| *** AimanA has joined #openstack | 00:01 | |
| *** AimanA has quit IRC | 00:04 | |
| *** pvo has joined #openstack | 00:08 | |
| *** ChanServ sets mode: +v pvo | 00:08 | |
| *** pvo has quit IRC | 00:10 | |
| *** pvo has joined #openstack | 00:11 | |
| *** ChanServ sets mode: +v pvo | 00:11 | |
| *** odyi is now known as ody | 00:12 | |
| *** ody is now known as Guest80220 | 00:13 | |
| *** AimanA has joined #openstack | 00:14 | |
| *** abecc has quit IRC | 00:14 | |
| *** Guest80220 is now known as odyi | 00:16 | |
| *** odyi is now known as odi | 00:16 | |
| *** kashyapc has quit IRC | 00:16 | |
| *** odi is now known as odyi | 00:16 | |
| *** kashyapc has joined #openstack | 00:17 | |
| *** jakedahn has quit IRC | 00:19 | |
| *** jakedahn has joined #openstack | 00:21 | |
| *** Ryan_Lane has joined #openstack | 00:23 | |
| *** kashyapc has quit IRC | 00:27 | |
| *** kashyapc has joined #openstack | 00:27 | |
| *** jc_smith has quit IRC | 00:28 | |
| *** gustavomzw has quit IRC | 00:28 | |
| *** jakedahn has quit IRC | 00:29 | |
| *** sirp1 has quit IRC | 00:29 | |
| *** kashyapc has quit IRC | 00:33 | |
| *** sebastianstadil has quit IRC | 00:35 | |
| *** jakedahn has joined #openstack | 00:36 | |
| *** jakedahn has quit IRC | 00:38 | |
| *** silassewell has joined #openstack | 00:44 | |
| *** kashyapc has joined #openstack | 00:52 | |
| *** pvo has quit IRC | 00:56 | |
| *** DubLo7 has quit IRC | 01:04 | |
| *** sophiap has quit IRC | 01:09 | |
| *** hisaki has joined #openstack | 01:12 | |
| *** sophiap has joined #openstack | 01:16 | |
| *** gondoi has joined #openstack | 01:23 | |
| *** miclorb has joined #openstack | 01:26 | |
| *** pvo has joined #openstack | 01:26 | |
| *** ChanServ sets mode: +v pvo | 01:26 | |
| *** pvo has quit IRC | 01:33 | |
| *** rnirmal has quit IRC | 01:34 | |
| *** hisaki has quit IRC | 01:34 | |
| *** hazmat has joined #openstack | 01:43 | |
| *** rnirmal has joined #openstack | 01:58 | |
| *** gaveen has quit IRC | 02:10 | |
| *** hazmat has quit IRC | 02:12 | |
| *** sirp1 has joined #openstack | 02:28 | |
| *** miclorb has quit IRC | 02:37 | |
| *** miclorb_ has joined #openstack | 02:44 | |
| *** gondoi has quit IRC | 02:46 | |
| *** localhost has joined #openstack | 02:53 | |
| *** DubLo7 has joined #openstack | 02:58 | |
| *** abecc has joined #openstack | 02:59 | |
| *** drizzle has quit IRC | 02:59 | |
| *** abecc has joined #openstack | 02:59 | |
| *** kashyapc has quit IRC | 03:02 | |
| *** localhost has quit IRC | 03:04 | |
| *** localhost has joined #openstack | 03:04 | |
| *** kashyapc has joined #openstack | 03:31 | |
| pjz | anyone awake? | 03:33 |
|---|---|---|
| notmyname | yo | 03:34 |
| pjz | does swift still have a 5GB blob limit? | 03:38 |
| notmyname | yes, but we have code that will allow arbitrary object sizes | 03:39 |
| notmyname | it's been written, but is being tested now | 03:39 |
| pjz | I see | 03:39 |
| pjz | I knew you were working on it :) | 03:39 |
| creiht | I doubt it will make it for the austin release | 03:39 |
| notmyname | https://code.launchpad.net/~gholt/swift/largefiles | 03:40 |
| *** pvo has joined #openstack | 03:40 | |
| *** pvo has joined #openstack | 03:40 | |
| *** ChanServ sets mode: +v pvo | 03:40 | |
| *** pvo has quit IRC | 03:46 | |
| *** pvo has joined #openstack | 03:47 | |
| *** ChanServ sets mode: +v pvo | 03:47 | |
| *** pvo_ has joined #openstack | 03:50 | |
| *** ChanServ sets mode: +v pvo_ | 03:50 | |
| *** aimon_ has joined #openstack | 03:52 | |
| *** pvo has quit IRC | 03:53 | |
| *** pvo_ is now known as pvo | 03:53 | |
| *** aimon has quit IRC | 03:56 | |
| *** aimon_ is now known as aimon | 03:56 | |
| *** sneumann has joined #openstack | 03:57 | |
| *** silassewell has quit IRC | 04:00 | |
| *** joearnold has quit IRC | 04:02 | |
| *** AimanA is now known as HouseAway | 04:02 | |
| *** sneumann has quit IRC | 04:11 | |
| *** pharkmillups has joined #openstack | 04:21 | |
| pjz | proxy/server.py has some big blocks of repeated code: ObjectController.POST/DELETE, ContainerController.PUT/DELETE, AccountController.PUT/POST ; should I bother to fix? | 04:26 |
| *** sirp1 has quit IRC | 04:27 | |
| pjz | I dunno how much you care about DRY | 04:27 |
| *** luke_ has joined #openstack | 04:33 | |
| *** kevnfx has joined #openstack | 04:36 | |
| *** coolhandluke has joined #openstack | 04:37 | |
| *** luke_ has quit IRC | 04:37 | |
| notmyname | pjz: we welcome patches | 04:48 |
| *** Cybodog has joined #openstack | 04:49 | |
| *** abecc has quit IRC | 05:00 | |
| *** joearnold has joined #openstack | 05:13 | |
| *** kevnfx has left #openstack | 05:17 | |
| *** pvo has quit IRC | 05:25 | |
| *** sneumann has joined #openstack | 05:25 | |
| *** zaitcev has quit IRC | 05:34 | |
| *** iammartian has left #openstack | 05:35 | |
| *** jsgotangco has joined #openstack | 05:39 | |
| *** pvo has joined #openstack | 05:40 | |
| *** guigui has joined #openstack | 05:43 | |
| *** pvo has quit IRC | 05:43 | |
| *** littleidea has joined #openstack | 05:47 | |
| *** rnirmal has quit IRC | 05:54 | |
| *** pharkmillups has quit IRC | 05:59 | |
| *** ibarrera has joined #openstack | 06:05 | |
| *** Orman has joined #openstack | 06:19 | |
| Orman | Does anyone know if OpenStack still needs documentation writers for documenting security or other stuff? | 06:20 |
| Orman | Anyone know? | 06:23 |
| soren | Orman: We can always use more people to write docs. | 06:26 |
| Orman | I am sort of a developer,but I develop security features. | 06:27 |
| Orman | Could I add documentation for security and include code along with it? | 06:27 |
| soren | Sure. | 06:28 |
| Orman | I am not the best coder,but I like designing features. | 06:28 |
| Orman | I have been blogging about Open Stack on my blog. | 06:29 |
| *** allsystemsarego has joined #openstack | 06:29 | |
| *** sophiap_ has joined #openstack | 06:30 | |
| *** jtimberman has quit IRC | 06:31 | |
| *** BK_man has quit IRC | 06:32 | |
| soren | Orman: Cool :) | 06:33 |
| Orman | So is everyone ready for the new release? I know I am counting down the days. | 06:34 |
| Orman | :D | 06:34 |
| *** jtimberman has joined #openstack | 06:34 | |
| *** sophiap has quit IRC | 06:34 | |
| *** sophiap_ is now known as sophiap | 06:34 | |
| soren | If we were ready, we'd release now :) | 06:36 |
| soren | ...but we're working on getting ready. | 06:36 |
| Orman | Ok I am just excited that's all. | 06:38 |
| Orman | Just glad to be making a different in the future of Cloud Computing. ;D | 06:39 |
| Orman | Hmm I see that from my view on the wiki that there is no security docs. Am I looking at the right page? | 06:41 |
| Orman | http://wiki.openstack.org/Documentation | 06:42 |
| soren | Yep. | 06:46 |
| *** coolhandluke has quit IRC | 06:47 | |
| Orman | Ok that makes me feel good. I think I can help with that. | 06:48 |
| jsgotangco | go Orman! | 06:49 |
| Orman | :) | 06:50 |
| Orman | Well security is were I love to focus my designing of new features and documenting security especially. ;D | 06:51 |
| *** coolhandluke has joined #openstack | 06:51 | |
| *** f4m8_ is now known as f4m8 | 06:53 | |
| Orman | Ok guys well I'm out for tonight. | 06:58 |
| Orman | God Bless you all and thanks for the warm welcome! | 06:58 |
| Orman | :D | 06:58 |
| *** Ryan_Lane has quit IRC | 07:00 | |
| *** Orman has quit IRC | 07:00 | |
| *** Ryan_Lane has joined #openstack | 07:02 | |
| *** joearnold has quit IRC | 07:04 | |
| *** khussein has joined #openstack | 07:09 | |
| *** coolhandluke has quit IRC | 07:10 | |
| *** Abd4llA has quit IRC | 07:10 | |
| *** joearnold has joined #openstack | 07:12 | |
| *** miclorb_ has quit IRC | 07:18 | |
| *** jsgotangco has quit IRC | 07:18 | |
| *** calavera has joined #openstack | 07:25 | |
| *** Abd4llA has joined #openstack | 07:32 | |
| *** coolhandluke has joined #openstack | 07:38 | |
| *** astrolin has joined #openstack | 07:52 | |
| *** befreax has joined #openstack | 07:53 | |
| *** befreax has quit IRC | 08:02 | |
| *** mgriffin_ has quit IRC | 08:09 | |
| *** mgriffin has joined #openstack | 08:09 | |
| *** khussein has quit IRC | 08:14 | |
| *** befreax has joined #openstack | 08:15 | |
| *** BK_man has joined #openstack | 08:21 | |
| zykes- | hmmm, libvirtd allready supports vmware | 08:28 |
| zykes- | can't one make openstack nova talk to vmware via libvirt ? | 08:28 |
| *** befreax has quit IRC | 08:30 | |
| *** miclorb has joined #openstack | 08:35 | |
| *** jonwood has joined #openstack | 08:38 | |
| *** befreax has joined #openstack | 08:44 | |
| soren | zykes-: Probably. | 08:45 |
| soren | zykes-: I, for one, just don't care much about VMWare. I'm one of these free software nuts :) | 08:46 |
| *** miclorb has quit IRC | 08:47 | |
| *** allsystemsarego has quit IRC | 08:49 | |
| * jonwood checks the time | 08:55 | |
| jonwood | Has the same conversation really been going on for 9 hours? | 08:55 |
| *** joearnold has quit IRC | 08:55 | |
| sneumann | I thkn there was a break somewhere :-) | 08:57 |
| *** allsystemsarego has joined #openstack | 09:02 | |
| zykes- | is there any place i should go to look on how to write a driver / backend for openstack towards vmware ? | 09:05 |
| *** sanjib has quit IRC | 09:06 | |
| *** miclorb has joined #openstack | 09:09 | |
| jonwood | zykes-: Have a look at nova/virt/xenapi.py in the source code. That seems to do a similar thing for XenServer. | 09:10 |
| *** coolhandluke has quit IRC | 09:10 | |
| jonwood | zykes-: Actually - it looks like libvirt supports vCenter | 09:11 |
| jonwood | http://libvirt.org/drvesx.html | 09:11 |
| jonwood | So you might be able to add support to the libvirt driver. | 09:12 |
| jonwood | However, this is largely conjecture based on a brief look at the source code. I'm not very familiar with nova yet. | 09:12 |
| *** irahgel has joined #openstack | 09:13 | |
| zykes- | soren: i too like opensource | 09:15 |
| zykes- | like setting up kvm == 10 mins | 09:15 |
| zykes- | but when my company has vmware allready it's not easy to do that :p | 09:15 |
| soren | 10 mins includes booting up and getting coffee. | 09:16 |
| soren | :) | 09:16 |
| zykes- | yeh :p | 09:16 |
| zykes- | and if you have redhat | 09:16 |
| zykes- | or centos it's even easier | 09:16 |
| zykes- | kickstart :o | 09:16 |
| soren | Yeah, but then you have a different set of problems :) | 09:18 |
| zykes- | how come ? ;) | 09:22 |
| soren | Because you've installed RedHat. | 09:46 |
| zykes- | hehe, Centos I primarely use | 09:49 |
| *** soosfarm_ is now known as soosfarm | 09:49 | |
| jonwood | CentOS/RedHat is improving. | 09:51 |
| jonwood | It doesn't make me want to jump out of 3rd floor windows any more at least. | 09:51 |
| *** ArdRigh has quit IRC | 10:08 | |
| *** irahgel has left #openstack | 10:11 | |
| *** ar1 has quit IRC | 10:24 | |
| *** coolhandluke has joined #openstack | 10:25 | |
| zykes- | i haven't done python before but | 10:41 |
| zykes- | gonna give it a try with the libvirt / vmware thing | 10:41 |
| *** irahgel has joined #openstack | 10:43 | |
| *** miclorb has quit IRC | 10:50 | |
| *** kashyapc has quit IRC | 10:57 | |
| soren | jonwood: Only 2nd floor now? :) | 10:59 |
| soren | Oh, /me forgot to have lunch.. | 10:59 |
| jonwood | Yeah, something like that ;) | 10:59 |
| jonwood | If you get a moment, can you have a look at the pages linked from http://wiki.openstack.org/NovaAdminManual? | 11:02 |
| jonwood | I'm trying to document configuration for Nova, but not really knowing it I'm working on assumptions made from reading the source. | 11:02 |
| soren | jonwood: That's what we all do :) | 11:03 |
| jonwood | I'm impressed by how easy the code is to read actually. I was expecting it to be much more complex then it is. | 11:04 |
| *** Kami_ has quit IRC | 11:10 | |
| *** Kami_ has joined #openstack | 11:10 | |
| *** astrolin has quit IRC | 11:24 | |
| *** gustavomzw has joined #openstack | 11:26 | |
| *** DubLo7 has quit IRC | 11:32 | |
| *** ctennis has quit IRC | 11:37 | |
| *** ctennis has joined #openstack | 11:37 | |
| *** ctennis has joined #openstack | 11:37 | |
| *** ctennis has quit IRC | 11:42 | |
| zykes- | how can I check out the src code ? | 11:52 |
| jonwood | http://wiki.openstack.org/GettingTheCode | 11:52 |
| soren | "bzr branch lp:nova" being the secret trick. | 11:54 |
| *** littleidea has quit IRC | 11:59 | |
| *** sugupta has joined #openstack | 12:07 | |
| coolhandluke | hi guys, I know y'all are busy with prep for the austin release.. I had a very quick question; we are testing out various components of nova; and we've got the installation to work on a single physical m/c controller + node(s) .... however, we still cant figure out a way to associate a compute node with a controller running on a separate physical machine.. | 12:09 |
| coolhandluke | briefly, how do we associate/register a compute node with an API/controller or vice versa? | 12:10 |
| jonwood | coolhandluke: It looks like it might work to have a central database and RabbitMQ instance, and then specify --rabbit_host and --sql_connection when starting up the daemons. | 12:18 |
| jonwood | I've not tried it though, I'm just working on the basis of available options. | 12:19 |
| *** sugupta has quit IRC | 12:20 | |
| soren | coolhandluke: jonwood is spot on. | 12:21 |
| soren | coolhandluke: The default is to use sqlite, because it's simple to get up an running. If you want multi-node deployment you need to set up a database (mysql or postgresql being obvious choices). | 12:22 |
| jonwood | soren: Does Nova create it's own database tables if they don't exist already? | 12:24 |
| soren | Yup. | 12:24 |
| coolhandluke | ok. thanks a lot for the pointers. jonwood: is some sort of documentation already in the making as far as multi-node deployment is concerned. We'll help out as much as we can. | 12:26 |
| jonwood | coolhandluke: There's the skeleton of a wiki page at http://wiki.openstack.org/NovaAdminManual | 12:26 |
| jonwood | I'm planning to write some documentation on getting Nova running on multiple nodes, but I'm flooded with client work, so I've only got an hour or so a day at the moment. | 12:27 |
| coolhandluke | jonwood: ok. i'll try this out and get back with updates. appreciate your help and contribution. | 12:29 |
| *** pvo has joined #openstack | 12:30 | |
| *** ChanServ sets mode: +v pvo | 12:30 | |
| *** pvo has quit IRC | 12:36 | |
| *** pvo has joined #openstack | 12:37 | |
| *** ChanServ sets mode: +v pvo | 12:37 | |
| *** perestrelka has quit IRC | 12:41 | |
| *** 17SAACUPS has joined #openstack | 12:50 | |
| *** aliguori has joined #openstack | 12:53 | |
| *** silassewell has joined #openstack | 12:53 | |
| soren | dendrobates: I needed to clear my head and do something different, so I whipped this up: http://wiki.openstack.org/Nova/EucalyptusFeatureComparison | 12:59 |
| *** perestrelka has joined #openstack | 13:00 | |
| *** kashyapc has joined #openstack | 13:04 | |
| dendrobates | soren: ooo | 13:07 |
| dendrobates | soren: wow, care if I share it? | 13:07 |
| *** coolhandluke has quit IRC | 13:12 | |
| *** abecc has joined #openstack | 13:15 | |
| *** abecc has joined #openstack | 13:15 | |
| *** coolhandluke has joined #openstack | 13:16 | |
| *** pvo has quit IRC | 13:20 | |
| *** pvo has joined #openstack | 13:21 | |
| *** ChanServ sets mode: +v pvo | 13:21 | |
| zykes- | hmms | 13:22 |
| zykes- | wonder if we could end up kicking vmware out the door | 13:23 |
| jonwood | Is it really that bad? | 13:23 |
| Ryan_Lane | features needed for that: storage vmotion, HA, fault tolerance, a web or fat client interface, a bunch of others :) | 13:24 |
| Ryan_Lane | I'd love to see it happen though | 13:25 |
| Ryan_Lane | jonwood: vmware requires windows server, and oracle or mssql. it's insanely expensive, and is very proprietary. even the management interface is windows only | 13:26 |
| jonwood | Ryan_Lane: I was talking about adding support for VMWare. | 13:26 |
| Ryan_Lane | :D | 13:26 |
| jonwood | But it doesn't require Windows server. ESXi and above run on their own. | 13:26 |
| zykes- | jonwood: my problem is that i don't know a bit of python | 13:27 |
| Ryan_Lane | jonwood: yeah. if you don't want any of the clustered features | 13:27 |
| zykes- | and my company is allready using vCenter etc | 13:27 |
| Ryan_Lane | what benefits does using vmware give over using kvm? | 13:27 |
| jonwood | Management have a warm fuzzy feeling because they paid thousands of dollars for their software. | 13:28 |
| Ryan_Lane | ah | 13:28 |
| jonwood | There may be technical advantages as well, but whenever I see companies using things like VMWare, that's generally the actual reason it got put in place. | 13:29 |
| Ryan_Lane | there are a bunch of features in vmware that don't exist elsewhere, or don't work as well | 13:29 |
| Ryan_Lane | also, it's insanely easy to use | 13:30 |
| piken | Ryan_Lane: VMware support infiniband and fabric networks. | 13:31 |
| piken | Main reason a lot of large companies use it for cloud infrastructure as it has 40GBps through put | 13:31 |
| zykes- | piken: i guess linux / xenserver does that as well ? :o | 13:31 |
| piken | zykes-: no, only vmware and intalio | 13:32 |
| piken | issue I am running into here. | 13:32 |
| Ryan_Lane | is infiniband supported in 4? I know it wasn't in 3.5 | 13:32 |
| piken | there is a test patch for it in 4, but it is not in the offical tree | 13:32 |
| Ryan_Lane | I don't see why openstack couldn't support infiniband, at least using ip over infiniband | 13:32 |
| piken | That is the point, intalio and VMware don't use ipoib, they use sdp. | 13:33 |
| Ryan_Lane | ah | 13:33 |
| piken | performance is more then 4 times better :( | 13:33 |
| Ryan_Lane | yeah | 13:33 |
| piken | with ipoib you need a dedicated router for traffic in the fabric | 13:33 |
| *** HouseAway is now known as AimanA | 13:33 | |
| Ryan_Lane | well, unless everything is in the same subnet, right? | 13:34 |
| piken | There are no subnets as defined by ethernet. | 13:34 |
| piken | fabric systems use system hardware id's and switch translation | 13:34 |
| piken | That is where the pain comes in as most software does not support that. | 13:35 |
| zykes- | Ryan_Lane: but then again i guess the tools to administer things aren't better then what we make them together? | 13:40 |
| Ryan_Lane | well, there is only command line tools right now, if you're talking about openstack ;) | 13:42 |
| Ryan_Lane | I know people are working on web and other tools, and I'm looking forward to it | 13:42 |
| *** jdarcy has joined #openstack | 13:43 | |
| *** f4m8 is now known as f4m8_ | 13:50 | |
| *** cdub has joined #openstack | 13:52 | |
| *** allsystemsarego has quit IRC | 13:58 | |
| *** xfaf has joined #openstack | 13:59 | |
| *** zul has quit IRC | 13:59 | |
| *** xfaf is now known as zul | 13:59 | |
| *** zul has joined #openstack | 14:00 | |
| *** kevnfx has joined #openstack | 14:00 | |
| *** coolhandluke has quit IRC | 14:01 | |
| *** ___mattf is now known as mattf | 14:02 | |
| *** ppetraki has joined #openstack | 14:03 | |
| *** sirp1 has joined #openstack | 14:04 | |
| *** sirp1 has left #openstack | 14:09 | |
| *** Podilarius has joined #openstack | 14:13 | |
| *** cdub has quit IRC | 14:18 | |
| *** cdub has joined #openstack | 14:20 | |
| *** pvo_ has joined #openstack | 14:20 | |
| *** ChanServ sets mode: +v pvo_ | 14:20 | |
| *** sugupta has joined #openstack | 14:22 | |
| *** ecarlin has joined #openstack | 14:23 | |
| *** pvo has quit IRC | 14:24 | |
| *** pvo_ is now known as pvo | 14:24 | |
| *** ecarlin has quit IRC | 14:25 | |
| *** gundlach has joined #openstack | 14:27 | |
| *** kevnfx has quit IRC | 14:27 | |
| *** daleolds has quit IRC | 14:28 | |
| *** sugupta has quit IRC | 14:30 | |
| *** guigui has quit IRC | 14:31 | |
| *** befreax has quit IRC | 14:36 | |
| *** pharkmillups has joined #openstack | 14:40 | |
| *** kakoni has joined #openstack | 14:42 | |
| *** littleidea has joined #openstack | 14:43 | |
| *** pvo has quit IRC | 14:44 | |
| kakoni | swift and authserver.. so theres atleast the devauth and ive seen traces of swauth. What is nova auth? | 14:47 |
| jonwood | kakoni: I believe it's authentication for Nova API access. | 14:48 |
| *** rnirmal has joined #openstack | 14:50 | |
| kakoni | but only for nova, not for swift? | 14:53 |
| soren | dendrobates: No, please do. That's why I put it there. | 14:54 |
| kakoni | also, could it be that devauth server (on swift) could become performance bottleneck? (when evaluating swift..) | 14:54 |
| *** kevnfx has joined #openstack | 14:55 | |
| notmyname | kakoni: probably not. the auth tokens are cached im memcached | 14:55 |
| jonwood | kakoni: Yes, only for Nova. | 14:55 |
| *** sophiap has quit IRC | 14:56 | |
| soren | Has anyone given any thought to consolidating user databases between Swift and Nova? | 14:56 |
| notmyname | kakoni: while we don't recommend using the devauth in a production swift cluster, it shouldn't cause any performance issues for evaluation purposes | 14:56 |
| notmyname | soren: could be an interesting idea, but we don't really have a user database in swift | 14:57 |
| notmyname | soren: the swauth server has the concept of users, but, if I recall correctly, any user info that needs to be stored is stored in the cluster and found via ring lookup | 14:59 |
| *** sirp1 has joined #openstack | 14:59 | |
| *** khussein has joined #openstack | 14:59 | |
| *** kevnfx has quit IRC | 15:00 | |
| notmyname | but I'm a little fuzzy on those details. I've not looked at it in a while | 15:00 |
| *** khussein has quit IRC | 15:00 | |
| notmyname | soren: actually, I think it would be a great idea to have some unified concept of users across all openstack projects | 15:01 |
| kakoni | so howabout swauth, is it being actively developed or whats the current status? | 15:01 |
| notmyname | swauth is being worked on, but a few other higher priority things have come up | 15:02 |
| *** sophiap has joined #openstack | 15:02 | |
| notmyname | gholt is the authority on swauth | 15:03 |
| notmyname | but he was online a 4:15 this morning, so I don't really expect him to chime in here now :-) | 15:04 |
| *** hazmat has joined #openstack | 15:10 | |
| *** jbryce has joined #openstack | 15:12 | |
| kakoni | So anyway, heres my current situation; 8 node cluster, 3 nodes as proxies/auth servers, rest of them as object storages. loadbalancer in front of the proxy nodes. | 15:13 |
| kakoni | Ran few simple tsung tests and managed to see performance like 75reqs/sec for 10kB file uploads | 15:14 |
| kakoni | anyway, about 10% of the requests are failing due to a 503 error.. | 15:14 |
| kakoni | (lots of connection timeout issues in proxy server logs) | 15:14 |
| kakoni | any...pointers?;) | 15:14 |
| *** BK_man has quit IRC | 15:16 | |
| kakoni | ERROR with Object server xxxx transaction tx157b626b-c46e-4eb7-9a96-c080c75ba9d7 re: Trying to get final status of PUT to /v1/AUTH_de235fbface04f13b15b45a7e9527d79/foobar/mjgyhxqpnigbmusxqich: Timeout (10s) | 15:16 |
| *** BK_man has joined #openstack | 15:19 | |
| *** ecarlin has joined #openstack | 15:19 | |
| *** BK_man has left #openstack | 15:19 | |
| *** gaveen has joined #openstack | 15:25 | |
| creiht | soren: I think xtoddx has added support for nova auth to be able to handle swift auth as well | 15:26 |
| *** dysinger has joined #openstack | 15:26 | |
| *** sneumann has quit IRC | 15:27 | |
| creiht | kakoni: how many disks per storage node, size, type? | 15:29 |
| creiht | kakoni: how many cores on each node? | 15:29 |
| creiht | kakoni: and what type of networking (1G, 10G)> | 15:29 |
| creiht | ? | 15:29 |
| *** 17SAACUPS has quit IRC | 15:29 | |
| *** scottie has joined #openstack | 15:30 | |
| kakoni | creicht: crappy 32bit servers;) with 4 cores. 1 HDD, storage node has its own swift partition. XFS partioning, about 120Gb worth of storage per node. 1Gig network. | 15:32 |
| creiht | ok | 15:32 |
| kakoni | set the worker conf to 2 | 15:32 |
| creiht | how many workers in the config? | 15:32 |
| creiht | ok | 15:32 |
| kakoni | i dont know if you can comment or not, but i dont know wether that 75reqs/sec is seen as good,bad or average figure;) | 15:33 |
| creiht | hehe | 15:33 |
| creiht | I can get the same on a cloud server running the swift all in one :) | 15:33 |
| creiht | so we should be able to do better | 15:34 |
| creiht | kakoni: what does cpu usage look like on the storage nodes? | 15:34 |
| creiht | I would suggest setting workers for all servers to 4 | 15:35 |
| creiht | to start with | 15:35 |
| *** ccustine has joined #openstack | 15:35 | |
| kakoni | under heavy tsung load about 50% | 15:35 |
| kakoni | ok, and probably putting 10gig conns to proxy servers would help. | 15:36 |
| creiht | :) | 15:36 |
| *** ecarlin has quit IRC | 15:36 | |
| kakoni | we've been also playing with the idea of introducing varnish/nginx in front of the proxies. (or perhaps get rid of proxy server complely and implement the same functionality inside nginx/varnish...) | 15:36 |
| *** ecarlin has joined #openstack | 15:36 | |
| creiht | kakoni: do you see a lot of those timeouts? or just intermittent? | 15:37 |
| notmyname | kakoni: I wouldn't recommend that.... | 15:37 |
| creiht | hehe | 15:37 |
| kakoni | creiht: under very heavy load, yes lots of these timeouts | 15:37 |
| kakoni | notmyname: ok, now why is that?;) | 15:37 |
| creiht | notmyname: you should just write up a faq :) | 15:38 |
| notmyname | kakoni: we looked at nginx to be our proxy servers. great server, but it spools data to the disk on uploads. that's very bad for storage systems that accept lots of uploads | 15:38 |
| kakoni | notmyname: ok. how about varnish? | 15:38 |
| creiht | now the idea of some sort of hot cache in front of swift for reads wouldn't be a bad idea for some use cases | 15:39 |
| notmyname | creiht: actually, if you get kakoni's stuff working much better, and the discussion is all here in the open, it would be a great start to such a FAQ | 15:39 |
| creiht | hehe | 15:39 |
| creiht | this is why I want to get a "refrence cluster" set up asap | 15:39 |
| creiht | kakoni: at what concurrency are you sending those uploads? | 15:40 |
| kakoni | I guess it ended up about 1000 concurrent session | 15:40 |
| kakoni | ended up with.. | 15:40 |
| *** cdub has quit IRC | 15:40 | |
| notmyname | kakoni: what creiht just said. varnish would be ok in certain use cases. essentially, a varnish cache in from of swift is exactly the same as the CDN in front of cloud files, minus the geographic dispersion | 15:40 |
| creiht | kakoni: and btw, I may be in and out a bit today as there is a lot going on today | 15:40 |
| creiht | kakoni: ok, so my guess is you are overloading what the single disk on each machine can do | 15:41 |
| kakoni | seems so. need to get better hw;) | 15:41 |
| creiht | and actually since you only have one drive, I would put workers more than 2 | 15:41 |
| creiht | wouldn't | 15:41 |
| creiht | that is | 15:41 |
| *** calavera has quit IRC | 15:42 | |
| AimanA | they have the next miner almost up out of the bore | 15:42 |
| AimanA | oops, wrong chan | 15:42 |
| creiht | in our testing setup, we were able to get about 2k puts/second for small puts per proxy | 15:42 |
| creiht | on a machine that had 2 quad core midlle of the road opterons | 15:43 |
| creiht | the storage nodes had 24 spindles, and same procs | 15:43 |
| creiht | 16 workers on the proxy, 8 on the storage nodes for each server | 15:43 |
| creiht | (to give you an idea of what type of performance should be possible) | 15:44 |
| kakoni | oh. cool. thanks | 15:44 |
| creiht | kakoni: and also the latest swift code has a swift-bench utility | 15:44 |
| kakoni | yep, been using it aswell. But u know with tsung I can get better...loads;9 | 15:45 |
| creiht | so that when people are benchmarking, we can be using the same thing | 15:45 |
| kakoni | oh | 15:45 |
| creiht | kakoni: that wouldn't surprise me :) | 15:45 |
| kakoni | didnt thought of that. makes sense. so from know one will be using swift-bench only. | 15:46 |
| creiht | I'm curious how far off it is though | 15:46 |
| creiht | still probably isn't a bad idea to use both | 15:46 |
| kakoni | (atleast when communicating with the mothership;) | 15:46 |
| creiht | yeah | 15:46 |
| *** jaypipes is now known as jaypipes-afk | 15:46 | |
| kakoni | theres also the other distributed storage stack that we've been following, elliptics network. They've shown some impressive numbers lately. | 15:47 |
| creiht | interesting... hadn't heard of that one yet | 15:48 |
| kakoni | http://www.ioremap.net/ | 15:48 |
| creiht | yeah looking now | 15:48 |
| notmyname | seems that he's claiming 20K requests per minute now with 10 proxies. that sounds low | 15:49 |
| notmyname | swift is...swifter | 15:50 |
| creiht | hah | 15:50 |
| creiht | GETs are easy | 15:50 |
| kakoni | " 5000 rps of random io requests in 1 tb of data (100 millions of objects)" | 15:50 |
| creiht | only 1tb? | 15:51 |
| kakoni | 2 servers with SAS shelves. | 15:51 |
| creiht | :) | 15:51 |
| kakoni | hah;) | 15:51 |
| creiht | yeah come back when you have some real scale :) | 15:51 |
| kakoni | anyway, need to dropoff now, thanks guys for the feedback, I'll tweak my cluster and lets see what happens. | 15:51 |
| creiht | kakoni: cool... keep us informed how it goes | 15:51 |
| kakoni | i will. alrighty.bye | 15:52 |
| *** kakoni has quit IRC | 15:52 | |
| creiht | well that gives me even more motivation now to get a reference cluster set up | 15:52 |
| *** mmalgeri has joined #openstack | 15:57 | |
| *** ecarlin has quit IRC | 15:58 | |
| *** Orman has joined #openstack | 16:00 | |
| Orman | annegentle: I just emailed you. | 16:02 |
| annegentle | Orman: ok, got it, thanks | 16:03 |
| Orman | Sure thing. Just thought I'd let you know. | 16:03 |
| *** pvo has joined #openstack | 16:04 | |
| *** ChanServ sets mode: +v pvo | 16:04 | |
| *** pvo has quit IRC | 16:05 | |
| *** pvo has joined #openstack | 16:05 | |
| *** ChanServ sets mode: +v pvo | 16:05 | |
| notmyname | Orman: call her to tell her that you sent a message in IRC that you sent her an email ;-) | 16:06 |
| Orman | Yeah | 16:08 |
| annegentle | any BZRers around? I'm not sure how to commit a fix that doesn't have a bug assigned to it, it's just spelling changes in .rst files. I guess I can log a bug if that's the only way? | 16:09 |
| creiht | annegentle: you should still be able to push your branch like you would before, and request a merge proposal | 16:10 |
| Orman | I agree with that. ;D | 16:11 |
| *** daleolds has joined #openstack | 16:11 | |
| creiht | annegentle: and you don't have to have a --fixes on your commit | 16:12 |
| *** khussein has joined #openstack | 16:12 | |
| vishy | rlucio: I'm here | 16:13 |
| annegentle | creiht: is it the --fixes that is telling it... ah can't type fast enough :) | 16:13 |
| vishy | soren: are you here? | 16:13 |
| dendrobates | I think it's dinner time in DK | 16:14 |
| Orman | Have a good dinner dendrobates. ;D | 16:15 |
| Orman | Lol | 16:15 |
| Orman | jk | 16:16 |
| *** khussein has quit IRC | 16:16 | |
| *** sandywalsh has quit IRC | 16:17 | |
| Orman | A lot to do before the release. We are moving towards our goal though. | 16:18 |
| dendrobates | it's almost lunch time for me | 16:18 |
| Orman | Oh ok. lol :D | 16:19 |
| dendrobates | Orman: I don;t recognize your nick. are you new? | 16:19 |
| Orman | Yes, I am. | 16:19 |
| vishy | My branches should merge ok now | 16:19 |
| dendrobates | vishy: cool thanks | 16:20 |
| vishy | I'm working on the fix-deprecation-warning branch as well, there are few more fixes that need to be backported/updated for new code | 16:20 |
| Orman | I'm a security dev,bug fixer, and I am just starting doc writing. ;D | 16:21 |
| *** faccenda has joined #openstack | 16:22 | |
| dendrobates | Orman: great to have you aboard | 16:24 |
| dendrobates | Orman: do you work for a cloud computing company? | 16:24 |
| Orman | No I am a College student in Florida. Trying to build my own Cloud Computing Security company. | 16:25 |
| dendrobates | where in florida? | 16:27 |
| dendrobates | I used to live in Gainesville | 16:27 |
| Orman | I used to live their to. Well High Springs really. | 16:28 |
| *** littleidea has quit IRC | 16:28 | |
| *** ElectricBill has quit IRC | 16:28 | |
| Orman | I live around SpringHill,FL now. | 16:29 |
| *** sandywalsh has joined #openstack | 16:30 | |
| *** daleolds has quit IRC | 16:32 | |
| *** daleolds has joined #openstack | 16:32 | |
| *** sirp1 has quit IRC | 16:33 | |
| dendrobates | soren: test failure for lp:~soren/nova/xen | 16:33 |
| *** kashyapc has quit IRC | 16:34 | |
| *** littleidea has joined #openstack | 16:35 | |
| Orman | dendrobates: I like around the New Port Richey area or close to it. | 16:36 |
| *** pvo has quit IRC | 16:40 | |
| uvirtbot | New bug: #660042 in nova "Better error when no virtualization method is available on the server" [Undecided,New] https://launchpad.net/bugs/660042 | 16:41 |
| *** sirp1 has joined #openstack | 16:42 | |
| Orman | Really? Hmm I'll have to look at it. | 16:42 |
| chmouel | what version of euca2ools is supposed to be supported by nova? | 16:44 |
| chmouel | i am getting this when trying to terminate instance http://paste.openstack.org/show/53/ | 16:47 |
| chmouel | this is on vanilla maverick with maverick packages | 16:47 |
| dendrobates | chmouel: vanilla maverick should work, file a bug | 16:48 |
| chmouel | dendrobates: cool. | 16:48 |
| *** jaypipes-afk is now known as jaypipes | 16:48 | |
| Orman | Yeah I agree with dendrobates. Vanilla should work. | 16:49 |
| dendrobates | chmouel: was this using the maverick packages or trunk? | 16:49 |
| chmouel | dendrobates: maverick packages | 16:49 |
| dendrobates | make sure it is broken in trunk before adding the bug | 16:49 |
| chmouel | dendrobates: cool or i guess i'll fill the bug in ubuntu | 16:50 |
| *** kashyapc has joined #openstack | 16:50 | |
| dendrobates | use vish's script to install from trunk. | 16:50 |
| dendrobates | it works great | 16:50 |
| *** burris has quit IRC | 16:50 | |
| chmouel | yeah i havent tried that do you have that github url handy? | 16:51 |
| jaypipes | chmouel: http://github.com/vishvananda/novascript | 16:52 |
| Orman | It's an Ec2 error. Now that's interesting too. | 16:52 |
| chmouel | cheers | 16:52 |
| chmouel | i get that on most of the tools | 16:52 |
| chmouel | and the method are indeed different | 16:52 |
| chmouel | http://paste.openstack.org/show/54/ | 16:53 |
| Orman | Ok wow hmm. | 16:53 |
| jaypipes | chmouel: ack-grep ftw ;) | 16:54 |
| chmouel | jaypipes: yeah i usually use that but that's a brand new install :) | 16:54 |
| jaypipes | chmouel: :) | 16:54 |
| *** rlucio has joined #openstack | 16:54 | |
| *** burris has joined #openstack | 16:54 | |
| jonwood | chmouel, dendrobates: soren mentioned that's a known bug the other day. | 16:57 |
| jonwood | It's fixed in trunk apparently | 16:57 |
| chmouel | jonwood: sweet i am about to test trunk... | 16:57 |
| *** daleolds has quit IRC | 16:59 | |
| *** daleolds has joined #openstack | 16:59 | |
| *** ibarrera has quit IRC | 17:00 | |
| *** khussein has joined #openstack | 17:00 | |
| *** scottie has quit IRC | 17:02 | |
| zykes- | what distributions come along with xen easily anymore ? | 17:02 |
| zul | opensuse | 17:05 |
| zul | debian | 17:05 |
| *** sneumann has joined #openstack | 17:06 | |
| *** daleolds has quit IRC | 17:08 | |
| *** kevnfx has joined #openstack | 17:08 | |
| *** alekibango has joined #openstack | 17:08 | |
| *** rlucio has quit IRC | 17:08 | |
| *** littleidea has quit IRC | 17:18 | |
| vishy | chmouel: it looks like you passed in an image_id | 17:19 |
| vishy | instance_ids start with i- | 17:19 |
| *** jc_smith has joined #openstack | 17:20 | |
| *** daleolds has joined #openstack | 17:21 | |
| vishy | chmouel: the method defs are different, boto converts to xml, and we convert back from xml | 17:21 |
| *** pharkmillups has quit IRC | 17:27 | |
| *** Ryan_Lane is now known as Ryan_Lane|foo | 17:41 | |
| *** Ryan_Lane|foo is now known as Ryan_Lane|food | 17:41 | |
| *** greenisus has joined #openstack | 17:41 | |
| *** jonwood has quit IRC | 17:45 | |
| *** jonwood has joined #openstack | 17:46 | |
| *** Ryan_Lane|food has quit IRC | 17:49 | |
| greenisus | does anyone mind helping me out with getting the OpenStack API in Nova to work? | 17:57 |
| gundlach | greenisus: what specifically are you looking at? | 17:59 |
| gundlach | i wrote some of the openstack API. | 17:59 |
| greenisus | gundlach: great! | 17:59 |
| greenisus | gundlach: so, i'm trying to GET http://my-ip-address:8773/v1.0 and i get a 404 back | 18:00 |
| greenisus | gundlach: i installed nova on ubuntu maverick on a rackspace cloud server with vish's install script on github | 18:00 |
| gundlach | try running nova-api again with --FAKE_subdomain=api and see what happens | 18:00 |
| greenisus | gundlach: i changed the code to set FAKE_subdomain to api and it does the same | 18:00 |
| gundlach | hmm, stop being a step ahead of me :) | 18:01 |
| gundlach | how about / -- does that give a 404? | 18:01 |
| greenisus | hehe, no, / gives me a list of versions | 18:01 |
| greenisus | consistent with nova/api/__init__.py | 18:01 |
| *** rlucio has joined #openstack | 18:02 | |
| rlucio | are there any other new deps besides gawk and ebtables since the ec2-security stuff was merged? I installed them but am still unable to launch instances, and still get the 'Could not get access to ACL tech driver 'ebiptables'' error | 18:02 |
| gundlach | greenisus: does /v1.0/ with a slash appended do anything? | 18:02 |
| greenisus | gundlach: whoa, this is weird | 18:03 |
| greenisus | okay, i killed nova-api and restarted with the --FAKE_subdomain flag | 18:03 |
| greenisus | and now it seems to work | 18:03 |
| greenisus | needs /v1.0/ instead of /v1.0, but before i would always get 404 | 18:03 |
| rlucio | greenisus: yea that is a workaround for ec2 iirc | 18:04 |
| gundlach | greenisus: ok, glad it started working for you :) | 18:04 |
| greenisus | rlucio: that's no big deal... just weird that setting FAKE_subdomain in code didn't work for me... | 18:04 |
| greenisus | great! thanks so much guys! | 18:05 |
| greenisus | one last question and i'll leave you alone :) | 18:05 |
| greenisus | will any username/api key work? | 18:05 |
| greenisus | or do i need to make a user or something | 18:05 |
| *** localhost has quit IRC | 18:06 | |
| greenisus | well, looks like it needs something | 18:06 |
| greenisus | i can grep the code | 18:06 |
| gundlach | greenisus: hmm | 18:06 |
| greenisus | don't want to bug you guys | 18:06 |
| gundlach | greenisus no problem -- until we have good docs, we deserve any "bugging" people want to give us :) | 18:06 |
| greenisus | haha, yeah i'm helping anne out to get the docs in better shape | 18:06 |
| gundlach | greenisus: yes, you need to use i believe nova-manage to create a user (and maybe a project)? | 18:06 |
| greenisus | and i'll start documenting/blueprinting the new iOS app i'm working on | 18:06 |
| gundlach | that will put entries into fakeldap that you can then use to hit the API | 18:07 |
| *** localhost has joined #openstack | 18:07 | |
| gundlach | greenisus: oh, cool! talk to mike mayo if you haven't yet -- he has built an iPhone/iPad app that is probably related to what you're working on. | 18:07 |
| _cerberus_ | haha | 18:07 |
| _cerberus_ | He *is* Mike Mayo :-P | 18:07 |
| greenisus | gundlach: lol, i am mike mayo :) | 18:07 |
| gundlach | greenisus: oh, even cooler! make sure to talk with yourself then | 18:08 |
| gundlach | :) | 18:08 |
| greenisus | gundlach: i promise i will :) | 18:08 |
| * gundlach is just about fed up with learning 3 names for every coworker ;) | 18:08 | |
| greenisus | yeah seriously. who are you? | 18:08 |
| *** khussein has joined #openstack | 18:08 | |
| gundlach | *i* chose an irc handle that corresponds to my last name :) i'm michael; I introduced myself to you at the SFO meetup (i wrote the groundcontrol extension) | 18:09 |
| greenisus | ah! sweet! | 18:09 |
| greenisus | gundlach: hate that i didn't get to hang with you guys more | 18:10 |
| gundlach | maybe irc needs headshots :) | 18:10 |
| *** kevnfx has quit IRC | 18:10 | |
| gundlach | greenisus: maybe we'll get to hang out more at the summit; are you coming? | 18:11 |
| greenisus | that would be nice... okay, so i created a user with nova-manage and it gave me back EC2 creds. do i use one of those as the API key? | 18:11 |
| uvirtbot | New bug: #660095 in nova "OpenStack "create instance" API call broken after r138" [High,In progress] https://launchpad.net/bugs/660095 | 18:11 |
| greenisus | gundlach: yes, i'll be there, but i'll miss the two hack days because i'm getting married nov 13 | 18:11 |
| gundlach | greenisus: hey, congratulations! :) | 18:11 |
| greenisus | gundlach: thanks!! | 18:12 |
| gundlach | _cerberus_: do you remember how to get the API key for use in the OpenStack API? | 18:12 |
| gundlach | greenisus: i forgot that nova-manage gives you an EC2 key | 18:12 |
| vishy | rlucio: you need the bleeding edge version of libvirt from the ppa | 18:13 |
| vishy | HEAD SHOT! | 18:15 |
| gundlach | greenisus: out of curiosity, does the EC2 ACCESS_KEY and SECRET_KEY work when you pass them in with OpenStack API requests? | 18:15 |
| rlucio | vishy: i actually installed that too, at least im pretty sure i did | 18:15 |
| gundlach | vishy: yeah, that kind of head shot would make irc more fun too | 18:15 |
| vishy | rlucio: I'm pretty sure that error is due to an old version of libvirt | 18:16 |
| vishy | dpkg -l | grep libvirt | 18:16 |
| greenisus | gundlach: no it gave me a 500 error | 18:17 |
| greenisus | then i restarted nova-api and it's not authenticating with either EC2 key | 18:17 |
| gundlach | greenisus: well that isn't very helpful | 18:17 |
| *** khussein has quit IRC | 18:17 | |
| rlucio | yea i have the new version | 18:17 |
| gundlach | greenisus: lemme go dig for a bit | 18:17 |
| *** dysinger has quit IRC | 18:18 | |
| greenisus | gundlach: yeah, haha, i get a 500 error with the access key, and nothing with the secret key | 18:18 |
| greenisus | gundlach: makes me think the access key is the way to go and i have another problem :) | 18:18 |
| greenisus | gundlach: i get a stack trace that ends with "TypeError: 'User' object is unsubscriptable" | 18:18 |
| jaypipes | mtaylor: around? | 18:19 |
| gundlach | greenisus: would you mind pasting that into pastebin? | 18:19 |
| gundlach | [that stacktrace] | 18:19 |
| greenisus | gundlach: sure, here you go: http://pastebin.com/2q2sG1K9 thanks a lot for the help! | 18:20 |
| gundlach | sure, thanks for the debugging ;) | 18:20 |
| *** Ryan_Lane has joined #openstack | 18:21 | |
| *** gaveen has quit IRC | 18:21 | |
| mtaylor | jaypipes: always | 18:22 |
| *** ecarlin has joined #openstack | 18:23 | |
| gundlach | greenisus: did you happen to override --auth-driver? | 18:23 |
| jaypipes | mtaylor: :) hi. I'd like to add something to our Tarmac job. Can we add a run of nosetests within /nova/tests/api/ please? | 18:24 |
| greenisus | gundlach: no, i'm running unmodified code from trunk yesterday | 18:24 |
| jaypipes | mtaylor: we've had a number of breakages of the RS API stuff get into trunk b/c Hudson only runs the "regular" test runner. | 18:24 |
| mtaylor | jaypipes: hrm. I don't see why not - I also don't see why nosetests isn't finding the API tests | 18:25 |
| jaypipes | mtaylor: you can't run nosetests in /nova/tests/. It works only in /nova/tests/api/ | 18:26 |
| rlucio | vishy: yea, i have gawk, ebtables, iptables, and the new libvirt stuff from the ppa and am still getting the same error | 18:26 |
| mtaylor | jaypipes: oh. right. nose still is broken in nova. gotcha | 18:27 |
| mtaylor | jaypipes: any chance that's ever gonna get fixed? :) | 18:27 |
| *** daleolds has quit IRC | 18:27 | |
| jaypipes | mtaylor: bexar. | 18:27 |
| vishy | rlucio: you checked it with dpkg -l ? | 18:27 |
| rlucio | 0.8.3-1ubuntu14~ppalucid2 for libvirt, libvirt0, libvirt-bin | 18:28 |
| rlucio | err, python-libvirt | 18:28 |
| *** hornbeck has quit IRC | 18:29 | |
| vishy | k | 18:29 |
| vishy | that is right | 18:29 |
| vishy | did you restart libvirt | 18:29 |
| vishy | after installing gawk and ebtables? | 18:29 |
| vishy | and restart nova | 18:30 |
| rlucio | i assumed libvirt restarted as part of the deb, but let me do that manually, and then try again | 18:31 |
| rlucio | vishy: uh-oh, i bet that's the issue. assuming things ftl | 18:32 |
| *** jfluhmann has joined #openstack | 18:32 | |
| vishy | libvirt should have a dependency on those two packages if it is using them imo | 18:32 |
| vishy | but it doesn't | 18:33 |
| vishy | :) | 18:33 |
| rlucio | vishy: arrgg you win, that did the trick, thanks :) | 18:33 |
| vishy | WINNAR! | 18:33 |
| *** gaveen has joined #openstack | 18:33 | |
| gundlach | greenisus: yep, looks like there's a bug in api/openstack/auth.py; i'm working on a branch. | 18:34 |
| gundlach | greenisus: if it helps you continue until i get the bug fixed, you can disable auth entirely by removing AuthMiddleware from api/openstack/__init__.py line 54. | 18:35 |
| soren | dendrobates: oh, ffs. | 18:35 |
| greenisus | gundlach: thanks! | 18:36 |
| vishy | soren: I assume the insane amount of physdev warnings in messages is related to the bug you filed? | 18:37 |
| soren | vishy: Sort of related. | 18:38 |
| vishy | soren: is there an easy way to squelch the errors? | 18:39 |
| soren | vishy: Don't look. | 18:39 |
| soren | :( | 18:39 |
| *** ecarlin has quit IRC | 18:39 | |
| vishy | :P | 18:39 |
| soren | Other than that, no. | 18:39 |
| vishy | they are filling up the logs under heavy load :) | 18:39 |
| soren | It's /extremely/ annoying. | 18:39 |
| vishy | i suppose i can add a filter rule to syslog-ng | 18:39 |
| soren | Oh, you can filter them in syslog. | 18:39 |
| soren | Right. | 18:39 |
| soren | You just can't stop the kernel from emitting them. | 18:39 |
| *** littleidea has joined #openstack | 18:39 | |
| soren | It only emits them, though, when new filters are added. | 18:40 |
| *** kashyapc has quit IRC | 18:42 | |
| greenisus | gundlach: so just comment out line 54? | 18:42 |
| gundlach | greenisus: no. replace line 54 with this: | 18:42 |
| gundlach | app = RateLimitingMiddleware(APIRouter()) | 18:43 |
| gundlach | line 53, sorry | 18:43 |
| greenisus | gundlach: okay thanks man! | 18:43 |
| gundlach | np :) | 18:43 |
| uvirtbot | New bug: #660115 in nova "OpenStack API auth manager not processing Users/Tokens correctly" [Critical,In progress] https://launchpad.net/bugs/660115 | 18:45 |
| *** electroniceagle has joined #openstack | 18:55 | |
| soren | jaypipes: Did you just retry my Xen branch when it failed? | 18:58 |
| *** daleolds has joined #openstack | 18:59 | |
| jaypipes | soren: yeah. was seeing if it was just a testing oops. | 19:01 |
| soren | jaypipes: Ok, that explains. Thanks. | 19:01 |
| *** kashyapc has joined #openstack | 19:02 | |
| *** gustavomzw has quit IRC | 19:03 | |
| greenisus | gundlach: it works! auth gives a 404, but other stuff works, so thanks! | 19:09 |
| gundlach | greenisus: glad to hear it. i'm working thru the auth bugs now. | 19:09 |
| Orman | Auth bugs. | 19:10 |
| greenisus | gundlach: another oddity; i have to pass X-Auth-User and X-Auth-Key with every request | 19:10 |
| gundlach | greenisus: even after removing the AuthMiddleware? | 19:11 |
| greenisus | gundlach: typically, you only pass those to the auth API, and then pass X-Auth-Token to all subsequent requests | 19:11 |
| gundlach | right | 19:11 |
| greenisus | gundlach: yeah, if i pass nothing or the token, i get a 500 error | 19:11 |
| gundlach | OK, I'll look into that too, thanks :) | 19:11 |
| gundlach | I believe the code was written intending to use X-Auth-Token after the 1st request, but something's obviously going wrong. | 19:11 |
| greenisus | gundlach: makes sense, also GET /servers blows up with KeyError: 'nova.context' | 19:12 |
| rlucio | vishy: for flat networking, do i still have to delete all my old entries (or maybe drop the table) in order to get nova-network to update with a new private_range? | 19:12 |
| gundlach | greenisus: that *does* make sense, because AuthMiddleware is supposed to create that :) | 19:12 |
| greenisus | gundlach: but that could very well be my fault in not having a complete setup | 19:12 |
| greenisus | gundlach: after all, i have no images | 19:13 |
| gundlach | greenisus: no, if you had AuthMiddleware back in place, nova.context would be created and you shouldn't see that error. | 19:13 |
| greenisus | gotcha | 19:13 |
| gundlach | one of AuthMiddleware's jobs is to create the nova.context container. | 19:13 |
| Orman | Image progress bug is still being fixed right? | 19:14 |
| greenisus | gundlach: okay, so i guess can you ping me when the auth bug is fixed? is that cool? | 19:14 |
| greenisus | gundlach: if i'm not here, i'm mike@openstack.org on email | 19:14 |
| gundlach | greenisus: absolutely. and if you want to subscribe to the bug itself, you can get email when it's updated. | 19:14 |
| gundlach | https://bugs.launchpad.net/nova/+bug/660115 | 19:14 |
| uvirtbot | Launchpad bug 660115 in nova "OpenStack API auth manager not processing Users/Tokens correctly" [Critical,In progress] | 19:14 |
| greenisus | gundlach: good call | 19:14 |
| Orman | So many bug fixes today. :D | 19:14 |
| Orman | This is an interesting bug | 19:19 |
| Orman | https://bugs.launchpad.net/swift/+bug/651598 | 19:19 |
| uvirtbot | Launchpad bug 651598 in swift "Write timeout bug in proxy server" [High,New] | 19:19 |
| soren | notmyname: Sorry for just leaving earlier. So, about a unified concept of users across swift and nova (and whatever else comes along): I think we're going to need it, but perhaps you can enlighten me. | 19:19 |
| notmyname | soren: not sure I have much to add, other than to agree that "we're going to need it" | 19:20 |
| soren | notmyname: to support something like shared AMI's for EC2 (VM images), it seems like a sound approach to use Swift's ACL mechanism to enforce it. | 19:20 |
| creiht | Orman: :) | 19:21 |
| notmyname | soren: gholt's swauth branch (not merged yet) is a good start to a swift-based identity management system | 19:21 |
| soren | notmyname: Without a shared user registry of some sort, I'm not sure how to achieve that. | 19:21 |
| Ryan_Lane | isn't nova using ldap? | 19:21 |
| Orman | creiht: It's like an interesting bug,but wondering how it came to be since it's in the proxy server. | 19:22 |
| creiht | It was an edge case that we didn't see very often in testing | 19:23 |
| creiht | or I guess at all :) | 19:23 |
| *** rlucio has quit IRC | 19:23 | |
| Orman | Ok so the fix is coming soon right? I see someone's already assigned to it. | 19:24 |
| creiht | https://code.launchpad.net/~clay-gerrard/swift/write_timeout_fix/+merge/38282 | 19:24 |
| creiht | fix will be in the release | 19:24 |
| Orman | Ok | 19:24 |
| Orman | cool | 19:24 |
| Orman | Jus thought I'd ask. :d | 19:25 |
| Orman | :D | 19:25 |
| creiht | not a problem :) | 19:25 |
| creiht | soren: the auth is an interesting question | 19:25 |
| creiht | as it is my impression that *most* places are going to either already have an auth that they are going to integrate with, or already have a pretty good idea what they want to use for auth | 19:25 |
| soren | Ryan_Lane: It /can/ use LDAP, but doesn't have to. | 19:26 |
| Orman | Not much auth docs on wiki I believe. I should see if I I can write some up when I get a chance. | 19:26 |
| Ryan_Lane | ah | 19:26 |
| creiht | How much need is there going to be for a cannonical auth for openstack? | 19:26 |
| soren | creiht: True, but I think there's more to it than that. | 19:26 |
| creiht | beyond a dev auth | 19:26 |
| soren | creiht: It's not just about auth, it's about user information, too. | 19:26 |
| Ryan_Lane | well, if swift and nova both have a common auth framework, it would be possible to write a plugin for both that unifies them, correct? | 19:27 |
| Ryan_Lane | A SAML plugin for both would likely do the job nicely | 19:27 |
| Orman | Yes | 19:27 |
| soren | creiht: Hm... Trying to think of a good example. | 19:27 |
| Ryan_Lane | or an OpenAM plugin | 19:27 |
| soren | creiht: Ah, yes. | 19:27 |
| jonwood | There needs to be some form dummy auth at least, so people can try things out without having to integrate with their existing auth. | 19:27 |
| creiht | jonwood: yeah that is already there | 19:28 |
| soren | creiht: So, EC2 mandates a specific authentication scheme. | 19:28 |
| soren | creiht: Using its own style of keys and whatnot. | 19:28 |
| *** ecarlin has joined #openstack | 19:28 | |
| Orman | Yes encryption would be in the auth plugin as well right? | 19:28 |
| Orman | OpenAM plugin seems like a very good idea. | 19:29 |
| soren | creiht: So, even though an organisation may have an auth system already (RADIUS, LDAP, Kerberos, whatever), there's a need for provisioning users for access to Nova's EC2 API. | 19:29 |
| Ryan_Lane | generic SAML would be even nicer :) | 19:29 |
| Ryan_Lane | write an openam plugin so that it handles it? :) | 19:30 |
| creiht | soren: hrm | 19:30 |
| *** dendrobates is now known as dendro-afk | 19:30 | |
| creiht | soren: It would be nice to have someone who has that problem write the code and donate it :) | 19:32 |
| creiht | I'm worried about trying to solve a problem we really don't know enough about | 19:32 |
| soren | I agree. | 19:32 |
| soren | I mean, I share your concern. | 19:33 |
| creiht | hah | 19:33 |
| *** dendro-afk is now known as dendrobates | 19:33 | |
| creiht | or at least assist somone who has that problem | 19:33 |
| creiht | so that we make sure it gets done right | 19:33 |
| Orman | Hmm I guess we will just have to find the right person. ;D | 19:33 |
| soren | I think a good start would just be to come up with a solution based on what we have now, and punt on the subject of external authentication for a little bit. | 19:34 |
| creiht | There is no better way to write bad software than to write software to solve a problem that you don't have | 19:34 |
| Ryan_Lane | really just making auth plugin frameworks for both is likely enough | 19:34 |
| creiht | Ryan_Lane: we pretty much have that already | 19:34 |
| soren | creiht: Well said. | 19:35 |
| alekibango | Orman: you want to help writing about security? | 19:35 |
| Orman | Yeah I could give it a shot. Why? | 19:36 |
| alekibango | anne wrote me some questions but is still not here, i will pvtmsg you the text ok? | 19:36 |
| Orman | Ok | 19:36 |
| notmyname | alekibango: keep in mind we also have etherpad.openstack.org | 19:37 |
| notmyname | keep as much in the open as you can ;-) | 19:37 |
| alekibango | notmyname: its just generic ideas about starting security related docs | 19:38 |
| alekibango | dont worry, i am not elitist working in dark room | 19:38 |
| notmyname | hehe, no worries | 19:38 |
| alekibango | its just i dont want to clutter this channel too much | 19:38 |
| alekibango | people have other things to do here :) | 19:38 |
| *** rlucio has joined #openstack | 19:43 | |
| *** rlucio has quit IRC | 19:48 | |
| mtaylor | jaypipes: running "nosetests -w nova/tests/api" on current nova trunk on the hudson box winds up with a failure | 19:51 |
| mtaylor | jaypipes: AttributeError: 'Controller' object has no attribute 'network_manager' | 19:51 |
| mtaylor | ERROR: test_create_instance (nova.tests.api.openstack.test_servers.ServersTest) | 19:51 |
| *** rlucio has joined #openstack | 19:53 | |
| jaypipes | mtaylor: already fixed. see bug660095. | 19:54 |
| soren | mtaylor: I don't know if I told you, but tarmac still doesn't change bugs to "fix committed" when their branches are merged. | 19:54 |
| soren | mtaylor: It gives a different error now, though. :) | 19:54 |
| mtaylor | jaypipes: I'm just wondering if you want me to go ahead and enable the nosetests for hudson even though it's failing for me? | 19:55 |
| mtaylor | soren: lovely | 19:55 |
| mtaylor | soren: I'll see what I can discover... | 19:55 |
| *** Orman has quit IRC | 19:56 | |
| jaypipes | mtaylor: you will have to enable nosetests *after* this is merged: https://code.launchpad.net/~jaypipes/nova/bug660095/+merge/38356 | 19:56 |
| mtaylor | jaypipes: ok. I'll wait then | 19:58 |
| jaypipes | mtaylor: cheers. | 19:59 |
| *** sneumann has quit IRC | 20:00 | |
| *** daleolds has quit IRC | 20:02 | |
| *** localhost has quit IRC | 20:05 | |
| *** daleolds has joined #openstack | 20:06 | |
| *** littleidea has quit IRC | 20:06 | |
| *** Ryan_Lane has quit IRC | 20:07 | |
| *** dizz has joined #openstack | 20:07 | |
| *** localhost has joined #openstack | 20:08 | |
| *** ecarlin has quit IRC | 20:10 | |
| *** zaitcev has joined #openstack | 20:10 | |
| vishy | rlucio: sorry i haven't really used flat networking much | 20:15 |
| vishy | i'd have to look it up to remember | 20:15 |
| vishy | :) | 20:15 |
| vishy | oh | 20:16 |
| vishy | flat-dhcp got merged | 20:16 |
| vishy | :) | 20:16 |
| rlucio | vishy: that's fine... i think that the 'changing the ip range fails' thing is the same issue i had a while back when we were still using redis | 20:17 |
| *** metoikos has joined #openstack | 20:17 | |
| rlucio | vishy: i just saw that a bit ago! thanks for that btw :) | 20:17 |
| vishy | i think you need to manually create networks | 20:17 |
| *** Ryan_Lane has joined #openstack | 20:17 | |
| rlucio | well, its more a matter of 1) i initially installed a recent build with mostly defaults, and the fixed_ips table in sql was filled with 10.0.0.X addresses | 20:19 |
| *** ecarlin has joined #openstack | 20:19 | |
| rlucio | and then 2) i change the network config to include '--private_range=192.168.2.64/26' (or something) and then restart nova-network, and fixed_ips table still shows 10.0.0.x | 20:20 |
| rlucio | and instances still launched with the 10.0.0.x ips | 20:20 |
| rlucio | what i had to do before to fix this was to delete the ip list in redis, and then restart nova-network | 20:22 |
| *** greenisus has quit IRC | 20:22 | |
| rlucio | https://bugs.launchpad.net/nova/+bug/638449 | 20:23 |
| uvirtbot | Launchpad bug 638449 in nova "Cannot update the flat network IP address list" [Medium,Triaged] | 20:23 |
| rlucio | (that is the old, redis related bug) | 20:24 |
| *** ccustine has quit IRC | 20:26 | |
| *** ecarlin has quit IRC | 20:30 | |
| *** deshantm has quit IRC | 20:31 | |
| *** dizz is now known as dizz|away | 20:32 | |
| *** spectorclan has joined #openstack | 20:32 | |
| spectorclan | I have a few quick nova questions, anyone out there who can help? | 20:32 |
| alekibango | shoot it | 20:33 |
| creiht | spectorclan: ask, and you will find out :) | 20:33 |
| spectorclan | Great i am working on a high level doc explaining the basic components of Nova and need to understand why we have the Redis service (server) and what data is being stored there | 20:34 |
| *** deshantm has joined #openstack | 20:34 | |
| alekibango | redis is used as distributed key/value db. | 20:35 |
| spectorclan | @alekibango, a bit more detail on what that means. Thanks. | 20:36 |
| larissa | spectorclan: Error: "alekibango," is not a valid command. | 20:36 |
| alekibango | :) | 20:36 |
| creiht | spectorclan: this isn't twitter :) | 20:36 |
| spectorclan | i typed too fast, sorry | 20:36 |
| spectorclan | too many communication systems for 1 day | 20:37 |
| alekibango | that means it can run on more servers in cluster (reliable) and it stores some values accessible by a key. | 20:37 |
| spectorclan | ok,what are those values we are storing? | 20:37 |
| alekibango | good one. | 20:37 |
| *** kevnfx has joined #openstack | 20:39 | |
| alekibango | its for fast, shared state data | 20:39 |
| alekibango | ReDIS == Remote Dictionary Store | 20:39 |
| spectorclan | ok, so I can assume that by state we mean all the critical data of what servers are running on what machines, etc? | 20:40 |
| alekibango | for example, fake ldap data | 20:40 |
| alekibango | (users, etc) | 20:40 |
| alekibango | spectorclan: i dont know yet, looking for it in sources | 20:40 |
| *** deshantm has quit IRC | 20:40 | |
| alekibango | but you are right, there are some critial data which we need to have safely stored | 20:40 |
| spectorclan | Great. I appreciate the help, trying to create a simple easy to follow intro tech document listing all the components and what they do | 20:41 |
| alekibango | spectorclan: thats very much needed. i will help any way i can | 20:41 |
| spectorclan | The Swift document was much easier to write as there are less pieces and it is easier to understand; this is much tricker | 20:41 |
| spectorclan | tricker? Now I can't spell | 20:41 |
| creiht | spectorclan: got a link? | 20:41 |
| alekibango | spectorclan: yes, its pretty complicated | 20:42 |
| *** pothos_ has joined #openstack | 20:42 | |
| spectorclan | I have the first draft on my machine and not posted anywhere. Send me an email to stephen.spector@openstack.org and I can email you the .odt version for feedback and update | 20:42 |
| *** pothos_ has quit IRC | 20:43 | |
| *** pothos_ has joined #openstack | 20:43 | |
| creiht | spectorclan: how about copy/paste to http://etherpad.openstack.org/ ? | 20:43 |
| alekibango | spectorclan: it could be much better to use etherpad | 20:44 |
| alekibango | heh creiht: ) | 20:44 |
| alekibango | spectorclan: i hate mailing | 20:44 |
| *** pothos_ has quit IRC | 20:44 | |
| gundlach | greenisus: you around? | 20:44 |
| spectorclan | doing etherpad thing now; hold on | 20:44 |
| *** pothos_ has joined #openstack | 20:45 | |
| *** mmalgeri has quit IRC | 20:45 | |
| *** pothos_ has quit IRC | 20:45 | |
| spectorclan | http://etherpad.openstack.org/FFBkFjikbp | 20:46 |
| alekibango | i have more than 2^14 mails waiting for processing right now | 20:46 |
| *** pothos_ has joined #openstack | 20:46 | |
| *** pothos has quit IRC | 20:46 | |
| spectorclan | FYI, the very poor pictures I created did not copy over but the text is there | 20:46 |
| *** pothos_ has quit IRC | 20:46 | |
| alekibango | do you need help with making images? | 20:46 |
| *** pothos has joined #openstack | 20:46 | |
| *** dwight has quit IRC | 20:47 | |
| spectorclan | alekibango: Todd Morey is going to do that are part of laying out the full design of the document but thanks. If he gets too backed up I will let you know | 20:47 |
| *** pothos has quit IRC | 20:47 | |
| *** pothos has joined #openstack | 20:47 | |
| alekibango | spectorclan: put the nova doc on the etherpad, so we can help you by live editing | 20:48 |
| alekibango | and that can become some new wiki page | 20:48 |
| *** pothos has quit IRC | 20:48 | |
| alekibango | we already did that with nova deployment/admin/enduser manual | 20:48 |
| alekibango | and it worked well | 20:48 |
| *** pothos has joined #openstack | 20:48 | |
| alekibango | see how it evolved using time slider: http://etherpad.openstack.org/ep/pad/view/NovaDocsNotes/latest | 20:49 |
| *** pothos has quit IRC | 20:49 | |
| *** pothos has joined #openstack | 20:49 | |
| *** pothos has joined #openstack | 20:50 | |
| *** pothos has quit IRC | 20:51 | |
| *** pothos has joined #openstack | 20:51 | |
| *** dizz|away has quit IRC | 20:51 | |
| *** pothos has quit IRC | 20:51 | |
| *** metoikos has quit IRC | 20:52 | |
| *** dizz has joined #openstack | 20:52 | |
| *** pothos has joined #openstack | 20:52 | |
| spectorclan | alexibango: yep, I see and that makes sense. I will leave the document up where it is on etherpad and get community feedback and changes before formal pdf and placement on the website, | 20:52 |
| gundlach | anyone want to review https://code.launchpad.net/~gundlach/nova/lp660115/+merge/38365 ? | 20:52 |
| *** pothos has quit IRC | 20:52 | |
| gundlach | fixes https://bugs.launchpad.net/nova/+bug/660115 . | 20:52 |
| uvirtbot | Launchpad bug 660115 in nova "OpenStack API auth manager not processing Users/Tokens correctly" [Critical,Fix committed] | 20:52 |
| *** pothos has joined #openstack | 20:53 | |
| *** deshantm has joined #openstack | 20:53 | |
| *** metoikos has joined #openstack | 20:53 | |
| *** pothos has quit IRC | 20:53 | |
| *** pothos has joined #openstack | 20:54 | |
| alekibango | spectorclan: you can create named etherpad page by typing it in url: http://etherpad.openstack.org/OpenStackComputeTechnologyOverview :) | 20:54 |
| *** pothos has quit IRC | 20:54 | |
| spectorclan | alekibango - thanks EtherPad is new to me. | 20:54 |
| *** pothos has joined #openstack | 20:54 | |
| alekibango | also to me, but its nice to play with | 20:54 |
| alekibango | i would like to have similar wysiwyg editor in wiki... | 20:55 |
| *** pothos has quit IRC | 20:55 | |
| alekibango | one should be part of the wiki software, but is not used (enabled)... | 20:55 |
| *** pothos has joined #openstack | 20:55 | |
| spectorclan | alexkibango - we are looking at a new wiki for openstack, submit your favorites to myself or Anne | 20:55 |
| *** pothos has quit IRC | 20:56 | |
| spectorclan | Thanks for your help everyone; I am off to propose a new openstack.org web structure; once I get some OKs here I will put on EtherPad for community review as well | 20:56 |
| alekibango | i cant really suggest good one. But i used docuwiki (php) with success on my small needs | 20:56 |
| *** pothos has joined #openstack | 20:56 | |
| *** spectorclan has quit IRC | 20:56 | |
| *** pothos has quit IRC | 20:57 | |
| alekibango | http://www.dokuwiki.org/dokuwiki | 20:57 |
| *** pothos has joined #openstack | 20:57 | |
| *** khussein has joined #openstack | 20:58 | |
| *** pothos has joined #openstack | 20:58 | |
| *** pothos has quit IRC | 20:58 | |
| alekibango | and i wrote few wiki implementations in different frameworks (lately in turbogears) -- just by following 20 minute tutorials :) | 20:59 |
| *** pothos has joined #openstack | 20:59 | |
| *** pothos has joined #openstack | 21:00 | |
| alekibango | http://www.wikimatrix.org/ <-- comparison :) | 21:00 |
| annegentle | alekibango: thanks, we'll add it to the list | 21:03 |
| alekibango | choice wizard might help there | 21:03 |
| alekibango | please include WYSIWYG, history and maybe even RCS | 21:04 |
| *** localhost has quit IRC | 21:05 | |
| *** dizz has quit IRC | 21:06 | |
| gundlach | jaypipes: approved your merge | 21:06 |
| alekibango | if you will choose dokuwiki, i have some experience with it, so i can help a bit | 21:06 |
| alekibango | but its really easy one to use | 21:07 |
| *** localhost has joined #openstack | 21:07 | |
| *** gustavomzw has joined #openstack | 21:10 | |
| *** dizz has joined #openstack | 21:11 | |
| *** Sanjib has joined #openstack | 21:11 | |
| dendrobates | fyi: for everyone that is testing openstack in the rackspace cloud, like me. you may notice that vlans do not work. This is not a bug, it is because the 8021q kernel module is not available | 21:15 |
| *** greenisus has joined #openstack | 21:15 | |
| dendrobates | I have asked the ops team to include it in the next kernel build, so we can have a fully functioning cloud within the cloud. | 21:16 |
| *** jdarcy has quit IRC | 21:16 | |
| *** mmalgeri has joined #openstack | 21:18 | |
| alekibango | annegentle: the current wiki should be able to use wysiwyg editor -- like this one: http://moinmo.in/4ct10n/edit/Python?action=edit&editor=gui | 21:18 |
| dendrobates | alekibango: tell me where to get it and I'll install it on the wiki | 21:20 |
| dendrobates | If annegentle approves | 21:20 |
| alekibango | dendrobates: not sure, they claim thats included in core | 21:20 |
| dendrobates | hmm? | 21:20 |
| alekibango | from certain version | 21:20 |
| alekibango | dendrobates: but i never installed that one | 21:21 |
| dendrobates | I'll look into it. | 21:21 |
| dendrobates | after the release! | 21:21 |
| alekibango | :) ok | 21:21 |
| alekibango | i will stop whining for 2 weeks | 21:22 |
| *** cachedout has quit IRC | 21:23 | |
| *** faccenda has quit IRC | 21:23 | |
| *** tonywolf has joined #openstack | 21:23 | |
| *** faccenda has joined #openstack | 21:23 | |
| zaitcev | Is anyone working on S3 front-end for Swift? I have a few ideas, roughly centered around taking the Hail implementation and using Swift instead of Hail's chunkserver. | 21:27 |
| zaitcev | Seems quite workable at first blush. | 21:27 |
| zaitcev | It's not going to be awfully well performing at first but it's better than nothing, right? | 21:28 |
| *** zykes- has quit IRC | 21:28 | |
| *** zykes- has joined #openstack | 21:29 | |
| alekibango | zaitcev: s3 api is not there for swift | 21:33 |
| alekibango | but some people were talking about making some s3 api around swift, maybe that will happen later | 21:34 |
| zaitcev | alekibango: OK, the race is on, then. | 21:34 |
| alekibango | zaitcev: please if you want it to happen. write blueprint soon | 21:34 |
| alekibango | there will be call for blueprints with the release | 21:34 |
| alekibango | i am sure some people would help you | 21:35 |
| zaitcev | I'll poke at the mailing list when I have something. | 21:36 |
| alekibango | rather write the blueprint | 21:37 |
| alekibango | its a plan -- people will read it there | 21:38 |
| alekibango | and do it soon (around release) | 21:38 |
| zaitcev | Thanks, I'll examine the blueprint mechanism. | 21:38 |
| alekibango | i never wrote one, but its not hard | 21:38 |
| alekibango | (i am also new on LP) | 21:38 |
| *** Orman has joined #openstack | 21:39 | |
| dendrobates | alekibango: I lied the gui editor is now functional on the wiki | 21:39 |
| alekibango | :) wow | 21:39 |
| alekibango | moving fast :) | 21:39 |
| alekibango | looks nice, will be much easier to edit | 21:40 |
| alekibango | (esp. headers are awkwards for me, i used to have =h6= and ==h5== ) | 21:40 |
| alekibango | easy link creation also helps | 21:41 |
| dendrobates | the only way I know to access it is to change &editor=text to &editor=gui | 21:41 |
| alekibango | gj... | 21:41 |
| dendrobates | is their an easier way. | 21:42 |
| alekibango | there is a button | 21:42 |
| dendrobates | ah | 21:42 |
| alekibango | [gui] | 21:42 |
| alekibango | and edit [gui] link on side | 21:42 |
| alekibango | so it will work | 21:42 |
| creiht | zaitcev: I don't think the api's differ a whole lot, so one approach that we talked about was having a bit of middleware that you could put in front of swift that would convert the s3 api calls to swift calls | 21:42 |
| alekibango | dendrobates: dont worry, it works well | 21:42 |
| *** miclorb has joined #openstack | 21:42 | |
| alekibango | creiht: seems nice idea for me, but i dont know swift nor s3 much | 21:43 |
| zaitcev | creiht: That's what I have in mind, by borrowing Hail/tabled code. But, it's a separate process (daemon). | 21:43 |
| *** morfeas has quit IRC | 21:43 | |
| creiht | with wsgi middleware it would be in the same process | 21:43 |
| creiht | it acts like a pass through | 21:43 |
| alekibango | yes i like the idea of using messaging | 21:44 |
| zaitcev | Sadly Hail is written in C, so I cannot just copy-paste it into Swift | 21:44 |
| creiht | hehe | 21:44 |
| gundlach | g'night all! | 21:45 |
| alekibango | zaitcev: now is prerelease fever, so most developers cant really talk much .... | 21:45 |
| creiht | heh | 21:45 |
| zaitcev | Right. | 21:45 |
| *** morfeas has joined #openstack | 21:45 | |
| creiht | zaitcev: you could probably make that work | 21:45 |
| alekibango | i think so | 21:46 |
| creiht | I don't know enough about hail though | 21:46 |
| *** jonwood has quit IRC | 21:46 | |
| creiht | zaitcev: does it support both the REST and SOAP S3 interface? | 21:46 |
| alekibango | zaitcev: this one is really demanded feature, so you would be hero for many people :) | 21:47 |
| zaitcev | creiht: REST only. | 21:47 |
| * alekibango hates soap with passion | 21:47 | |
| zaitcev | Amazon people said next to nobody uses SOAP with S3. | 21:47 |
| creiht | zaitcev: that is good to know | 21:48 |
| alekibango | soap is everything but 'simple' | 21:48 |
| zaitcev | Like 1% of incoming connections | 21:48 |
| zaitcev | Maybe even less. | 21:48 |
| alekibango | everything but 'protocol' | 21:48 |
| alekibango | :) | 21:48 |
| zaitcev | Still, some people do use SOAP. | 21:48 |
| alekibango | zaitcev: they could contribute code to make it work! | 21:49 |
| alekibango | ;) | 21:49 |
| alekibango | can* | 21:49 |
| creiht | I wont accept any SOAP code :) | 21:49 |
| zaitcev | Hail lags a bit. It has no ACL policies parsing, just the old ACLs and canned ACLs, for example. Partial GET is just coming online. etc. SOAP is way down the list here. | 21:49 |
| creiht | zaitcev: While that is one route, I still think the middleware would be a better route | 21:50 |
| zaitcev | Since Hail never had a wide following like OpenStack, I'm considering dumping Hail and switching to Swift altogether. Our new code has CF back-end option. So it's a way to salvage useable code from Hail and ditch the rest. | 21:51 |
| creiht | zaitcev: It would be better than nothing :) | 21:52 |
| annegentle | dendrobates: thumbs up on Edit (GUI) and Edit (Text) - thanks for doing that! | 21:56 |
| dendrobates | annegentle: I also installed the openid dependencies. We should be able to get launchpad logins working | 21:57 |
| annegentle | dendrobates: that would be great too. Could help match people, doc, and code. | 21:58 |
| dendrobates | yes and eventually help automate the CLA check. | 21:59 |
| annegentle | woo, that's a big help | 21:59 |
| *** localhost has quit IRC | 22:05 | |
| *** ecarlin has joined #openstack | 22:05 | |
| mtaylor | dendrobates: that CLA thing is still on my todo list if I can ever find the time... | 22:06 |
| * mtaylor smacks self in face | 22:06 | |
| dendrobates | mtaylor: we are going to do a manual work around, if we can get around to it. | 22:07 |
| mtaylor | dendrobates: excellent | 22:07 |
| *** localhost has joined #openstack | 22:07 | |
| *** Ryan_Lane is now known as Ryan_Lane|away | 22:09 | |
| *** jsgotangco has joined #openstack | 22:09 | |
| dendrobates | mtaylor: we are going to make the LP teams moderated and only allow people that have signed the CLA and added the name and signing code to the wiki to join. | 22:09 |
| mtaylor | dendrobates: cool. that'll make that easier | 22:09 |
| dendrobates | we are also planning on adding a code of conduct to the CLA and calling the new doc a Contributers agreement. | 22:10 |
| dendrobates | so everyone will sign it not just devs | 22:10 |
| mtaylor | dendrobates: and quicker to get going than the "sign a CoC through launchpad and then require a person to have a signed CoC before they can submit a merge request" patch to lp done | 22:10 |
| mtaylor | I think that's an excellent choice | 22:10 |
| dendrobates | but still we need time to get it done. | 22:11 |
| dendrobates | the painful part will be the clean up of all the people that already joined the teams | 22:11 |
| *** Orman has quit IRC | 22:12 | |
| mtaylor | ugh. that's going to suck | 22:12 |
| *** ppetraki has quit IRC | 22:12 | |
| dendrobates | yep | 22:12 |
| dendrobates | we should probably close the teams asap. | 22:13 |
| dendrobates | but we need some text explaining what we are doing. | 22:13 |
| dendrobates | maybe annegentle can help with that | 22:13 |
| mtaylor | annegentle can help with EVERYTHING | 22:15 |
| *** faccenda has quit IRC | 22:15 | |
| * dendrobates takes a family break. See you in a little while | 22:18 | |
| *** Ryan_Lane|away is now known as Ryan_Lane | 22:23 | |
| *** joearnold has joined #openstack | 22:24 | |
| *** electroniceagle has quit IRC | 22:25 | |
| *** gundlach has quit IRC | 22:26 | |
| *** AimanA is now known as HouseAway | 22:27 | |
| *** greenisus has quit IRC | 22:29 | |
| *** mtaylor has quit IRC | 22:30 | |
| *** jakedahn has joined #openstack | 22:32 | |
| *** ecarlin has quit IRC | 22:33 | |
| *** jakedahn has joined #openstack | 22:33 | |
| *** kevnfx has quit IRC | 22:35 | |
| *** littleidea has joined #openstack | 22:37 | |
| *** electroniceagle has joined #openstack | 22:39 | |
| *** ArdRigh has joined #openstack | 22:41 | |
| *** electroniceagle has quit IRC | 22:43 | |
| *** jfluhmann has quit IRC | 22:45 | |
| *** abecc has quit IRC | 22:50 | |
| *** Orman has joined #openstack | 22:51 | |
| *** Podilarius has left #openstack | 22:52 | |
| *** abecc has joined #openstack | 22:52 | |
| *** metoikos has quit IRC | 22:53 | |
| *** Ryan_Lane has quit IRC | 22:56 | |
| *** mtaylor has joined #openstack | 23:00 | |
| *** mtaylor has joined #openstack | 23:00 | |
| *** ChanServ sets mode: +v mtaylor | 23:00 | |
| *** aliguori has quit IRC | 23:03 | |
| *** ctennis has joined #openstack | 23:05 | |
| *** localhost has quit IRC | 23:05 | |
| *** localhost has joined #openstack | 23:08 | |
| *** irahgel has left #openstack | 23:10 | |
| *** ecarlin has joined #openstack | 23:11 | |
| *** aliguori has joined #openstack | 23:16 | |
| dendrobates | quiet tonight | 23:26 |
| dendrobates | I'm am being forced into chores. I'll catch you all later. | 23:27 |
| *** rnirmal has left #openstack | 23:28 | |
| *** jsgotangco has quit IRC | 23:31 | |
| *** tonywolf has quit IRC | 23:33 | |
| *** ecarlin has quit IRC | 23:33 | |
| *** miclorb has quit IRC | 23:40 | |
| *** ppetraki has joined #openstack | 23:43 | |
| *** rlucio has quit IRC | 23:47 | |
| *** littleidea has quit IRC | 23:47 | |
| *** Orman has quit IRC | 23:52 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!