:22 | |
*** jc_smith has quit IRC | 00:33 | |
*** miclorb has quit IRC | 00:33 | |
*** londo has quit IRC | 00:33 | |
*** londo has joined #openstack | 00:33 | |
*** pvo has quit IRC | 00:43 | |
*** michaeldreamhost has joined #openstack | 00:46 | |
*** joearnold has quit IRC | 00:52 | |
*** Ryan_Lane is now known as Ryan_Lane|away | 00:52 | |
*** miclorb_ has joined #openstack | 00:59 | |
*** abecc has quit IRC | 01:09 | |
*** abecc has joined #openstack | 01:11 | |
*** abecc has joined #openstack | 01:11 | |
*** londo has quit IRC | 01:14 | |
*** londo__ has joined #openstack | 01:14 | |
*** maplebed has quit IRC | 01:24 | |
*** dragondm has quit IRC | 01:35 | |
*** stewart has quit IRC | 01:46 | |
michaeldreamhost | I've set my blueprints to pending approval. should I be having them reviewed by someone other than rick before I set them to that? | 01:52 |
*** littleidea has quit IRC | 02:12 | |
*** littleidea has joined #openstack | 02:12 | |
*** littleidea has quit IRC | 02:17 | |
*** littleidea has joined #openstack | 02:24 | |
*** BK_man has quit IRC | 02:27 | |
*** daleolds has quit IRC | 02:37 | |
*** littleidea has quit IRC | 02:40 | |
*** littleidea_ has joined #openstack | 02:59 | |
creiht | for the nasa peoples: http://askaninja.com/taxonomy/term/909 | 03:05 |
*** annegentle-web has quit IRC | 03:11 | |
*** littleidea_ has quit IRC | 03:17 | |
*** littleidea has joined #openstack | 03:17 | |
vishy | creiht: laff | 03:18 |
*** littleidea has quit IRC | 03:21 | |
*** joearnold has joined #openstack | 03:28 | |
*** schisamo has joined #openstack | 03:30 | |
*** localhost2 has quit IRC | 03:34 | |
*** localhost2 has joined #openstack | 03:38 | |
*** joearnold has quit IRC | 03:38 | |
*** abecc has quit IRC | 03:49 | |
*** AimanA has quit IRC | 03:54 | |
*** AimanA has joined #openstack | 03:54 | |
*** mdomsch has joined #openstack | 04:11 | |
*** littleidea has joined #openstack | 04:16 | |
*** kashyapc has quit IRC | 04:20 | |
*** kashyapc_ has joined #openstack | 04:20 | |
*** schisamo_ has joined #openstack | 04:26 | |
*** schisamo has quit IRC | 04:26 | |
*** schisamo_ is now known as schisamo | 04:26 | |
*** ar1 has quit IRC | 04:28 | |
*** ar1 has joined #openstack | 04:28 | |
*** BK_man has joined #openstack | 04:32 | |
*** localhost2 has quit IRC | 04:34 | |
*** omidhdl has joined #openstack | 04:35 | |
*** localhost2 has joined #openstack | 04:36 | |
*** omidhdl1 has joined #openstack | 04:36 | |
*** omidhdl has quit IRC | 04:37 | |
*** kashyapc_ has quit IRC | 04:42 | |
*** stewart has joined #openstack | 04:47 | |
*** AimanA is now known as HouseAway | 04:48 | |
*** joearnold has joined #openstack | 04:55 | |
*** hazmat has quit IRC | 05:01 | |
*** stewart has quit IRC | 05:08 | |
*** joearnold has quit IRC | 05:14 | |
*** omidhdl1 has quit IRC | 05:19 | |
*** kashyapc has joined #openstack | 05:19 | |
*** ramkrsna has joined #openstack | 05:19 | |
*** ramkrsna has joined #openstack | 05:19 | |
*** ArdRigh has joined #openstack | 05:20 | |
*** littleidea has quit IRC | 05:22 | |
*** omidhdl has joined #openstack | 05:25 | |
*** littleidea has joined #openstack | 05:28 | |
*** mdomsch has quit IRC | 05:38 | |
*** schisamo has quit IRC | 05:44 | |
*** f4m8_ is now known as f4m8 | 05:46 | |
*** sirp2 has quit IRC | 05:51 | |
*** hazmat has joined #openstack | 06:13 | |
*** stewart has joined #openstack | 06:13 | |
*** randybias has joined #openstack | 06:18 | |
*** littleidea has quit IRC | 06:20 | |
*** arthurc has joined #openstack | 06:26 | |
dubsquared | anyone home :O | 06:30 |
*** randybias has quit IRC | 06:32 | |
*** zaitcev has quit IRC | 06:33 | |
*** joearnold has joined #openstack | 06:40 | |
*** littleidea has joined #openstack | 06:42 | |
*** miclorb_ has quit IRC | 06:43 | |
*** JordanRinke_ has joined #openstack | 06:46 | |
*** JordanRinke has quit IRC | 06:49 | |
*** dfg_ has quit IRC | 06:50 | |
*** dfg_ has joined #openstack | 06:51 | |
*** joearnold has quit IRC | 06:52 | |
*** joearnold has joined #openstack | 07:01 | |
*** ibarrera has joined #openstack | 07:02 | |
*** jcannava has joined #openstack | 07:04 | |
*** larissa has quit IRC | 07:07 | |
*** larissa has joined #openstack | 07:08 | |
*** alekibango has quit IRC | 07:08 | |
*** jcannava has quit IRC | 07:08 | |
*** alekibango has joined #openstack | 07:08 | |
*** chmouel has quit IRC | 07:11 | |
*** jc_smith has joined #openstack | 07:13 | |
*** littleidea has quit IRC | 07:42 | |
*** allsystemsarego has joined #openstack | 07:43 | |
*** allsystemsarego has joined #openstack | 07:43 | |
*** littleidea has joined #openstack | 07:43 | |
*** littleidea has quit IRC | 07:47 | |
*** almaisan-away is now known as al-maisan | 07:54 | |
*** littleidea has joined #openstack | 07:55 | |
*** sandywalsh has quit IRC | 08:03 | |
*** BK_man has quit IRC | 08:04 | |
*** chmouel has joined #openstack | 08:04 | |
*** perestrelka has quit IRC | 08:04 | |
*** perestrelka has joined #openstack | 08:07 | |
*** BK_man has joined #openstack | 08:09 | |
ttx | michaeldreamhost: you don't need to. | 08:13 |
ttx | michaeldreamhost: only if you need some other people's input on them. | 08:14 |
*** sandywalsh has joined #openstack | 08:16 | |
*** addu has joined #openstack | 08:18 | |
*** joearnold has quit IRC | 08:18 | |
addu | hello | 08:18 |
*** addu has left #openstack | 08:20 | |
*** miclorb has joined #openstack | 08:21 | |
*** calavera has joined #openstack | 08:23 | |
*** BK_man has quit IRC | 08:26 | |
*** sandywalsh has quit IRC | 08:35 | |
*** sandywalsh has joined #openstack | 08:36 | |
*** perestrelka has quit IRC | 08:41 | |
*** matiu has quit IRC | 08:41 | |
*** perestrelka has joined #openstack | 08:41 | |
*** littleidea has quit IRC | 09:07 | |
*** iammartian_ has quit IRC | 09:12 | |
*** JordanRinke_ has quit IRC | 09:17 | |
*** colinnich has joined #openstack | 09:20 | |
*** jc_smith has quit IRC | 09:37 | |
*** patri0t_ has joined #openstack | 09:38 | |
*** doude has joined #openstack | 09:42 | |
[CnC] | ping | 09:44 |
*** BK_man has joined #openstack | 09:47 | |
*** shahid_ has joined #openstack | 09:50 | |
shahid_ | hi | 09:50 |
shahid_ | I want to know whether there is any option for taking instance backup and restoring | 09:52 |
*** ptremblett has quit IRC | 09:52 | |
joshuamckenty | that would be snapshotting | 09:54 |
joshuamckenty | it's a proposed blueprint, it's not started yet | 09:54 |
*** patri0t_ has quit IRC | 09:56 | |
*** patri0t_ has joined #openstack | 09:58 | |
doude | Hi all, is it possible to use the VLAN network mode on multiple server installation ? | 09:58 |
doude | Or the flat network mode is the only way actually ? | 09:58 |
soren | VLAN mode is very likely what you'd be using on any proper, multi-server installation. | 10:09 |
soren | doude: * | 10:09 |
soren | annegentle: I've fixed your man page builds. | 10:12 |
*** shahid_ has quit IRC | 10:12 | |
soren | annegentle: http://bazaar.launchpad.net/~soren/nova/novaman/revision/404 | 10:14 |
soren | annegentle: If you're cool with that change, you can merge it into your tree ("bzr merge lp:~soren/nova/novaman" followed by "bzr commit"). | 10:15 |
doude | soren: In the doc manual http://wiki.openstack.org/NovaInstall/MultipleServer, the config network is flat network and I saw on the doc http://etherpad.openstack.org/NovaMultinodeInstall, only one nova-network should run in the cloud. | 10:18 |
doude | How the netwok configuration (bridge, VLAN) is set on the slave node ? | 10:19 |
doude | soren: The admin shiould set it ? | 10:20 |
*** miclorb has quit IRC | 10:21 | |
soren | doude: Yeah, this is configuration of the physical network topology. | 10:21 |
soren | doude: It is not something the consumers of the cloud service should worry about. | 10:21 |
joshuamckenty | when he's awake, you can bug vishy for details | 10:24 |
soren | doude: The fact that http://wiki.openstack.org/NovaInstall/MultipleServer uses flat networking doesn't mean that you have to. It's just an example. | 10:25 |
doude | soren: ok, it's what I'm thinking. But if I want use VLAN mode, I set only one instance of nova-network ? | 10:27 |
soren | doude: I'm not completely sure, to be honest. :( | 10:27 |
soren | doude: In my test environment, I can't really use VLAN.s | 10:28 |
*** shahid_ has joined #openstack | 10:29 | |
*** miclorb_ has joined #openstack | 10:31 | |
doude | soren: ok, and do you use the flatDHCP mode ? | 10:32 |
*** dizz|away is now known as dizz | 10:42 | |
*** ptremblett has joined #openstack | 10:53 | |
*** ArdRigh has quit IRC | 10:53 | |
*** ramkrsna has quit IRC | 10:55 | |
*** doude has quit IRC | 10:56 | |
*** GasbaKid has joined #openstack | 11:06 | |
*** kashyapc has quit IRC | 11:20 | |
*** miclorb_ has quit IRC | 11:25 | |
*** ramkrsna has joined #openstack | 11:35 | |
*** ramkrsna has joined #openstack | 11:35 | |
*** doude has joined #openstack | 11:41 | |
*** littleidea has joined #openstack | 11:41 | |
*** littleidea_ has joined #openstack | 11:41 | |
*** littleidea has quit IRC | 11:45 | |
*** littleidea_ is now known as littleidea | 11:45 | |
*** iammartian_ has joined #openstack | 12:23 | |
*** joshuamckenty has left #openstack | 12:26 | |
*** aryko has quit IRC | 12:28 | |
*** ctennis has quit IRC | 12:33 | |
*** localhost2 has quit IRC | 12:34 | |
*** localhost2 has joined #openstack | 12:35 | |
*** nijaba has quit IRC | 12:47 | |
*** doude has quit IRC | 12:52 | |
sandywalsh | g/day! | 12:53 |
sandywalsh | g'day even | 12:54 |
*** ctennis has joined #openstack | 12:57 | |
soren | sandywalsh: o/ | 12:59 |
sandywalsh | soren, what's your timezone vs. GMT? | 12:59 |
sandywalsh | +1 hr? | 13:00 |
soren | sandywalsh: Yup. | 13:02 |
soren | It's 14:02 here. | 13:02 |
soren | (That means 2:02PM) | 13:02 |
sandywalsh | :) well, it's nice to have someone else online when I get in. I'm -4 gmt (9am) | 13:03 |
soren | Oh. I somehow thought you were in SAT. | 13:03 |
sandywalsh | nova scotia | 13:04 |
_0x44 | All of the best people are in CET right now. :) | 13:04 |
ttx | _0x44: isn't that always the case ? | 13:05 |
sandywalsh | hipsters | 13:05 |
sandywalsh | ;) | 13:05 |
_0x44 | ttx: No, not always. Sometimes the best people are in other time zones. | 13:06 |
_0x44 | :) | 13:06 |
soren | sandywalsh: Ah, not even in the US. Imagine that :) | 13:07 |
ttx | _0x44: we'll all agree that last week, that was the case. :P | 13:07 |
_0x44 | Nah, last week all the best people were still in CET. :D | 13:09 |
sandywalsh | haha | 13:09 |
dendrobates | jaypipes: Was the teller-api blueprint completed in Austin, or is there still work to do? | 13:10 |
*** smoser has joined #openstack | 13:14 | |
*** nijaba has joined #openstack | 13:35 | |
*** nijaba has joined #openstack | 13:35 | |
*** kashyapc has joined #openstack | 13:36 | |
*** kashyapc has quit IRC | 13:38 | |
*** kashyapc has joined #openstack | 13:38 | |
*** doude has joined #openstack | 13:39 | |
*** westmaas has joined #openstack | 13:40 | |
*** ar1 has quit IRC | 13:42 | |
*** littleidea has quit IRC | 13:47 | |
*** littleidea has joined #openstack | 13:47 | |
*** littleidea has quit IRC | 13:52 | |
*** hggdh has quit IRC | 13:52 | |
*** hggdh has joined #openstack | 14:01 | |
*** littleidea has joined #openstack | 14:04 | |
*** annegentle_ has joined #openstack | 14:05 | |
*** iammartian_ has quit IRC | 14:10 | |
jaypipes | dendrobates: still work to do | 14:14 |
dendrobates | jaypipes: thanks | 14:15 |
*** jcannava has joined #openstack | 14:15 | |
*** schisamo has joined #openstack | 14:17 | |
*** omidhdl has left #openstack | 14:18 | |
dendrobates | gholt: is the https://blueprints.launchpad.net/swift/+spec/client-side-chunking targeted at bexar or cactus? | 14:19 |
*** schisamo has quit IRC | 14:20 | |
*** aliguori has quit IRC | 14:21 | |
dendrobates | jaypipes: will you be working on the teller api? and if so bexar or cactus? I would prefer bexar. | 14:22 |
*** jcannava has quit IRC | 14:23 | |
gholt | dendrobates: bexar, definitely | 14:23 |
dendrobates | gholt: thanks | 14:23 |
*** Podilarius has joined #openstack | 14:24 | |
*** nijaba has quit IRC | 14:27 | |
*** nijaba has joined #openstack | 14:31 | |
*** nijaba has joined #openstack | 14:31 | |
*** abecc has joined #openstack | 14:33 | |
*** littleidea has quit IRC | 14:34 | |
*** littleidea has joined #openstack | 14:34 | |
*** abecc has joined #openstack | 14:34 | |
*** jdarcy has joined #openstack | 14:34 | |
*** piken_ is now known as piken | 14:34 | |
*** sirp1 has joined #openstack | 14:35 | |
*** gundlach has joined #openstack | 14:42 | |
*** annegentle_ has quit IRC | 14:48 | |
ttx | For those using Google calendar or such, I created a .ics link for the Openstack release schedule, see http://wiki.openstack.org/BexarReleaseSchedule | 14:48 |
ttx | Let me know if it works for you. | 14:48 |
jaypipes | dendrobates: that's not really up to me :) I'll work on whatever is higher priority... | 14:49 |
* ttx pauses | 14:50 | |
*** littleidea has quit IRC | 14:51 | |
*** f4m8 is now known as f4m8_ | 14:51 | |
dendrobates | jaypipes: ok, bexar it is. | 14:53 |
*** littleidea has joined #openstack | 14:54 | |
*** blamar has quit IRC | 14:58 | |
sandywalsh | with bzr/lp if I'm working on a blueprint, there is no lp### associated correct? I should branch with a human-readable name, like "admin-only-api" ? | 15:03 |
*** damon__ has joined #openstack | 15:04 | |
sandywalsh | I also assume this means I can't use --fixes when commiting? | 15:05 |
*** dubsquared1 has joined #openstack | 15:06 | |
*** dizz has quit IRC | 15:11 | |
alekibango | ttx: ics work is korganizer | 15:12 |
alekibango | gj | 15:12 |
*** pvo has joined #openstack | 15:13 | |
*** ChanServ sets mode: +v pvo | 15:13 | |
*** deshantm has quit IRC | 15:16 | |
*** deshantm has joined #openstack | 15:18 | |
ttx | sandywalsh: correct. Though you could create one one more bugs and link them to the blueprint... if you really want to :) | 15:18 |
* ttx is awaiting the Great Revelation, when Launchpad will merge bugs and blueprints. | 15:19 | |
sandywalsh | cool, thanks ttx | 15:19 |
sandywalsh | ... are they any notes on creating a nova dev environment (separate from my running nova env)? | 15:20 |
sandywalsh | couldn't see any on the wiki | 15:20 |
sandywalsh | specifically the virtual env stuff in run_tests | 15:20 |
ttx | sandywalsh: annegentle plans to add one doc about "getting from the branching the code to testing your feature/fix" | 15:20 |
sandywalsh | excellent ... perhaps I can work with her on that since I'm going through it now | 15:21 |
*** JordanRinke has joined #openstack | 15:21 | |
ttx | alekibango: thanks for the feedback. I'll add Cactus dates to that ics feed soon. | 15:22 |
dabo | sandywalsh: keep me in the loop - I just connected to my dev box that ant set up in the lab | 15:23 |
sandywalsh | dabo, will do | 15:23 |
*** ctennis has quit IRC | 15:25 | |
*** ramkrsna has quit IRC | 15:25 | |
*** dendrobates is now known as dendro-afk | 15:34 | |
*** theron has quit IRC | 15:35 | |
spy | where do you specifiy what connection nova uses? i want to flip to xenapi and play around with that | 15:36 |
*** Ryan_Lane|away has quit IRC | 15:36 | |
annegentle | soren: thank you very very much. Went well beyond my knowledge so I needed the help. :) Man page now building, can someone merge https://code.launchpad.net/~annegentle/nova/novaman? | 15:36 |
*** deshantm has quit IRC | 15:36 | |
soren | spy: --connection_type=xenapi | 15:37 |
annegentle | mtaylor: Nova now contains a man page that's built from rst, when I build locally it appears in build/man. It's pending a merge, but wanted to give you a heads up to look for it. | 15:37 |
spy | ah, thanks, does that get added to nova-manage.conf? | 15:38 |
*** deshantm has joined #openstack | 15:38 | |
*** littleidea has quit IRC | 15:38 | |
sandywalsh | annegentle, I'm just setting up a dev env for myself (separate from my working nova vm) is there a doc in-progress on this? | 15:39 |
sandywalsh | I can add my notes | 15:39 |
*** schisamo has joined #openstack | 15:41 | |
gundlach | spy: that would work | 15:41 |
gundlach | spy: anybody who doesn't need that flag would ignore it | 15:41 |
*** ctennis has joined #openstack | 15:42 | |
spy | gotcha | 15:42 |
spy | thanks | 15:42 |
*** doude has quit IRC | 15:42 | |
annegentle | sandywalsh: you know, I don't think we have one yet. It's on the list for ttx and I to write, but I hadn't started it yet. How about http://wiki.openstack.org/PythonDevelopmentEnvironment for starters? | 15:42 |
*** doude has joined #openstack | 15:43 | |
*** littleidea has joined #openstack | 15:43 | |
spy | well that exceptioned heh | 15:43 |
annegentle | sandywalsh: Also, see http://wiki.openstack.org/GettingTheCode | 15:43 |
sandywalsh | annegentle, works for me, I'll add some notes there. thx | 15:43 |
sandywalsh | annegentle, the pythondevelopmentenvironment page is exactly up the right alley | 15:44 |
annegentle | sandywalsh: good, my first guess was my gut :) | 15:47 |
piken | I would vote that the pythondevelopmentenvironment page should link of a pythondevelopment page which embed's pep8 instilling that it is the standard used in the project. | 15:47 |
sandywalsh | piken, +1 | 15:47 |
*** littleidea has quit IRC | 15:49 | |
gundlach | spy: paste? | 15:51 |
gundlach | spy: are you ant? | 15:51 |
piken | I would even go as far as suggesting we provide in the source tree a predefined project file for at least one major python ide allowing people that have never worked with development using vi. | 15:52 |
sandywalsh | piken, perhaps not in the source tree, but a separate project. The #vim comments in the source code irk me currently | 15:53 |
sandywalsh | but I like the idea | 15:53 |
soren | gundlach: I've at least been working under the assumption that "spy" == "ant". Otherwise, there are a few conversations I need to reread to make sense of them :) | 15:53 |
gundlach | soren: in my half-paying-attention haze i was answering spy as if he were someone new to the project, and then i thought "wait, i've seen that name before" and then "wait, didn't ant rename himself to spy one day?" and decided i couldn't leave it unresolved :) | 15:54 |
gundlach | it's bad enough that everybody has random irc handles, now we've got different handles on different IRC servers :) | 15:55 |
soren | gundlach: IRC is meant to be confusing. It's ok. | 15:55 |
gundlach | good, i'll just go with the flow. | 15:55 |
spy | spy: yeah, i'm ant, looks like it was looking for the XenApi python bindings, i snagged it, just need to figure out where it's looking for it | 15:57 |
spy | i should just flip back to my other nick heh | 15:57 |
spy | tried dropping it into site-packages but it doesn't appear to be picking it up there | 15:58 |
*** littleidea has joined #openstack | 15:58 | |
sandywalsh | SPY! | 15:58 |
sandywalsh | virtual_env is pretty cool | 15:59 |
spy | ant was taken and i like three letter nicks | 15:59 |
creiht | dendro-afk: so which mailing list is supposed to be the one list to rule them all? | 15:59 |
sandywalsh | spy, god taken too? that would be fitting. | 15:59 |
spy | maybe dog hah | 16:00 |
sandywalsh | hah | 16:00 |
dabo | spy: yeah, I like 'hah' | 16:01 |
*** damon__ has quit IRC | 16:04 | |
*** littleidea has quit IRC | 16:05 | |
piken | soren: you have time to take a look at a new libvirt log. I think I got it further along, but now see an ebtables error that I am not sure of. | 16:07 |
*** Ryan_Lane has joined #openstack | 16:07 | |
piken | http://paste.openstack.org/show/164/ if you have time | 16:08 |
soren | piken: Not right now. Am cooking :) | 16:09 |
*** littleidea has joined #openstack | 16:09 | |
*** dubsquared1 has quit IRC | 16:10 | |
*** ibarrera has quit IRC | 16:11 | |
*** littleidea_ has joined #openstack | 16:12 | |
*** dubsquared1 has joined #openstack | 16:15 | |
*** littleidea has quit IRC | 16:15 | |
*** damon__ has joined #openstack | 16:16 | |
*** littleidea_ is now known as littleidea | 16:29 | |
*** dendro-afk is now known as dendrobates | 16:34 | |
hazmat | creiht, i get the sense that mailing lists aren't much used in the openstack projects | 16:36 |
dendrobates | creiht: the main openstack one. | 16:36 |
dendrobates | we need to use the ml more. | 16:37 |
hazmat | dendrobates, +1 | 16:37 |
creiht | dendrobates: ok | 16:37 |
heckj | which is the main openstack one? (sorry - not trying to be a jerk) | 16:37 |
hazmat | i heard about lots of activity at the summit with lots of attendees, but actual communication of what happened is pretty minimal on the lists | 16:37 |
creiht | hazmat: It was decided at the conference that we would consolidate all the mailinglists to one | 16:38 |
annegentle | heckj: no worries, I couldn't find it either | 16:38 |
creiht | heckj: the one pointed to here: https://launchpad.net/~openstack | 16:38 |
hazmat | heckj, openstack@lists.launchpad.net afaik | 16:38 |
eday | who was it that was saying we shouldn't be putting code in __init__.py? And was the reason due to something around recursive dependencies? | 16:39 |
*** jcannava has joined #openstack | 16:39 | |
hazmat | annegentle, that should be mailing list should be prominent on https://launchpad.net/openstack fwiw | 16:39 |
heckj | Thanks - at least I am subscribed to that one. | 16:39 |
gundlach | eday: me | 16:39 |
hazmat | creiht, sounds good to me, if only they said it on the same on the ml ;-) | 16:39 |
creiht | lol | 16:39 |
gundlach | eday: part of the argument is so that if you want to import a module to look at it, you don't want side effects. | 16:39 |
annegentle | hazmat: agreed. dendrobates, can you give me permissions to edit that front page for the Launchpad page? | 16:39 |
gholt | That's anne, always wanting more power. ;) | 16:40 |
creiht | If only we had someone who's job it was to communitcate with the community.... :) | 16:40 |
gundlach | eday: but there's weird stuff with importing in a thread as well -- paramiko tends to barf when you import it under certain conditions because you get deadlocks -- can't remember the details. | 16:40 |
eday | gundlach: hmm, ok | 16:40 |
annegentle | heh | 16:41 |
gundlach | eday: esp. as we have eventlet breaking that rule -- 'import eventlet' has side effects -- putting code into __init__ that requires eventlet means that any subpackages being imported will run eventlet's side effects | 16:41 |
jaypipes | dendrobates, ttx, heckj: https://blueprints.launchpad.net/openstack-ci/+spec/basic-build-automation | 16:41 |
gundlach | eday: to clarify: i think somebody else originally said "let's not put code in __init__", but i dunno if they meant not even including classes and methods. i'm talking about *running* code in __init__. | 16:42 |
* creiht likes minimal code in __init__ | 16:42 | |
gundlach | creiht: reason? | 16:42 |
gholt | Yeah, I think more generally: __init__.py should not have side-effects or circular dependency problems. :) | 16:43 |
gholt | Past that, whatever goes goes. Hehe | 16:43 |
creiht | gundlach: what gholt said, and also a lot is just personal preference | 16:43 |
* jaypipes puts stuff in __init__.py for module configuration stuff (like module-level constants that apply to a number of files in the module directory) and when there are name shadowing concerns (like naming a module "json" or something...) | 16:44 | |
creiht | Not sure why it bugs me when someone creates a module, and puts all the code for the module in __init__.py | 16:44 |
vishy | eday: that was me | 16:44 |
gundlach | creiht, gholt: can you get circular dependency problems if you don't use the 'from module import word' syntax and don't have side effects? | 16:45 |
jaypipes | creiht: lots of stuff bothers you. I just assume something bothers you, and I am surprised if you don't get bothered by something. :) | 16:45 |
creiht | hah | 16:45 |
gholt | lol | 16:45 |
jaypipes | :) | 16:45 |
eday | vishy: any more reasons to add? | 16:46 |
eday | creiht: well, in that case you should just make name.py, not name/__init__.py so it would bug me too :) | 16:47 |
creiht | hah | 16:47 |
creiht | Maybe we should define a process for determining if something bugs me or not | 16:48 |
creiht | ;) | 16:48 |
eday | if (True) ? | 16:48 |
creiht | lol | 16:48 |
gundlach | creiht is bugged by the unneeded parens | 16:48 |
creiht | hahaha | 16:48 |
gholt | gundlach: Dunno, lotsa folks use from _ import _. Never really fiddled with with all the permutations personally. I just tend to not put anything in __init__.py Heh | 16:48 |
creiht | but to __init__.py it doesn't bother me enough to really say someone *must* or *should* do things a certain way | 16:49 |
eday | in any event, I think it's a good rule we should enforce more in nova | 16:49 |
gundlach | gholt: from package import module is fine, but from module import name can barf (because if that module hasn't been imported all the way down to the definition of that name, you're in trouble). leaving __init__ empty will certainly help avoid problems, but i think it's OK to do when it makes the package structure make sense. | 16:49 |
sandywalsh | got the tests running in my dev env ... wiki updated http://wiki.openstack.org/PythonDevelopmentEnvironment | 16:49 |
gundlach | eday: i would argue that we continue putting methods or classes in __init__ when they make sense there in the package structure. e.g. nova.api.openstack.ApiRouter needs to live in .../openstack/__init__.py, lest we have to import nova.api.openstack.router.ApiRouter or something dumb | 16:50 |
annegentle | sandywalsh: thanks for updating the wiki, I'll add testing stuff once I know more | 16:51 |
gholt | gundlach: Yeah, I mean doing a from eventlet import sleep is better than trying to remember that it's really from eventlet.greenthread.sleep. | 16:51 |
jaypipes | creiht: don't you mean making a process to see if something *doesn't* bother you? ;P | 16:52 |
creiht | hah | 16:52 |
piken | with all the talk about people making modules that irk you someone answer this one. | 16:53 |
* jaypipes has a feeling if creiht ever read the source code for MySQL he would explode into little pieces... | 16:53 | |
sandywalsh | annegentle, hmm, I made changes to testing, but it's not showing up for me ... odd caching going on. sec | 16:53 |
gundlach | sandywalsh: there's a useful trick with virtualenvs that i use to avoid having to install the virtualenv every time i branch: move your newly-created virtualenv one directory up, and it'll work for all your branches | 16:53 |
* heckj chuckles | 16:53 | |
creiht | jaypipes: well C is meant to be like that :) | 16:53 |
piken | Do you rather see a module used with imports the norm with import mod.name or where the morons put from name import * in __init__.py and do just import mod in the main. | 16:53 |
jaypipes | creiht: lol | 16:53 |
sandywalsh | gundlach, good point! | 16:53 |
jaypipes | heckj: feel free to add to that spec page on the CI stuff... | 16:54 |
gundlach | sandywalsh: and then i made a bash function called 'venv' that searches up the directory tree to turn on the virtualenv no matter what dir i'm currently in... i'll paste to the wiki | 16:54 |
jaypipes | heckj: here too, since it's mostly just placeholders right now: http://wiki.openstack.org/ContinuousIntegration#preview | 16:54 |
sandywalsh | gundlach, that would be excellent! | 16:54 |
heckj | jaypipes - will do. getting my head around paramiko & paste right now - and writing doc pages based on reading the code outlined to date | 16:55 |
gundlach | heckj: careful when you import paramiko ;) | 16:55 |
jaypipes | heckj: awesomeness. | 16:55 |
*** calavera has quit IRC | 16:55 | |
heckj | Also trying to wrangle a set of servers at my office to replicate the openstack-ci work into - want to run a functional test lab with my team here if we can | 16:55 |
jaypipes | vishy: what was the word that we decided to bring out of 1980s retirement? was it "rad"? | 16:56 |
heckj | gunlach: do I start immediately floating to the ceiling or something? | 16:56 |
sandywalsh | gundlach, shouldn't nova.sh be in trunk if it's taking on such an important role in dev? | 16:56 |
jaypipes | heckj: nice | 16:56 |
gholt | no, .... rad | 16:56 |
jaypipes | :) | 16:58 |
annegentle | sandywalsh: Ah, I see it now, I was looking under "Running Tests" heading. I'll remove that section. | 16:59 |
gundlach | sandywalsh: that's a q for vishy | 16:59 |
gundlach | i'd say "yes" -- let's get it productionized | 16:59 |
comstud | jaypipes- bodacious? | 16:59 |
gundlach | so you can say "nova.sh install --network=flat" and have it Do The Right Thing | 16:59 |
sandywalsh | comstud, jaypipes "most savory" ? | 16:59 |
sandywalsh | gundlach, +1 | 16:59 |
*** mdomsch has joined #openstack | 17:00 | |
sandywalsh | vishy, can we get nova.sh in trunk? seems to be the defacto tool for install | 17:00 |
gundlach | sandywalsh: you could paste it into trunk and propose for merge and then it'd be in trunk and we could all start hacking on it :) | 17:00 |
gundlach | or rather, it would be more visible and we *would* all start hacking on it. | 17:00 |
hazmat | is anyone else bothered by the half sync/async code mixing.. ie sometimes using twisted but with sync communications, or sometimes using eventlet, it makes things very hard to audit cleanly. | 17:00 |
piken | gundlach: That would be nice in the nova.sh scripts for now, but I am working on the unified-installer that will handle a direct install on the system. I was also thinking of adding a config option to it that will allow you to have it auto config things like flat networking based on options you pass. | 17:01 |
gundlach | hazmat: yes. we're getting rid of Twisted, as decided at the recent summit. | 17:01 |
gundlach | hazmat: all sync-looking code is the goal | 17:01 |
gundlach | (though some of it is async under the covers) | 17:02 |
gundlach | piken: cool! | 17:02 |
hazmat | gundlach, good to know, thanks, are there any blog posts regarding the summit decisions? | 17:02 |
piken | sandywalsh: putting nova.sh into trunk is actually counter productive as it is limited to ubuntu only, reason I opened the unified-install blueprint and have been working on a POC | 17:02 |
gundlach | hazmat: no idea. annegentle: ^^ ? | 17:02 |
*** doude has quit IRC | 17:02 | |
gundlach | piken: i must have missed the blueprint. did it get approved for bexar? | 17:02 |
piken | gundlach: not as of yet. | 17:02 |
annegentle | hazmat: thinking, looking, just a sec. | 17:03 |
piken | It most likely will be post bexar due to the holiday season | 17:03 |
piken | gundlach: take a look at the spec so far. https://github.com/piken/unified-install/wiki/Nova-Unified-Install-System | 17:03 |
sandywalsh | piken, gotcha ... that's great that a universal installer will be in place. Could we bring the python setup portion from tools/install_venv.py into it as well? | 17:03 |
dabo | piken: coming in late on your question about: from _ import * | 17:03 |
dabo | IMO, that pollutes the current namespace so that the dev can save typing a few keystrokes; e.g., mymethod() instead of myclass.mymethod() | 17:03 |
sandywalsh | piken, for dev environments | 17:04 |
dabo | from _ import * should be avoided | 17:04 |
piken | sandywalsh: that is a possibility, I am working to handle dependencies and such first, then want to work other things like that into the spec. | 17:04 |
sandywalsh | piken, great ... thanks for the heads up! | 17:04 |
piken | dabo: I agree, was curious what others though. | 17:05 |
piken | *thought | 17:05 |
piken | sandywalsh: also the start of the POC code is in that github repo that the current spec wiki is on. | 17:05 |
gundlach | piken: you may wish to talk w/ jaypipes about config format, as he's working on a replacement for gflags that uses .ini configs. we probably don't want a JSON manifest that specifies all the same options as our .ini configs do. | 17:05 |
*** maplebed has joined #openstack | 17:06 | |
piken | the JSON manifest is for the install process only. dependencies, configuration of the dependencies, and such | 17:06 |
annegentle | hazmat: I don't see a "summary decisions" post/page anywhere. It's a good idea. Thinking.... not sure if I can write it but I could attempt it. | 17:06 |
piken | The config portion would be seperate. | 17:06 |
gundlach | dabo: in a package, from _ import * can be useful, because you're saying "I defined all my foo-related stuff in a foo submodule for ease of development, but for my users, i want all those names available in the top-level package." | 17:06 |
*** piken is now known as piken_lunch | 17:06 | |
gundlach | piken_lunch: my apologies for scanning your blueprint too quickly :) | 17:07 |
dabo | gundlach: true, but that's not namespace pollution, as you are doing that in your module, and you control the namespace. | 17:07 |
gundlach | dabo: right. i thought that was piken's original q -- about package authors doing that. of course 'from _ import *' is deprecated in user code -- didn't the BDFL even proclaim that somewhere? | 17:07 |
comstud | totally tubular, dude | 17:08 |
dabo | gundlach: don't know if it's officially deprecated, but strongly frowned upon, yes. | 17:08 |
dabo | I thought that piken_lunch's question was at the nova level, not the individual packages. If not, my mistake. | 17:09 |
hazmat | annegentle, maybe send a draft around to the list for others to offer input | 17:10 |
uvirtbot | New bug: #677093 in swift "stats system won't process the last hour's stats" [Low,New] https://launchpad.net/bugs/677093 | 17:11 |
*** maplebed has quit IRC | 17:11 | |
*** maplebed has joined #openstack | 17:12 | |
jaypipes | gundlach: FYI, dendrobates placed the common options processing as a Low priority, so I will be focusing on Glance Teller API and continuous integration, not openstack-common at least for Bexar... | 17:13 |
annegentle | hazmat: an excellent idea. Adding to my "todo" and will collaborate with ttx (since I don't know all the decisions but he just might) | 17:13 |
hazmat | annegentle, awesome | 17:14 |
dendrobates | jaypipes: I am waiting for stakeholder feedback on priorities. they could change, and you could influence them. | 17:14 |
gundlach | jaypipes: right then. hey, how's the unit tests branch? have you touched it since presummit? i haven't. | 17:16 |
jaypipes | dendrobates: like I said, I'm happy to work on anything...doesn't matter :) | 17:16 |
jaypipes | gundlach: no I haven't...need to look again at it. | 17:16 |
gundlach | jaypipes: there's a stack of RS blueprints that need doing. would you be crushed if i didn't return to unit test work? | 17:17 |
vishy | sandywalsh: it is in trunk | 17:17 |
gundlach | jaypipes: i wouldn't want to crush you, is all. | 17:17 |
vishy | sandywalsh: contrib/nova.sh | 17:17 |
vishy | sandywalsh: the newest version is in my fix-flatdhcp branch, which will hopefully get merged as soon as people test it | 17:18 |
dendrobates | jaypipes: so, you have no opinion about what is more important to the project? | 17:18 |
vishy | jaypipes: yup | 17:18 |
dendrobates | jaypipes: we should talk about how much work each of these things are. | 17:18 |
vishy | secondarily it was to flip the meaning of grody to something good | 17:18 |
jaypipes | gundlach: :) I'm cool with that, brutha. :) | 17:21 |
gundlach | jaypipes: v00t. maybe by the time we return to that branch Twisted will have disappeared and the tests will get yet simpler. | 17:22 |
jaypipes | dendrobates: my personal opinion is that CI is top priority, then Teller API, then openstack-common stuff, simply because of impact to devs and users. | 17:22 |
*** colinnich has quit IRC | 17:22 | |
gundlach | jaypipes: my brutha from anotha mutha | 17:22 |
jaypipes | gundlach: :) | 17:22 |
*** littleidea has quit IRC | 17:23 | |
jaypipes | dendrobates: the reason I rank CI over all others is that without CI, we are going to be implementing features and not have a clue before we release those features whether they have impacted system performance.. | 17:23 |
creiht | jaypipes: While I agree that CI is a good thing, can we really do enough at that level to ensure that we don't impact system performance at scale? | 17:24 |
creiht | Well there are some things that could be detected | 17:25 |
creiht | I just wouldn't want people to think that since we have CI, that they don't need to still test the stuff at scale | 17:25 |
eday | dendrobates: https://blueprints.edge.launchpad.net/nova/+spec/unified-service-architecture want to make sure this doesn't get missed. termie is traveling at the moment, but I think this is pretty high to get things cleaned up (makes other work easier, less code that needs to be changed later). I'd actually like to see this before I get too deep in the distributed scheduler since it simplifies that work, and I might even work on it if termie doesn't have time | 17:28 |
jaypipes | creiht: the CI system *is* going to be testing things at scale... | 17:29 |
creiht | jaypipes: ahh... awesome then! :) | 17:30 |
jaypipes | creiht: on a 200-node cluster eventually. | 17:30 |
creiht | nice | 17:30 |
creiht | jaypipes: then I say rock on my friend, rock on! | 17:30 |
creiht | :) | 17:30 |
jaypipes | creiht: we're starting smaller, of course, but the goal is to have *real* deployment huddles doing the testing. | 17:30 |
creiht | cool | 17:30 |
jaypipes | creiht: and not just RS. | 17:30 |
jaypipes | creiht: that isn't yet in scope for Bexar, but I think it's do-able for Cactus. | 17:31 |
jaypipes | creiht: but that might be a prickly goal to meet ;) | 17:31 |
creiht | indeed :) | 17:31 |
gundlach | <groan/> | 17:31 |
jaypipes | gundlach: hehe | 17:32 |
jaypipes | gundlach: I'm slowly becoming my father... ugh. | 17:32 |
*** kashyapc has quit IRC | 17:32 | |
gundlach | same here. so far one of my legs and some of my internal organs have fully converted. i kind of wish the transformation was purely mental. | 17:32 |
*** joearnold has joined #openstack | 17:45 | |
*** littleidea has joined #openstack | 17:48 | |
*** GasbaKid has quit IRC | 17:52 | |
*** heckj has quit IRC | 17:53 | |
*** littleidea has quit IRC | 17:57 | |
eday | vishy: are you fmailiar with the ec2/s3 image wrapper you guys are using? curious if you actually use 'update_image', since this is not part of the ec2 api | 18:02 |
*** al-maisan is now known as almaisan-away | 18:05 | |
*** daleolds has joined #openstack | 18:06 | |
vishy | yeah we use it | 18:06 |
vishy | from dashboard | 18:06 |
eday | vishy: hmm, ok. does it just update metadata? any reason you don't use ModifyImageAttribute' ? | 18:09 |
*** piken_lunch is now known as piken | 18:09 | |
*** Ryan_Lane has quit IRC | 18:10 | |
*** filler has joined #openstack | 18:14 | |
*** johnpur has joined #openstack | 18:16 | |
*** ChanServ sets mode: +v johnpur | 18:16 | |
vishy | eday: it updates metadata that doesn't exist in amazon world | 18:22 |
vishy | like description | 18:22 |
piken | Soren: you still here or done for the evening? | 18:23 |
vishy | xtoddx did it, I'm not sure why he didn't hack up ModifyImageAttribute. Perhaps just to keep a clean distinction between existing aws commands and extensions | 18:23 |
eday | ahh, ok | 18:24 |
xtoddx | eucatools issues | 18:24 |
xtoddx | why name it the same if eucatools/boto won't support it | 18:24 |
eday | makes sense | 18:25 |
*** zaitcev has joined #openstack | 18:27 | |
creiht | zaitcev: I'm finally nailing down what should be done for the first pass of the s3 emulation layer here: http://etherpad.openstack.org/swift3 | 18:28 |
creiht | If you are interested :) | 18:28 |
zaitcev | creiht: I am interested, but I may not be in a position to offer any constructive help for now. Being boggled down with iwhd at work... | 18:29 |
creiht | zaitcev: no worries, and I understand :) | 18:29 |
sandywalsh | vishy, well, look at that! | 18:29 |
zaitcev | How long do Etherpads live? | 18:29 |
creiht | I think pretty much as long as we want them to :) | 18:30 |
zaitcev | "Some tools only use SOAP, for example amazon api tools" -- what? | 18:30 |
zaitcev | I need to look into that. I don't think it's true, but could be. | 18:31 |
creiht | zaitcev: I think that was from the discussion at the conference | 18:31 |
eday | xtoddx: any reason you didn't just use the S3 api directly for those other attributes? | 18:32 |
*** rlucio has joined #openstack | 18:33 | |
xtoddx | eday: boto choked on doing anything outside of modifying privacy attribute | 18:34 |
eday | xtoddx: yeah, but instead of routing through ec2 with new method, why not just POST directly to the s3 clone? | 18:35 |
xtoddx | eday: i dunnow. i guess i was following the existing pattern. | 18:36 |
eday | ok, just trying to understand the image stuff, trying to consolidate it with the other image service APIs that were added | 18:36 |
xtoddx | plus, this could work with glance as well | 18:36 |
jdarcy | Will the S3 middleware as-is pass through ETags and If-None-Match, or would code have to be added for that? | 18:36 |
*** gondoi has joined #openstack | 18:38 | |
vishy | still would love people to test lp:~vishvananda/nova/fix-flatdhcp/ especially on the rack cloud and maybe even with multiple servers | 18:41 |
creiht | jdarcy: good question... I believe that everything will be passed on through | 18:41 |
*** arthurc has quit IRC | 18:42 | |
vishy | we have a little hack to make metadata work with HAProxy and multiple apis. Is that something that we should add in? | 18:43 |
vishy | to trunk? | 18:43 |
zaitcev | Does anyone know how much the cheapest Rackspace VMs go for? Ditto for CF. I was kinda thinking about running a developer's cloud out of pocket, but the non-permanency of EC2s scares me. | 18:48 |
zaitcev | Maybe if I have an instance of CLD up and running, everyone will love it and adopt it into OpenStack :-) | 18:49 |
dabo | vishy: I'm setting up my lab env, but when that's done, I'm planning on checking out your flat fix | 18:49 |
*** GasbaKid has joined #openstack | 18:49 | |
vishy | dabo: cool, let me know how it goes | 18:50 |
dabo | will do | 18:50 |
*** GasbaKid has joined #openstack | 18:50 | |
jdarcy | I've been running one of the smallest RSC instances for several months. About $10/month. | 18:53 |
zaitcev | That sounds manageable. | 18:53 |
jdarcy | It's handy so that people don't look in logs and see "someone from Red Hat" visiting their site, in addition to occasional cloud experiments. | 18:54 |
*** mdomsch has quit IRC | 18:55 | |
zaitcev | I use ani-nouto.animeblogger.net for that | 18:55 |
jdarcy | Yeah, if you already have a solution (or just work from home) it's probably not an issue. I was surprised how much people tended to notice, though. | 18:56 |
zaitcev | But I'm not supposed to use that one to run Swift or other cloud stuff on that, obviously. | 18:56 |
jdarcy | Probably an issue for anyone who works at a big company nowadays. | 18:56 |
uvirtbot | New bug: #677138 in nova "Error messages from api that include unicode fail to display" [Undecided,New] https://launchpad.net/bugs/677138 | 18:56 |
*** ptremblett has quit IRC | 18:57 | |
*** ptremblett has joined #openstack | 18:57 | |
*** mdomsch has joined #openstack | 19:00 | |
*** pvo_ has joined #openstack | 19:05 | |
*** ChanServ sets mode: +v pvo_ | 19:05 | |
*** pvo has quit IRC | 19:07 | |
*** pvo_ is now known as pvo | 19:07 | |
zaitcev | Apparently, no IPv6. Oh well, tunnelbroker.net to the rescue. | 19:07 |
*** Ryan_Lane has joined #openstack | 19:08 | |
*** BK_man has quit IRC | 19:18 | |
*** cbuben has joined #openstack | 19:20 | |
*** damon__ has quit IRC | 19:31 | |
*** damon__ has joined #openstack | 19:32 | |
sandywalsh | gundlach, with your .bash script for virtual env, do you have to do run_tests -N ? | 19:33 |
sandywalsh | gundlach, actually that won't work either. Does run_tests.sh work with .nova-venv moved? | 19:35 |
openstackhudson | Project nova build #247: SUCCESS in 1 min 2 sec: http://173.203.104.204:8080/job/nova/247/ | 19:35 |
soren | piken: I'm back for a little bit right now. | 19:36 |
soren | piken: Exact same error: 11:37:13.800: debug : virRunWithHook:836 : Command stderr: iptables v1.4.10: unknown option `--ctdir' | 19:36 |
*** ctennis has quit IRC | 19:38 | |
gundlach | sandywalsh: no, i just always say No | 19:39 |
gundlach | sandywalsh: and yes, it does. whaterror are you getting? | 19:39 |
*** dendrobates is now known as dendro-afk | 19:44 | |
gundlach | anotherjesse, annegentle: http://paste.openstack.org/show/165/ is my first draft of a tutorial for extending the openstack API. i'm sure it's terribly formatted. what should i change? | 19:45 |
gundlach | [that q was for annegentle, i threw anotherjesse in there just to get your attn in case you wanted an early peek anotherjesse] | 19:45 |
_cerberus_ | Question for the room: https://code.launchpad.net/~jk0/nova/improve-pylint-scores/+merge/41094 failed to merge because of a random test failure. All pass locally, and for what it's worth, I merged the branch into trunk locally and they all passed. Should we just re-attempt the merge? | 19:46 |
gundlach | yes. | 19:46 |
gundlach | if it fails 2 or 3 times, shout at the author of the test | 19:47 |
_cerberus_ | gundlach: thanks | 19:47 |
*** dendro-afk is now known as dendrobates | 19:49 | |
annegentle | gundlach: reading now | 19:50 |
soren | vishy: ^ (see _cerberus_' comment) | 19:50 |
soren | _cerberus_: It's all vishy's fault. | 19:50 |
soren | :) | 19:50 |
jk0 | :) | 19:50 |
_cerberus_ | soren: I can get behind that idea | 19:51 |
soren | _cerberus_: I saw the same failure yesterday, but it didn't reproduce. Another one did, and vishy fixed it. We just need more of his magic. | 19:51 |
_cerberus_ | Can we change the topic to represent it? :-) | 19:51 |
openstackhudson | Project nova-tarmac build #36,147: FAILURE in 3 min 8 sec: http://173.203.104.204:8080/job/nova-tarmac/36147/ | 19:51 |
_cerberus_ | hah, ok | 19:52 |
soren | jk0: You need to set a commit message or a description. | 19:52 |
soren | jk0: The commit message is what will end up in the ChangeLog file. | 19:52 |
jk0 | hm, thought I did that | 19:52 |
jk0 | oh, I see, it must not pick it up from the bzr commit message? | 19:52 |
soren | No. | 19:53 |
*** gundlach has quit IRC | 19:53 | |
soren | Typically, there are many commits in a single merge. | 19:53 |
_cerberus_ | jk0: think of it as a git squash-merge | 19:53 |
dendrobates | jk0: and it breaks the build, if you leave it off | 19:53 |
jk0 | I smell what'cher cookin | 19:53 |
soren | But now at least openstackhudson will tell us about it, instead of someone having to go and look at hudson every once in a while. | 19:54 |
soren | (Just enabled that 10 minutes ago) | 19:54 |
*** ChanServ sets mode: +v _cerberus_ | 19:55 | |
*** Gibby_away has quit IRC | 19:56 | |
* JordanRinke waves at jk0 | 19:56 | |
jk0 | heya champ | 19:56 |
openstackhudson | Project nova-tarmac build #36,148: STILL FAILING in 2 min 59 sec: http://173.203.104.204:8080/job/nova-tarmac/36148/ | 19:56 |
jk0 | ^ that should go away now that I've added a commit message | 19:58 |
uvirtbot | jk0: Error: "that" is not a valid command. | 19:58 |
_cerberus_ | hah | 19:58 |
creiht | heh | 19:58 |
creiht | how many bots do we have in this room? :) | 19:58 |
dendrobates | o/ | 19:58 |
creiht | heh | 19:59 |
* soren takes a break | 19:59 | |
openstackhudson | Yippie, build fixed! | 20:02 |
openstackhudson | Project nova-tarmac build #36,149: FIXED in 3 min 28 sec: http://173.203.104.204:8080/job/nova-tarmac/36149/ | 20:02 |
jk0 | \o/ | 20:02 |
openstackhudson | Project nova build #248: SUCCESS in 56 sec: http://173.203.104.204:8080/job/nova/248/ | 20:05 |
*** pvo has quit IRC | 20:05 | |
*** gundlach has joined #openstack | 20:06 | |
*** sirp1 has quit IRC | 20:10 | |
*** sirp1 has joined #openstack | 20:10 | |
*** pvo has joined #openstack | 20:17 | |
*** ChanServ sets mode: +v pvo | 20:17 | |
*** Gibby_away has joined #openstack | 20:18 | |
*** dendrobates is now known as dendro-afk | 20:21 | |
*** dendro-afk is now known as dendrobates | 20:22 | |
*** rsampaio has joined #openstack | 20:22 | |
*** ArdRigh has joined #openstack | 20:23 | |
*** iammartian has left #openstack | 20:26 | |
*** pvo has quit IRC | 20:27 | |
*** pvo has joined #openstack | 20:30 | |
*** ChanServ sets mode: +v pvo | 20:30 | |
*** brd_from_italy has joined #openstack | 20:32 | |
*** iammartian has joined #openstack | 20:33 | |
vishy | _cerebus_, soren: annoying | 20:34 |
vishy | guess we need a longer wait | 20:35 |
vishy | just rerunning the tests should work though | 20:35 |
vishy | s/tests/merge | 20:35 |
vishy | ah looks like you already reran it | 20:36 |
*** rlucio has quit IRC | 20:42 | |
*** gundlach has quit IRC | 20:47 | |
*** ctennis has joined #openstack | 20:50 | |
*** ctennis has joined #openstack | 20:50 | |
*** dendrobates is now known as dendro-afk | 20:54 | |
*** littleidea has joined #openstack | 21:04 | |
*** pvo has quit IRC | 21:10 | |
jaypipes | _0x44, sirp1: around? I'm wondering if you would be opposed to getting rid of the Parallax/Teller split and consolidating to just a single Glance API? | 21:11 |
*** pvo has joined #openstack | 21:11 | |
*** ChanServ sets mode: +v pvo | 21:11 | |
JordanRinke | interesting | 21:13 |
sirp1 | jaypipes: hmm, *thinking* | 21:14 |
*** gundlach has joined #openstack | 21:14 | |
*** rlucio has joined #openstack | 21:15 | |
*** dragondm has joined #openstack | 21:19 | |
*** spectorclan has joined #openstack | 21:20 | |
sirp1 | jaypipes: so GlanceAPI would just be the union of the two APIs? I guess some operations wouldn't make sense if we're only talking to Parallax outside of the context of a specific OpenStack installation (e.g. no Teller) | 21:20 |
*** ptremblett_ has joined #openstack | 21:26 | |
*** BK_man has joined #openstack | 21:28 | |
*** pvo has quit IRC | 21:28 | |
*** ptremblett has quit IRC | 21:29 | |
spectorclan | Design Summit Feedback - If you attended, please take 10 minutes and provide your thoughts at http://www.surveymonkey.com/s/openstackdesignsummit2010 | 21:31 |
spectorclan | Also, if you want to chat 1:1 please let me know and I can setup a call to get your thoughts. Early thinking on 2011 Design Summits undreway | 21:31 |
alekibango | spectorclan: do you think you would like to look on my spec etherpad page now for few minutes ? i would like to get some comments | 21:34 |
alekibango | spectorclan: use my diagrams in any way you please if you want them, please help to improve them if you can: http://alekiba.dyndns.org/nova/ | 21:35 |
*** pothos_ has joined #openstack | 21:35 | |
alekibango | my spec etherpad page (resource partitioning) is http://etherpad.openstack.org/M9RemacWog -- comments welcome | 21:36 |
annegentle | alekibango: I've got a merge request to get the diagrams into nova now, we all like them muchly. Next I need to place them in the .rst files, working on that now. | 21:36 |
alekibango | annegentle: seen the request, will look on that later tonight | 21:37 |
*** pothos has quit IRC | 21:37 | |
*** jc_smith has joined #openstack | 21:38 | |
*** pothos_ has quit IRC | 21:38 | |
*** pothos has joined #openstack | 21:38 | |
annegentle | alekibango: thanks for the review | 21:38 |
alekibango | thanks for award :) was glad to help in my small way... | 21:39 |
*** HouseAway is now known as AimanA | 21:40 | |
*** dendro-afk is now known as dendrobates | 21:40 | |
*** piken is now known as pikenAFK | 21:40 | |
gundlach | eday: re side-effects in __init__: jaypipes provides a nice example of why you don't want them, in https://code.launchpad.net/~jaypipes/openstack-common/options/+merge/40985 . To generate --help output, he imports every module he can find, to read their config options. | 21:43 |
alekibango | annegentle: i would like someone to improve my C1/C2 diagrams to include more deails... | 21:44 |
hazmat | gundlach, how is that avoidable outside of centralizing gflag usage? | 21:44 |
gundlach | hazmat: no, my point is that if any modules had side effects, his approach wouldn't work. it's normal to expect, as jaypipes does, that you can import classes and not have crazy side effects. | 21:45 |
hazmat | ah | 21:45 |
eday | gundlach: yeah, def agree to avoid those side effects. but in this case I don think importing every module you can find for --help a good idea :) | 21:46 |
hazmat | i find randomly importing everything in site, also worthy of something not to do | 21:46 |
gundlach | eday: like hazmat said, how else might we do it, if we're defining config options in various modules? | 21:46 |
eday | gundlach: as you ask for those modules to be loaded, you see the --help | 21:46 |
jaypipes | gundlach: one sec, on phone with sirp1 | 21:47 |
eday | gundlach: this is a user preference really, and mine is to only see the help relevant to what I'm doing, not everything under the sun (like bin/nova-* tools are now) | 21:47 |
jaypipes | gundlach: thx for putting me on the spot :P | 21:55 |
gundlach | jaypipes: glad to help! | 21:55 |
jaypipes | hehe | 21:55 |
hazmat | also as the codebase gets bigger it means waiting, proportionally longer waits on --help options, tools like bzr for example ended up having to optimize lazyloads so they could get output out faster to the user, by not loading everything in sight. | 21:56 |
*** spectorclan has quit IRC | 21:56 | |
*** spectorclan has joined #openstack | 21:57 | |
*** pvo has joined #openstack | 22:00 | |
*** ChanServ sets mode: +v pvo | 22:00 | |
*** rsampaio has quit IRC | 22:00 | |
*** johnpur has quit IRC | 22:02 | |
annegentle | should --help be replaced by a man page? Or do you generally expect --help to be sort of Oregontrail-adventure-text-like? (That's what it seems like to me now) | 22:02 |
gundlach | lol | 22:03 |
annegentle | it's an honest question, I really don't know :) | 22:03 |
gundlach | that's what eday is advocating -- that we make --help more terse by default. i think it should still exist, as many people will try --help to learn about a binary | 22:03 |
gundlach | i'm waiting for jaypipes to hang up the phone and start arguing with hazmat and eday about it | 22:03 |
pvo | as long as I don't die of dysentery | 22:04 |
gundlach | too late | 22:04 |
*** miclorb has joined #openstack | 22:04 | |
jaypipes | gundlach: sorry, I'm off phone now :) | 22:05 |
annegentle | okay, yeah all the options made for a really lonnng man page | 22:05 |
creiht | perhaps it now brings up to question weather it is worth having a command line option for *every* single configurable item in a config file? | 22:06 |
jk0 | would it be acceptable/ideal to sys.exit(1) a nova-* process if rabbit isn't running upon startup? | 22:08 |
jaypipes | creiht: I think it makes sense to allow any option you can put in a configuration file to be specified on the command line...but maybe I'm in the minority. | 22:08 |
jk0 | or do we want to keep it running and polling for rabbit? | 22:08 |
dendrobates | jk0: how would you know rabbit isn't running? | 22:08 |
*** Ryan_Lane has quit IRC | 22:08 | |
hazmat | jaypipes, at some point though the sheer number of options that are useful in the cli vs. in a config file, means that the cli ux suffers | 22:09 |
jk0 | dendrobates: right now I'm leaning towards a try/except when the connection is initiated | 22:09 |
gundlach | jaypipes: maybe we could have --help print out the 'common' options, and man page/--help can point people to the config options and say 'oh yeah you can specify any of these via -- too.' | 22:09 |
creiht | yup | 22:09 |
sandywalsh | is the plan that the nova api will always support both openstack *and* ec2 commands? Or will it be a configuration switch (--api=openstack)? | 22:09 |
gundlach | creiht: yup to me or hazmat? | 22:09 |
jk0 | I'm not seeing a way to handle it more gracefully using carrot | 22:09 |
gundlach | sandywalsh: there are two separate APIs. | 22:10 |
jaypipes | hazmat: true enough, yes. I'm on the fence, really... | 22:10 |
creiht | gundlach: to hazmat sorry :) | 22:10 |
gundlach | sandywalsh: i think the current idea is that they could run on different ports. | 22:10 |
sandywalsh | gundlach, yes, but it seems that api answers to both, but openstack prefixes with /1.0/... | 22:10 |
sandywalsh | gundlach, so the api is chosen by the uri | 22:11 |
jaypipes | hazmat: it's pretty easy to adapt the work I've already done to make options visible or not visible in --help... and have a --help-extended kind of thing (like MySQL does) | 22:11 |
* annegentle gets ready to edit the man page, it doesn't indicate that config options can be set in files. | 22:11 | |
jaypipes | hazmat: I don't want to force my opinion on anyone (if I did, I wouldn't have come up with this solution ;) ) | 22:11 |
dendrobates | jk0: rabbit could be on another box, and the network cold hicup | 22:11 |
dendrobates | you should not create a dependency on a service starting. | 22:12 |
hazmat | jaypipes, true, but i have some other comments earlier re load time to display help.. i don't really see other options besides centralizing flag usage though, or instituting additional lazyloading.. or at least localizing flags within a package to a known module to reduce the imports at cli startup | 22:12 |
* jaypipes thinks my pug is signaling a desire for dinner... | 22:12 | |
creiht | if you have to use map reduce to display your --help, you are doing it wrong :) | 22:12 |
jk0 | dendrobates: that's what I'm trying to decide, would it be better to just stop the process from starting if it can't find rabbit, or just continue to retry the connection | 22:13 |
dendrobates | jk0: continue | 22:13 |
hazmat | creiht, :-) | 22:13 |
dendrobates | and log it occasionally | 22:13 |
jk0 | excellent, thanks | 22:13 |
JordanRinke | it would be nice if the connection refused errors from the nova services, said what ip/port they were trying. | 22:14 |
JordanRinke | jk0's comment just made me think of that | 22:14 |
jk0 | that's what I have in my branch now | 22:14 |
jk0 | more detailed error | 22:14 |
JordanRinke | I <3 you | 22:14 |
gundlach | sandywalsh: erm, i do believe there's a flag that lets you set which subdomains serve which APIs -- e.g. you can say that requests coming in on ec2.yourdomain.com only serve the ec2 API. i wrote that, but eday followed behind me to clean some stuff up. | 22:15 |
jaypipes | crisis averted...pug has been fed. | 22:16 |
sandywalsh | gundlach, ah yes, I saw the subdomain stuff ... will need to see how that's enforced. thx | 22:16 |
gundlach | sandywalsh: nova.openstack.api.__init__.ApiRouter.__init__'s mapper conditions. | 22:16 |
gundlach | whoops, sorry, s/openstack// | 22:17 |
sandywalsh | gundlach, cool. thx | 22:18 |
hazmat | also i wonder if folks now that carrot has been deprecated | 22:21 |
eday | hazmat: in favor of what? celery? | 22:22 |
creiht | I heard it was broccoli | 22:22 |
hazmat | eday, kombu is the name i think, fix all the design flaws in celery | 22:22 |
gundlach | ta ta all, see you monday | 22:23 |
hazmat | i never really understood the value of celery vs. raw amqp.. but that's a different story | 22:23 |
*** gundlach has quit IRC | 22:24 | |
soren | jk0: Awesome, I've wanted to add that for a while. | 22:24 |
hazmat | eday, http://groups.google.com/group/carrot-users/browse_thread/thread/8c7dc50d4f0a0822 | 22:24 |
eday | hazmat: interesting, thanks | 22:29 |
*** littleidea has quit IRC | 22:33 | |
*** jimbaker has joined #openstack | 22:34 | |
*** localhost2 has quit IRC | 22:34 | |
JordanRinke | Any Rackers in here want to weigh in on the linux distro email on SAFun in re to development for OpenStack ? Curious to see if anyone suggests anything but Ubuntu | 22:37 |
*** littleidea has joined #openstack | 22:37 | |
spectorclan | JordanRinke: I saw someone say they use Mac OSX | 22:37 |
JordanRinke | yeah, talking to him about that right now | 22:37 |
* jk0 votes Windows | 22:37 | |
jk0 | :D | 22:38 |
JordanRinke | jk0: well I am a windows engineer, but I have my RHCE so I feel obligated to dual boot something lol | 22:38 |
alekibango | JordanRinke: i use debian | 22:39 |
dubsquared1 | I've OSX as my primary OS for the last 12 months or so here, and have had great success with custom tools and general dev/admin work | 22:39 |
dubsquared1 | used^ | 22:39 |
dubsquared1 | and helpdesk is buying licenses for the good stuff now, so you get that all for free :D | 22:39 |
alekibango | i dont use windows for 12 years now | 22:39 |
alekibango | but i am not racker | 22:40 |
alekibango | ;) | 22:40 |
*** localhost2 has joined #openstack | 22:40 | |
vishy | i was developing on osx for a while, but now I do it all in an ubuntu vm | 22:40 |
vishy | lack of libvirt / kvm on osx makes it tough | 22:40 |
creiht | JordanRinke: I've used linux as my primary desktop for about 5 years now (and Ubuntu for many of those) | 22:41 |
*** littleidea has quit IRC | 22:41 | |
vishy | developing on a vm seems a lot safer in general when dealing with network configs and stuff that can hose your system | 22:41 |
hazmat | not a racker, but ubuntu ftw.. (minorly biased though) | 22:41 |
spectorclan | I ran Fedora for 2 years in a previous life | 22:42 |
jk0 | Ubuntu seems to be working pretty well so far for nova dev | 22:43 |
hazmat | vishy, good point, it found it helpful as well if i'm not on a vm but playing with network for vm routing, to only use the eth devices, and save the wireless for sanity, vms in vms.. turtles all the way down ;-) | 22:43 |
jk0 | I like how all of the python modules are available in 10.10, no need to compule gflags | 22:43 |
creiht | JordanRinke: and yeah I do most of my swift dev in a vm (local and on a slice) | 22:43 |
JordanRinke | I am expecting to do primary dev on box and kick it in to a VM for testing | 22:43 |
*** brd_from_italy has quit IRC | 22:44 | |
jk0 | *compile | 22:44 |
vishy | hazmat: yo, dawg, I heard you like vms... | 22:44 |
hazmat | :-) bring back the vax | 22:44 |
creiht | hah | 22:44 |
*** GasbaKid has quit IRC | 22:46 | |
*** littleidea has joined #openstack | 22:46 | |
*** jimbaker has quit IRC | 22:47 | |
*** jimbaker has joined #openstack | 22:47 | |
*** littleidea has quit IRC | 22:51 | |
*** spectorclan has quit IRC | 22:51 | |
*** jimbaker has quit IRC | 22:52 | |
*** allsystemsarego has quit IRC | 22:55 | |
*** perestrelka has quit IRC | 22:56 | |
*** westmaas has quit IRC | 22:56 | |
*** gondoi has quit IRC | 22:56 | |
*** perestrelka has joined #openstack | 22:56 | |
JordanRinke | OK so... best IDE (or whatever you use) for Python dev on a *nix platform? | 22:57 |
creiht | vim | 22:57 |
creiht | and ipython in another shell | 22:58 |
creiht | :) | 22:58 |
JordanRinke | really just going the editor with syntax highlighting route? | 22:58 |
JordanRinke | interesting | 22:58 |
creiht | yeah | 22:58 |
creiht | I've tried the ide route several times, and always end up back with vim | 22:58 |
JordanRinke | I am used to .Net /heavy IDE dev | 22:58 |
jk0 | vim ftw | 22:59 |
creiht | If I am doing a lot of HTML/JS then I use Komodo Edit | 22:59 |
pvo | JordanRinke: yea, vim with some good syntax highlighting is pretty nice | 23:01 |
*** damon__ has quit IRC | 23:02 | |
creiht | JordanRinke: have you ever played with ipython? | 23:02 |
creiht | http://ipython.scipy.org/moin/ | 23:02 |
*** mdomsch has quit IRC | 23:02 | |
creiht | It is a must have for python dev :) | 23:02 |
mtaylor | vim | 23:03 |
JordanRinke | python will be a new venture for me | 23:03 |
JordanRinke | (don't worry, I won't be contributing for a while lol) | 23:03 |
JordanRinke | mostly I want to learn it, to be able to read the source for nova/swift to be able to figure out integration/direction etc while it is still in such rapid development | 23:04 |
vishy | creiht: s/ipython/bpython | 23:05 |
creiht | or that as well | 23:05 |
creiht | :) | 23:05 |
creiht | I like that I can run shell commands in ipython and easily mess with the results in python | 23:07 |
creiht | well with python code | 23:07 |
vishy | i like that bpython pops up docstrings and params for bultin methods | 23:07 |
creiht | I like that ipython gives me that when I ask it :) | 23:08 |
vishy | very helpful when i can't remember if it is .strip() or .trim() | 23:08 |
vishy | creiht: is there a way to pop up stuff like that in ipython? | 23:09 |
creiht | hitting tab will auto complete | 23:09 |
creiht | it doesn't pop up, but you can get doc strings by appending ? to the method | 23:10 |
vishy | cool. No help on args though :( | 23:10 |
creiht | yeah it shows you args as well with ? | 23:10 |
creiht | though they don't pop up | 23:10 |
vishy | ? is cool | 23:10 |
creiht | I get annoyed with things always popping up, but I like to have it when I do need it | 23:11 |
*** jimbaker has joined #openstack | 23:11 | |
creiht | but don't get me wrong, bpython is pretty cool too :) | 23:14 |
*** cbuben has left #openstack | 23:18 | |
*** spike has left #openstack | 23:18 | |
*** ptremblett has joined #openstack | 23:22 | |
*** burris has quit IRC | 23:24 | |
hazmat | i'm a big fan of rlcompleter2 its a bit more minimal than bpython or ipython.. but it does nice tab completion on attrs/methods without the bloat | 23:27 |
*** jcannava has quit IRC | 23:31 | |
*** Podilarius has left #openstack | 23:38 | |
*** schisamo has quit IRC | 23:48 | |
*** maplebed has quit IRC | 23:54 | |
*** jaypipes has quit IRC | 23:56 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!