| *** dizz is now known as dizz|away | 00:09 | |
| *** sophiap has quit IRC | 00:12 | |
| *** pharkmillups has quit IRC | 00:15 | |
| *** pharkmillups has joined #openstack | 00:16 | |
| *** Phacops_ has quit IRC | 00:24 | |
| *** Phacops has joined #openstack | 00:24 | |
| *** burris has quit IRC | 00:24 | |
| *** jc_smith has quit IRC | 00:34 | |
| *** pvo has quit IRC | 00:34 | |
| *** Cybodog has quit IRC | 00:36 | |
| *** jakedahn has quit IRC | 00:42 | |
| *** DubLo7 has joined #openstack | 00:48 | |
| *** tobym has joined #openstack | 00:50 | |
| *** ptremblett has joined #openstack | 00:51 | |
| *** iammartian_ has joined #openstack | 00:52 | |
| *** mtaylor has quit IRC | 00:57 | |
| *** ar1 has joined #openstack | 01:02 | |
| *** khussein has joined #openstack | 01:04 | |
| *** aimon has joined #openstack | 01:09 | |
| *** pharkmillups has quit IRC | 01:12 | |
| *** burris has joined #openstack | 01:13 | |
| *** codejunkie has joined #openstack | 01:27 | |
| *** DubLo71 has joined #openstack | 01:28 | |
| *** orman has joined #openstack | 01:31 | |
| orman | Hello | 01:31 |
|---|---|---|
| *** DubLo7 has quit IRC | 01:32 | |
| orman | Anybody here | 01:33 |
| orman | I have a question. | 01:35 |
| notmyname | orman: ask. someone will answer as soon as they can | 01:35 |
| orman | Ok | 01:35 |
| *** dagger_ has quit IRC | 01:36 | |
| *** dagger has joined #openstack | 01:38 | |
| *** dagger has joined #openstack | 01:39 | |
| *** aimon_ has joined #openstack | 01:39 | |
| *** burris has quit IRC | 01:40 | |
| orman | I added a feature request to the Open Stack launchpad site and I was wondering if I proposed after Oct 1? | 01:41 |
| alekibango | orman: url? | 01:41 |
| orman | One sec getting it. | 01:42 |
| orman | I hope it's good,but it's probably been implemented. | 01:43 |
| orman | https://blueprints.launchpad.net/nova/+spec/request-objectstores-validation-inside-ec2-group | 01:43 |
| *** aimon has quit IRC | 01:43 | |
| *** aimon_ is now known as aimon | 01:43 | |
| alekibango | orman: cant comment that, but you will see tomorrow | 01:44 |
| alekibango | btw austin release is very soon (21st) | 01:44 |
| alekibango | and no new features are added | 01:44 |
| *** burris has joined #openstack | 01:45 | |
| orman | Ok,but in the future or the next release do you think new features will be added? | 01:45 |
| orman | I would think so. | 01:46 |
| orman | Hopefully the release will a good one. ;) | 01:48 |
| *** pvo has joined #openstack | 01:49 | |
| *** ChanServ sets mode: +v pvo | 01:49 | |
| orman | Thanks for checking alekibango. ;) | 01:49 |
| *** orman has quit IRC | 01:53 | |
| *** Orman has joined #openstack | 01:55 | |
| *** codejunkie has left #openstack | 01:56 | |
| *** iammartian_ has quit IRC | 02:04 | |
| alekibango | Orman: first release will be a pretty nice one, and second will be even better :) | 02:06 |
| Orman | Cool. I just hope I can help out. | 02:06 |
| Orman | ; | 02:06 |
| Orman | :) | 02:06 |
| Orman | Everyone will have to celebrate when the first release comes out. :P | 02:09 |
| alekibango | everyone will be writing documentation which is pretty much missing now >:) | 02:10 |
| *** aimon has quit IRC | 02:10 | |
| *** aimon has joined #openstack | 02:11 | |
| *** Cybodog has joined #openstack | 02:11 | |
| *** sirp1 has quit IRC | 02:12 | |
| Orman | What about documenting security? Is anyone doing that currently? | 02:12 |
| alekibango | Orman: manual pages just started on wiki... more or less empty skeletons | 02:13 |
| Orman | Ok | 02:13 |
| alekibango | there are some devel docs, but user/admin/deployment nova docs are still pretty much empty | 02:13 |
| alekibango | we need help with that | 02:13 |
| alekibango | see wiki and please try help writing if you will feel | 02:14 |
| alekibango | (be bold, it can be edited later) | 02:14 |
| Orman | Ok, I manly write about security,but I'll see if I can chip in. ;D | 02:15 |
| alekibango | that would be nice | 02:18 |
| *** Orman has quit IRC | 02:24 | |
| *** Orman has joined #openstack | 02:25 | |
| Orman | Alekibango: Thanks for your help so far. ;D | 02:28 |
| *** hazmat has quit IRC | 02:30 | |
| *** guynaor has joined #openstack | 02:48 | |
| *** jasonb365 has joined #openstack | 02:53 | |
| Orman | Can't wait to see what the release has and how many downloads. ;D | 02:54 |
| *** khussein has left #openstack | 02:54 | |
| pvo | orman: the new release should have everything it has now, but with less bugs. : ) | 02:54 |
| Orman | Yeah hopefully it will. I have been on bug duty in the last few weeks and have found just a few. | 02:55 |
| *** ptremblett has quit IRC | 02:58 | |
| *** silassewell has quit IRC | 03:01 | |
| *** iammartian_ has joined #openstack | 03:03 | |
| *** mikhail has joined #openstack | 03:04 | |
| *** pcrews has quit IRC | 03:10 | |
| *** ZNS has joined #openstack | 03:14 | |
| *** ZNS has quit IRC | 03:19 | |
| *** mmalgeri has joined #openstack | 03:28 | |
| *** sophiap has joined #openstack | 03:29 | |
| *** aimon_ has joined #openstack | 03:29 | |
| *** aimon has quit IRC | 03:33 | |
| *** aimon_ is now known as aimon | 03:33 | |
| *** mikhail_ has joined #openstack | 03:35 | |
| *** kashyapc has joined #openstack | 03:36 | |
| *** Orman has quit IRC | 03:37 | |
| *** aliguori has quit IRC | 03:38 | |
| *** mikhail has quit IRC | 03:38 | |
| *** aimon has quit IRC | 03:41 | |
| *** aimon has joined #openstack | 03:41 | |
| *** sirp1 has joined #openstack | 03:43 | |
| *** jasonb365 has quit IRC | 04:09 | |
| *** aimon_ has joined #openstack | 04:22 | |
| *** cw has quit IRC | 04:23 | |
| *** aimon has quit IRC | 04:25 | |
| *** aimon_ is now known as aimon | 04:25 | |
| *** mikhail_ has quit IRC | 04:38 | |
| *** guigui1 has joined #openstack | 04:38 | |
| *** mikhail_ has joined #openstack | 04:45 | |
| *** mmalgeri has quit IRC | 04:55 | |
| *** aimon has quit IRC | 04:58 | |
| *** aimon has joined #openstack | 04:59 | |
| *** cw has joined #openstack | 05:06 | |
| *** pharkmillups has joined #openstack | 05:15 | |
| soren | vishy: Just making sure: You guys are expecting to try to get your fixes from your deploy branch merged into trunk, right? | 05:22 |
| *** mikhail__ has joined #openstack | 05:27 | |
| *** mikhail_ has quit IRC | 05:30 | |
| *** jtimberman has quit IRC | 05:32 | |
| vishy | soren: some of them for sure. If we can get them all in it would be rad, but I don't know if we'll make it | 05:34 |
| soren | vishy: Ok. Do you have a timeframe for that? | 05:34 |
| vishy | soren: we're still fighting with launch, but we plan on focusing on review and merging all this week. We're working through a bit of an issue with eventlet and failures | 05:35 |
| soren | vishy: I'm just trying to plan my week, and if 37 merge proposals are going to turn up at some point, that's probably something I should take into account. | 05:35 |
| *** jtimberman has joined #openstack | 05:35 | |
| *** kevnfx has left #openstack | 05:37 | |
| vishy | soren: we need to rewrite a bit of it against trunk, so I don't think we're going to try and ninja merge all of them | 05:38 |
| vishy | soren: just the important ones | 05:38 |
| *** anotherjesse has joined #openstack | 05:38 | |
| anotherjesse | soren: hopefully we can figure out the failure, since after about running several thousand instances within a couple of hours, we see weird stuff | 05:41 |
| soren | anotherjesse: Like what? | 05:41 |
| *** huz has joined #openstack | 05:42 | |
| anotherjesse | soren: on compute, iptables can give "resource unavailable" | 05:42 |
| anotherjesse | the exception isn't handled in a sane way from the user's point of view | 05:43 |
| anotherjesse | (it is left in a state that says "launching" but it isn't continuing to launch) | 05:43 |
| anotherjesse | also on compute we can get to a state where all the "process pool" are all hung and the node is sitting there doing nothing | 05:44 |
| huz | hi all | 05:45 |
| huz | i got a problem when i try to setup a dev VM according the documentation | 05:46 |
| huz | the command in the manual: "swift-auth-add-user -K devauth -a test tester testing" seems doesn't work because swift-auth-add-user was not found | 05:47 |
| huz | i checked the source folder of swift(version 1.0.2) and the install folder, the command file was not there, is there anyone where can help me ? | 05:49 |
| anotherjesse | soren: also on API we are having mysql / rabbitmq connection issues (where it "goes away" or duplicate read detected) and nova-api doesn't always die when you call it with stop - which means that process control via monit or puppet is hampered | 05:49 |
| soren | anotherjesse: :( | 05:51 |
| anotherjesse | soren: we are spending a lot of time stress testing it by hitting it as hard as we can and then fix it | 05:51 |
| *** f4m8_ is now known as f4m8 | 05:53 | |
| anotherjesse | works good until you launch ~13,000 instances :) | 05:53 |
| anotherjesse | in 2 hours | 05:53 |
| *** pharkmillups has quit IRC | 05:53 | |
| anotherjesse | soren: did you see the patch to send instances to a specific host by using the availility zone? | 05:55 |
| anotherjesse | (based on http://wiki.openstack.org/blueprint-hardware-staging) | 05:56 |
| soren | anotherjesse: i don't think so, no. :( | 05:56 |
| anotherjesse | basically we want to be able to hit a specific node during buring / testing | 05:57 |
| soren | anotherjesse: In that wiki page, are 1) and 2) different options for achieving the same goal or are they supposed to work together? | 05:58 |
| anotherjesse | then we use that with http://github.com/vishvananda/novatools/raw/master/run-instances | 05:58 |
| anotherjesse | together | 05:58 |
| anotherjesse | since if you add new compute nodes as "enabled" by default, then they will be sent instances before they have been determined sane | 05:59 |
| soren | anotherjesse: Sure, if those are the semantics. | 06:00 |
| soren | anotherjesse: Ok. | 06:00 |
| soren | anotherjesse: We should really have a test suite that we could run on new boxes that would pound on them and make sure they act properly and only after passing this test, they could get added to the pool. | 06:01 |
| soren | anotherjesse: Having to do that manually would really suck. | 06:01 |
| anotherjesse | soren: I'm currently thinking about launching a "tiny" instance periodically on each node | 06:04 |
| soren | anotherjesse: Reserving a bit of ressources on each for testing? | 06:05 |
| soren | Sounds reasonable enough. | 06:06 |
| * soren has always believed that integration testing isn't just for development | 06:06 | |
| *** Ryan_Lane|away has quit IRC | 06:08 | |
| soren | huz: swift-auth-create-account might be what you're looking for. | 06:12 |
| huz | soren: i notice that swift-auth-add-user has some options such as -K -a, so in order to replace the command line:swift-auth-add-user -K devauth -a test tester testing, how can i do this by swift-auth-create-account? it doesn't has these options. | 06:25 |
| soren | huz: Sorry, I don't know. | 06:44 |
| huz | soren:ok, thank you all the same :) | 06:46 |
| *** anotherjesse has quit IRC | 06:47 | |
| *** mikhail_ has joined #openstack | 07:04 | |
| *** allsystemsarego has joined #openstack | 07:04 | |
| *** befreax has joined #openstack | 07:07 | |
| *** mikhail__ has quit IRC | 07:07 | |
| *** anotherjesse has joined #openstack | 07:11 | |
| *** mikhail_ has quit IRC | 07:17 | |
| *** omidhdl has joined #openstack | 07:20 | |
| *** zykes- has joined #openstack | 07:25 | |
| *** perestrelka has joined #openstack | 07:26 | |
| *** sirp1 has quit IRC | 07:32 | |
| *** omidhdl1 has joined #openstack | 07:46 | |
| *** omidhdl has quit IRC | 07:48 | |
| *** dizz|away is now known as dizz | 07:59 | |
| *** dizz is now known as dizz|away | 08:10 | |
| *** irahgel has joined #openstack | 08:10 | |
| *** krish has joined #openstack | 08:13 | |
| *** dizz|away is now known as dizz | 08:15 | |
| *** miclorb has quit IRC | 08:18 | |
| *** miclorb_ has joined #openstack | 08:23 | |
| *** dizz is now known as dizz|away | 08:35 | |
| *** aimon_ has joined #openstack | 08:37 | |
| *** aimon has quit IRC | 08:39 | |
| *** aimon_ is now known as aimon | 08:39 | |
| *** dizz|away is now known as dizz | 08:48 | |
| *** sneumann has joined #openstack | 08:59 | |
| *** ptremblett has joined #openstack | 09:08 | |
| *** ptremblett has quit IRC | 09:12 | |
| *** gasbakid has joined #openstack | 09:13 | |
| *** omidhdl has joined #openstack | 09:24 | |
| *** omidhdl1 has quit IRC | 09:24 | |
| *** huz has quit IRC | 09:25 | |
| *** jonwood has joined #openstack | 09:38 | |
| *** hazmat has joined #openstack | 09:48 | |
| *** hazmat has quit IRC | 09:54 | |
| *** tsharju has joined #openstack | 09:55 | |
| tsharju | Hi, all... I'm trying to setup Swift cluster and I'm wondering how/where I should run the replicators? I'm running separate proxy servers and storage servers (account,container,object)... | 09:58 |
| *** Kami_ has quit IRC | 10:04 | |
| *** Kami_ has joined #openstack | 10:04 | |
| *** krish has quit IRC | 10:12 | |
| *** Abd4llA has joined #openstack | 10:20 | |
| *** ctennis has quit IRC | 10:49 | |
| *** hazmat has joined #openstack | 10:52 | |
| soren | How incredibly annoying. | 11:06 |
| soren | My security groups patch finally got approved, and I just learned that it doesn't actually work. | 11:07 |
| *** DubLo71 has quit IRC | 11:09 | |
| zykes- | :p | 11:19 |
| *** miclorb_ has quit IRC | 11:23 | |
| *** hazmat has quit IRC | 11:23 | |
| jonwood | I may be doing something wrong, but it seems that euca_terminate_instances doesn't work with Nova. I get an error message about terminate_instances expecting 3 arguments, but only getting 2. | 11:29 |
| *** ctennis has joined #openstack | 11:31 | |
| *** ctennis has joined #openstack | 11:31 | |
| jonwood | jon@myfanwy:~$ euca-terminate-instances i-1oag1wr | 11:32 |
| jonwood | TypeError: terminate_instances() takes exactly 3 non-keyword arguments (2 given) | 11:32 |
| *** kashyapc has quit IRC | 11:36 | |
| dizz | jonwood: I've have the same issue too | 11:47 |
| jonwood | I guess that's something then. | 11:48 |
| jonwood | I'll have a dig around once I have a bit of spare time I guess. | 11:49 |
| *** DubLo7 has joined #openstack | 12:04 | |
| soren | jonwood: It's fixed in trunk. | 12:14 |
| jonwood | Excellent, guess I won't bother digging around then ;) | 12:18 |
| *** omidhdl has left #openstack | 12:23 | |
| *** aliguori has joined #openstack | 12:23 | |
| *** hazmat has joined #openstack | 12:24 | |
| *** pvo has quit IRC | 12:28 | |
| *** pvo has joined #openstack | 12:28 | |
| *** pvo has joined #openstack | 12:28 | |
| *** ChanServ sets mode: +v pvo | 12:28 | |
| *** pvo_ has joined #openstack | 12:35 | |
| *** ChanServ sets mode: +v pvo_ | 12:35 | |
| uvirtbot | New bug: #659135 in nova "Security Group ingress filtering doesn't work if traffic is routed" [Undecided,New] https://launchpad.net/bugs/659135 | 12:37 |
| *** pvo has quit IRC | 12:38 | |
| *** pvo_ has quit IRC | 12:39 | |
| *** dizz is now known as dizz|away | 12:39 | |
| *** gustavomzw has joined #openstack | 12:44 | |
| *** hazmat has quit IRC | 12:46 | |
| *** pvo has joined #openstack | 12:48 | |
| *** pvo has joined #openstack | 12:48 | |
| *** ChanServ sets mode: +v pvo | 12:48 | |
| *** hazmat has joined #openstack | 12:51 | |
| *** abecc has joined #openstack | 12:57 | |
| *** sophiap has quit IRC | 13:04 | |
| jaypipes | yawn. | 13:09 |
| *** jdarcy has joined #openstack | 13:09 | |
| *** hazmat has quit IRC | 13:11 | |
| *** abecc has joined #openstack | 13:11 | |
| *** daleolds has quit IRC | 13:13 | |
| *** gustavomzw has quit IRC | 13:13 | |
| *** sophiap has joined #openstack | 13:14 | |
| *** gustavomzw has joined #openstack | 13:16 | |
| *** cdub has quit IRC | 13:17 | |
| soren | jaypipes: My security groups branch didn't merge yesterday. Can you give it another quick look and re-approve? | 13:19 |
| * soren has to run, will be back later | 13:19 | |
| *** dendro-afk is now known as dendrobates | 13:20 | |
| *** iammartian_ has quit IRC | 13:34 | |
| *** rnirmal has joined #openstack | 13:34 | |
| *** mdomsch has joined #openstack | 13:37 | |
| *** johnbergoon has joined #openstack | 13:38 | |
| *** abecc has quit IRC | 13:40 | |
| *** abecc has joined #openstack | 13:40 | |
| *** abecc has joined #openstack | 13:40 | |
| *** f4m8 is now known as f4m8_ | 13:41 | |
| *** dizz|away is now known as dizz | 13:41 | |
| creiht | tsharju: the replicators run on the same server as the service it is replicating | 13:45 |
| creiht | so if you have storage servers that are running account, container and object, then the replicators for all of those should be run on the storage machines as well | 13:46 |
| creiht | doh... huz quit before I can answer his questions :/ | 13:49 |
| notmyname | creiht: the obvious answer here is to never try to answer questions ;-) | 13:55 |
| creiht | heh | 13:56 |
| jaypipes | soren: sure will! | 13:57 |
| *** kashyapc has joined #openstack | 13:57 | |
| zykes- | does nova require packages installed on the nodes ? | 13:58 |
| *** mdomsch has quit IRC | 14:04 | |
| *** kashyapc has quit IRC | 14:04 | |
| jonwood | zykes-: Not on the VMs it starts, no. | 14:04 |
| *** kashyapc has joined #openstack | 14:04 | |
| jonwood | Although I think there's an agent that you can run to allow extra functionality. | 14:04 |
| *** hisaki has joined #openstack | 14:06 | |
| *** iammartian_ has joined #openstack | 14:12 | |
| *** Cybodog has quit IRC | 14:13 | |
| *** Cybodog has joined #openstack | 14:14 | |
| dendrobates | soren: your security groups merge has conflicts. Please fix. | 14:21 |
| *** kevnfx has joined #openstack | 14:21 | |
| *** Cybodog has quit IRC | 14:24 | |
| *** Cybodog has joined #openstack | 14:25 | |
| dendrobates | jaypipes: add-glance has merge conflicts. | 14:27 |
| jaypipes | dendrobates: I told you yesterday that's an obselete branch :) Also, it's not even on the Active Reviews page :) https://code.launchpad.net/nova/+activereviews | 14:29 |
| *** khussein has joined #openstack | 14:30 | |
| dendrobates | hmm, I thought you deleted that one. | 14:30 |
| dendrobates | I deleted it. | 14:30 |
| jaypipes | :) | 14:30 |
| dendrobates | it showed up here: https://code.edge.launchpad.net/~hudson-openstack/nova/trunk/+activereviews | 14:31 |
| dendrobates | this is where I am looking. | 14:31 |
| jaypipes | dendrobates: hmm, I don't look there... | 14:32 |
| dendrobates | it seems hudson or tarmac or whatever, is not deleting merge proposals when the get deleted from nova | 14:32 |
| jaypipes | dendrobates: mtaylor? :) | 14:32 |
| jaypipes | dendrobates: I guess Tarmac/Hudson should be on that A&E show "Hoarders" | 14:33 |
| dendrobates | :) | 14:33 |
| *** piken has joined #openstack | 14:34 | |
| piken | Hi all, has anyone tried to run Nova on an infiniband mesh network? | 14:35 |
| jaypipes | piken: hmm, not that I know of... you could try the mailing list and you will prolly get some responses there... | 14:36 |
| dendrobates | piken: my guess is no. Should not be a problem,, though | 14:36 |
| dendrobates | piken: it has been run over 10Ge | 14:37 |
| piken | my concern is it running in an ip'less fabric network | 14:38 |
| * jaypipes slaps dendrobates for putting ec2-secuity-groups to Approve over and over again.. | 14:38 | |
| jaypipes | dendrobates: lightly slaps, of course ;) | 14:38 |
| dendrobates | I got confused by sorens quoting of the source in a comment. | 14:39 |
| *** johnpur has joined #openstack | 14:39 | |
| *** ChanServ sets mode: +v johnpur | 14:39 | |
| dendrobates | I could not see the merge conflicts | 14:39 |
| dendrobates | I realized it right after I clicked. | 14:39 |
| jaypipes | dendrobates: jk, man :) | 14:40 |
| *** Cybodog has quit IRC | 14:40 | |
| dendrobates | well we now know why I was seeing more merge conflicts than you yesterday. We were looking in different places. | 14:41 |
| *** Cybodog has joined #openstack | 14:41 | |
| creiht | or somebody was looking in the wrong place? :) | 14:41 |
| *** gundlach has joined #openstack | 14:42 | |
| dendrobates | I think wrong here is subjective, but it was probably me. | 14:42 |
| dendrobates | jaypipes: I don't know about you, but I think the nova code is looking better all the time. I think it would be much easier to be a new dev now than at the beginning of the cycle | 14:44 |
| *** ppetraki has joined #openstack | 14:44 | |
| *** Cybodog has quit IRC | 14:46 | |
| *** KanGouLya has joined #openstack | 14:46 | |
| jaypipes | dendrobates: heck yeah. | 14:47 |
| *** Cybodog has joined #openstack | 14:48 | |
| *** hazmat has joined #openstack | 14:53 | |
| *** sneumann has quit IRC | 14:54 | |
| *** sirp1 has joined #openstack | 14:58 | |
| *** Ryan_Lane has joined #openstack | 15:00 | |
| *** sneumann has joined #openstack | 15:00 | |
| *** khussein_ has joined #openstack | 15:02 | |
| jaypipes | vishy: how did the nebula release go? | 15:03 |
| *** khussein has quit IRC | 15:04 | |
| *** khussein_ is now known as khussein | 15:04 | |
| *** tobym has quit IRC | 15:09 | |
| *** anticw has quit IRC | 15:13 | |
| *** anticw has joined #openstack | 15:13 | |
| *** befreax has quit IRC | 15:14 | |
| *** sneumann has quit IRC | 15:15 | |
| dendrobates | damn launchpad just told me it was down for maintenance, when I tried to send a team email, then sent it 3 times when I reloaded the page. | 15:15 |
| zul | hah | 15:15 |
| *** kevnfx has joined #openstack | 15:16 | |
| *** pharkmillups has joined #openstack | 15:16 | |
| dendrobates | it literally printed 'Message sent to OpenStack Team' 3 times, and I checked and sure enough. | 15:16 |
| *** sneumann has joined #openstack | 15:16 | |
| dendrobates | so sorry for the shotgun email | 15:16 |
| annegentle | heh. yep I've got 3 copies. Bang! :) | 15:16 |
| dendrobates | maybe someone will actually read it, this time! | 15:17 |
| dendrobates | damn, I am having one of those days. | 15:17 |
| *** hazmat has quit IRC | 15:19 | |
| *** AimanA has joined #openstack | 15:20 | |
| *** guynaor has left #openstack | 15:23 | |
| *** guigui1 has quit IRC | 15:27 | |
| *** tobym has joined #openstack | 15:27 | |
| *** DubLo7 has quit IRC | 15:29 | |
| dendrobates | vishy: are you guys around? | 15:29 |
| *** DubLo7 has joined #openstack | 15:29 | |
| *** aliguori has quit IRC | 15:35 | |
| zul | umm...trunk seems to be broken http://paste.ubuntu.com/511738/ | 15:40 |
| *** jsgotangco has joined #openstack | 15:41 | |
| johnpur | dendrobates: read all 3 copies of your email... did the xml namespace get updated as part of the rename of the OpenStack API? | 15:43 |
| dendrobates | no | 15:43 |
| johnpur | ok, that's what i thought, from your email it looked like it did | 15:44 |
| dendrobates | we need the doc url and handoff from rackspace to do that. | 15:44 |
| dendrobates | I just replied to your email | 15:44 |
| johnpur | thx | 15:44 |
| dendrobates | btw, that was launchpad that sent 3 copies not me. | 15:44 |
| johnpur | ha ha, sure sure... my dog hit the keyboard! | 15:45 |
| dendrobates | zul: what revision? | 15:46 |
| zul | 339 | 15:47 |
| *** Erik__ has joined #openstack | 15:49 | |
| *** ecarlin has joined #openstack | 15:51 | |
| *** jonwood has quit IRC | 15:53 | |
| *** jonwood has joined #openstack | 15:54 | |
| *** sneumann has quit IRC | 15:54 | |
| *** anticw has quit IRC | 15:54 | |
| *** anticw has joined #openstack | 15:55 | |
| *** rlucio has joined #openstack | 15:56 | |
| *** ecarlin has quit IRC | 15:58 | |
| *** ecarlin has joined #openstack | 15:59 | |
| *** jonwood_ has joined #openstack | 15:59 | |
| *** jonwood has quit IRC | 15:59 | |
| *** rlucio has quit IRC | 16:02 | |
| *** jonwood_ is now known as jonwood | 16:06 | |
| *** sebastianstadil has joined #openstack | 16:12 | |
| *** hazmat has joined #openstack | 16:13 | |
| *** khussein has quit IRC | 16:15 | |
| *** gasbakid has quit IRC | 16:20 | |
| *** aliguori has joined #openstack | 16:21 | |
| *** Ram___ has joined #openstack | 16:27 | |
| *** Ram___ has left #openstack | 16:27 | |
| gundlach | zul: looks like this has to do with turning internal ids into integers instead of strings... the instance name is set to the internal_id if i recall correctly, and something must be unhappy with that | 16:29 |
| zul | gundlach: agreed | 16:29 |
| gundlach | zul: i did the change to integers; today i was going to make them 64 bit in the database. i'll take a look at fixing the issue you found as well. | 16:30 |
| zul | gundlach: thanks | 16:30 |
| dendrobates | zul: create a bug for it. | 16:32 |
| *** sanjib has joined #openstack | 16:34 | |
| zul | dendrobates: done | 16:35 |
| dendrobates | gracias | 16:35 |
| *** iammartian_ has quit IRC | 16:36 | |
| *** spectorclan has joined #openstack | 16:37 | |
| spectorclan | OpenStack Developers - If you have any questions about Design Summit or need registration help, please let me know | 16:38 |
| openstack | spectorclan: Error: "Developers" is not a valid command. | 16:38 |
| *** hazmat has quit IRC | 16:38 | |
| *** iammartian_ has joined #openstack | 16:38 | |
| gundlach | zul: link? | 16:39 |
| zul | gundlach: https://bugs.edge.launchpad.net/nova/+bug/659330 | 16:40 |
| uvirtbot | Launchpad bug 659330 in nova "Unable to create an instance." [Undecided,New] | 16:40 |
| gundlach | tyvm | 16:40 |
| creiht | whoa | 16:40 |
| creiht | http://www.supermicro.com/products/chassis/4U/847/SC847E16-RJBOD1.cfm | 16:40 |
| uvirtbot | New bug: #659330 in nova "Unable to create an instance." [Undecided,New] https://launchpad.net/bugs/659330 | 16:41 |
| jonwood | I bet the rear drives are a bitch to swap out once you've got it racked. | 16:42 |
| dendrobates | holy crap that is dense | 16:42 |
| *** daleolds has joined #openstack | 16:43 | |
| *** rlucio has joined #openstack | 16:44 | |
| jonwood | But what if 45 hard disks isn't enough? | 16:45 |
| jonwood | http://www.supermicro.com/products/chassis/4U/847/SC847E16-RJBOD1.cfm | 16:45 |
| sanjib | can someone help with a pointer to some link explaining the openstack netwoking model? | 16:45 |
| *** mtaylor has joined #openstack | 16:45 | |
| *** ChanServ sets mode: +v mtaylor | 16:45 | |
| *** ecarlin has quit IRC | 16:46 | |
| *** rlucio_ has joined #openstack | 16:46 | |
| annegentle | sanjib: for OpenStack Compute, which is Nova, here's some info http://nova.openstack.org/network.html | 16:46 |
| sanjib | thanks annegentle, appreciate it | 16:47 |
| creiht | jonwood: yeah but those are 2.5" | 16:47 |
| jonwood | Yeah, I know, not as fast but I guess still handy if you just want a ton of storage for archival. | 16:47 |
| *** ecarlin has joined #openstack | 16:47 | |
| creiht | it isn't so much about spead, but 2.5" drives are expensive, and limited on how much storage you can get out of them | 16:48 |
| dendrobates | 2.5" 15k SaS drives are very fast. | 16:48 |
| *** iammartian_ has quit IRC | 16:48 | |
| *** rlucio has quit IRC | 16:48 | |
| *** rlucio_ is now known as rlucio | 16:48 | |
| dendrobates | they are also very expensive | 16:49 |
| jonwood | Fill it up with SSDs ;) | 16:49 |
| creiht | jonwood: on rear accessability: http://www.supermicro.com/products/Rack/ | 16:49 |
| creiht | hah | 16:49 |
| jonwood | Ok, that's quite cool. | 16:50 |
| sanjib | I know this question may sound stupid to and pardon me for that,is rabbitmq the only available way to communicate between computing nodes and storage? | 16:53 |
| *** pvo has quit IRC | 16:53 | |
| *** khussein has joined #openstack | 16:58 | |
| annegentle | sanjib: no stupid questions around here. :) I think that's true, but I'd defer to others for confirmation. | 16:59 |
| *** tobym is now known as tobym-away | 17:00 | |
| *** khussein has quit IRC | 17:03 | |
| *** khussein has joined #openstack | 17:03 | |
| *** dizz is now known as dizz|away | 17:04 | |
| kevnfx | Is there something that folks here are using to search nova bugs other than https://bugs.edge.launchpad.net/nova? | 17:06 |
| *** tobym-away has quit IRC | 17:07 | |
| annegentle | kevnfx: as far as I know, edge may have outdated data. try https://bugs.launchpad.net/nova instead? | 17:08 |
| rlucio | any word on bug 656012? thats the one that is causing instance launching to fail | 17:11 |
| uvirtbot | Launchpad bug 656012 in nova "euca-run-instances fails" [Undecided,In progress] https://launchpad.net/bugs/656012 | 17:11 |
| rlucio | uh, yea that one | 17:11 |
| kevnfx | Hmm ... I'm seeing the same problem there. That is, 659330 looks very similar to 656012, but when I search for virDomainLookupByName I only see the former (even though the latter has the same method in its trace) ... are there usually inconsistencies like this in the bugs search engine? | 17:11 |
| rlucio | not sure; but yea it 659330 certainly looks like a dupe | 17:12 |
| rlucio | dont forget, they were logged on different days, with different versions of nova, so the trace may be slightly differnt | 17:13 |
| gundlach | rlucio: sorry if i missed the dupe; I just pushed a mergeprop with a fix to both bugs (i assume). https://code.launchpad.net/~gundlach/nova/lp659330/+merge/38251 | 17:14 |
| alekibango | rlucio: sorry for being late with patch | 17:14 |
| alekibango | its mine | 17:14 |
| gundlach | hmm | 17:14 |
| alekibango | but i can tell you the merge request is there | 17:14 |
| alekibango | and it makes it work | 17:14 |
| gundlach | alekibango: what's the link to your patch? did we just step on each other's toes? | 17:14 |
| alekibango | https://code.launchpad.net/~alekibango-danix/nova/lp656012/+merge/37809 | 17:15 |
| alekibango | down there | 17:15 |
| alekibango | i was requested to make some testCase | 17:15 |
| alekibango | so it will be tested, but i am still fighting with many dogs out there | 17:16 |
| rlucio | in regard to this bug? or you mean you didnt have time to write the test? | 17:16 |
| alekibango | well i wrote it, but it given strange results :) | 17:16 |
| alekibango | when debugging it, many people started to distract me, so i had not enough time to finish it | 17:16 |
| alekibango | will try now | 17:17 |
| gundlach | alekibango: i'm going to go ahead and push my (approved) mergeprop to get this over with. if you can get a working test case, that'll be a great addition; but zul just verified that this fixes his trace. | 17:17 |
| alekibango | gundlach: you also fix this in your code? | 17:17 |
| alekibango | how? which? | 17:17 |
| alekibango | i have problem debugging tests... :) | 17:18 |
| gundlach | alekibango: yep -- i missed rlucio's report and your code entirely. | 17:18 |
| gundlach | https://code.launchpad.net/~gundlach/nova/lp659330/+merge/38251 | 17:18 |
| gundlach | fixed it the same way you did basically. | 17:18 |
| alekibango | this was my first attempt to post something on LP.. :) at least i learned how to do it | 17:19 |
| alekibango | imho your patch is better, | 17:19 |
| alekibango | i will write the test | 17:19 |
| gundlach | well it certainly has less contention in its comments! ;) | 17:19 |
| alekibango | so please make trunk working if you can | 17:20 |
| gundlach | alekibango: it should merge in a few minutes and trunk will be fixed. | 17:20 |
| rlucio | thanks guys! | 17:20 |
| alekibango | gundlach: they just took opportunity to hastle a novice to make him more productive, heh | 17:20 |
| gundlach | if you pull from trunk into your patch and resolve the conflict b/w your code and mine, then you'll be able to write the test and submit just that for merge to trunk. | 17:20 |
| gundlach | :) | 17:20 |
| alekibango | and i learned something, really | 17:20 |
| alekibango | gundlach: will do | 17:21 |
| alekibango | or another fix/merge? just for the test | 17:21 |
| alekibango | whatever, i need to finish the test first | 17:21 |
| alekibango | gundlach: howto debug failing test should be part of nova docs | 17:21 |
| alekibango | my code went well - but i got some errors elsewhere (heh) | 17:22 |
| gundlach | alekibango: i may be running into the same thing you were. i'm getting a test failure in test_run_instance_existing. | 17:23 |
| gundlach | time to dig... | 17:23 |
| alekibango | well, i am too new to nova to dig fast:) | 17:24 |
| *** DubLo7 has joined #openstack | 17:24 | |
| gundlach | alekibango: don't worry, i'm less new to it and i still dig very slowly ;) | 17:25 |
| dendrobates | gundlach: did you see the failed tests from hudson on your last merge? | 17:26 |
| *** jc_smith has joined #openstack | 17:26 | |
| gundlach | dendrobates: yep, digging | 17:26 |
| alekibango | gundlach: ack. life is not easy... :) | 17:26 |
| gundlach | for whatever reason, running the test suite on my box is way slower than i saw it run on todd's box once, so i'm twiddling my thumbs atm | 17:26 |
| sanjib | #openstack-meeting | 17:26 |
| alekibango | sanjib: few hours later... | 17:27 |
| *** hazmat has joined #openstack | 17:27 | |
| *** gustavomzw has quit IRC | 17:28 | |
| *** pvo has joined #openstack | 17:29 | |
| *** ChanServ sets mode: +v pvo | 17:29 | |
| *** rlucio has quit IRC | 17:30 | |
| *** pvo has quit IRC | 17:31 | |
| *** pvo has joined #openstack | 17:32 | |
| *** pvo has joined #openstack | 17:32 | |
| *** ChanServ sets mode: +v pvo | 17:32 | |
| *** guynaor has joined #openstack | 17:35 | |
| sanjib | if i write code base for nova in erlang will it be outright rejected? | 17:37 |
| *** kashyapc has quit IRC | 17:37 | |
| *** kashyapc has joined #openstack | 17:39 | |
| *** cdub has joined #openstack | 17:42 | |
| pvo | sanjib: I don't think it would get accepted | 17:42 |
| sanjib | we are ready to provide python integration interfaces, | 17:43 |
| sanjib | probably we could write up a mulitlanguage interface blueprint | 17:44 |
| sanjib | and see if there is traction to the idea. | 17:45 |
| pvo | sanjib: out of curiosity, what would be written in erlang? | 17:45 |
| sanjib | pvo: an auto scalable load balancer using an architecture similar to that of rabbitmq | 17:46 |
| pvo | sanjib: to load balance traffic to the VMs? or just to pass messages as rabbit does? | 17:47 |
| sanjib | pvo: idea is to provide scalability by design and not through api controls | 17:47 |
| sanjib | pvo: load balance traffic to instances | 17:47 |
| pvo | sanjib: I would think that would sit on top of nova. Those are two different problems that are being solved. I agree it sounds interesting and it is needed but I don't see that as a nova specific problem | 17:49 |
| dendrobates | are we talking about a load balancer as a service? | 17:49 |
| pvo | dendrobates: I thinkn so | 17:49 |
| *** burris has quit IRC | 17:50 | |
| dendrobates | sanjib: things that are consumed by openstack or are auxiliary can be written in other languages | 17:50 |
| dendrobates | sanjib: we just don't want to require a 1Gb download to get dependancies for 4 different languages to run openstack. | 17:51 |
| sanjib | pvo: here is the thing nginx load balancer can route traffic to instances, imagine a world where we will not need to route traffic if more nodes are available the server becomes aware of it and distributes load | 17:51 |
| sanjib | dendrobates: understood , but we are already using rabbitmq and hence are already using erlang | 17:52 |
| sanjib | dendrobates: the language dependency is already taken care of | 17:53 |
| pvo | that is a use case of nova. But again, i see that as something that would sit on top. | 17:53 |
| dendrobates | sanjib: I am saying it is fine. | 17:53 |
| dendrobates | yes, on top can be whatever. | 17:54 |
| *** dendrobates is now known as dendro-afk | 17:55 | |
| *** dendro-afk is now known as dendrobates | 17:58 | |
| *** kevnfx has quit IRC | 17:58 | |
| *** Ryan_Lane is now known as Ryan_Lane|food | 17:59 | |
| jaypipes | gundlach: heh, how many times has the str_id/instance.name changed in the last week? :) | 18:07 |
| gundlach | jaypipes: 6? | 18:07 |
| *** rlucio has joined #openstack | 18:07 | |
| *** daleolds has quit IRC | 18:08 | |
| gundlach | jaypipes: i bet it's 6. | 18:09 |
| jaypipes | gundlach: :) | 18:10 |
| *** daleolds has joined #openstack | 18:11 | |
| *** BK_man has joined #openstack | 18:13 | |
| BK_man | hi all! | 18:13 |
| BK_man | does Openstack support elastic IP addresses? | 18:14 |
| BK_man | (in AWS definition) | 18:14 |
| *** tobym-away has joined #openstack | 18:14 | |
| *** sirp1 has quit IRC | 18:15 | |
| *** tobym-away is now known as tobym | 18:15 | |
| *** burris has joined #openstack | 18:18 | |
| *** dendrobates is now known as dendro-afk | 18:20 | |
| alekibango | http://aws.amazon.com/ec2/ -- contains definition of "elastic ip address" | 18:21 |
| alekibango | An Elastic IP address is associated with your account not a particular instance, and you control that address until you choose to explicitly release it. | 18:21 |
| *** daleolds has quit IRC | 18:21 | |
| *** sneumann has joined #openstack | 18:22 | |
| *** dendro-afk is now known as dendrobates | 18:23 | |
| *** guynaor has quit IRC | 18:24 | |
| BK_man | alekibango: ok. so I suppose that there is no such feature in Openstack at the moment | 18:25 |
| alekibango | BK_man: imho there is not, but i must admit i do not know enough yet | 18:25 |
| alekibango | but i am trying to help you by forming the question | 18:25 |
| *** jsgotangco_ has joined #openstack | 18:26 | |
| *** jsgotangco_ has quit IRC | 18:26 | |
| *** jsgotangco_ has joined #openstack | 18:26 | |
| *** guynaor has joined #openstack | 18:26 | |
| *** jsgotangco has quit IRC | 18:26 | |
| *** jsgotangco_ is now known as jsgotangco | 18:26 | |
| *** cdub has quit IRC | 18:26 | |
| *** burris has quit IRC | 18:26 | |
| *** jsgotangco has quit IRC | 18:27 | |
| alekibango | BK_man: this might improve your chances to get good answers :) http://www.catb.org/esr/faqs/smart-questions.html | 18:27 |
| jonwood | It seemed like a reasonable question to me. | 18:27 |
| alekibango | yes it is | 18:27 |
| jonwood | I can't answer it, but given the audience here, I'd expect people to know what elastic IPs are. | 18:27 |
| alekibango | well i dont :) | 18:28 |
| alekibango | maybe thats only me, heh | 18:29 |
| *** herki has joined #openstack | 18:29 | |
| alekibango | jonwood: and as i would like to put that in wiki, i need good definition :) | 18:30 |
| dendrobates | expect shared ip groups in the next 2 releases, but elastic IPs are not on the roadmap yet. | 18:30 |
| BK_man | alekibango: thanks for suggestion. I actually translated that document to Russian some time ago :) | 18:30 |
| alekibango | BK_man: :) | 18:30 |
| alekibango | i dont like the author much, but this really inspired me | 18:31 |
| BK_man | we have a customer (a major one) that asked us to compare Openstack & vCloud. That's why I'm asking here - suppose to get answers faster than grep'ing sources :) | 18:32 |
| alekibango | dendrobates: shared ip groups? how that will work? | 18:32 |
| BK_man | next stage will be clarifying what need to be done in OpenStack and provide that funtionality in opensourse manner! | 18:33 |
| alekibango | BK_man: what i noticed is limits on scalability. | 18:33 |
| alekibango | (difference) | 18:33 |
| alekibango | openstack should be 100+ times scalable | 18:33 |
| openstack | alekibango: Error: "should" is not a valid command. | 18:34 |
| dendrobates | alekibango: http://www.rackspacecloud.com/blog/2010/03/03/networking-and-cloud-servers-%E2%80%93-shared-ip-groups/ | 18:34 |
| alekibango | oh that bot is funny | 18:34 |
| alekibango | it should have another name or expect openstack: at the start of the line | 18:34 |
| alekibango | dendrobates: ty | 18:34 |
| *** littleidea has joined #openstack | 18:36 | |
| sanjib | opencloud need not be yaaa( yet another aws aws ) | 18:37 |
| alekibango | BK_man: i believe that after comparison you will join the winning team :) | 18:38 |
| BK_man | alekibango: we (as a company) have our own commitment to opensource cloud platform development. Open is a key | 18:39 |
| alekibango | great :) | 18:39 |
| BK_man | alekibango: I'm team lead for IaaS at the moment and my task now is understand what missing in Openstack and what needs to be done to cover these gaps in functionality | 18:40 |
| alekibango | ok, good. please give us all your questions, :) i will do my best to make answers documented in openstack wiki | 18:41 |
| alekibango | (which is pretty boring atm) | 18:41 |
| alekibango | i mean nova manuals | 18:41 |
| BK_man | is Openstack supporting VMs live migration at the moment (suppose that we are using KVM) | 18:41 |
| BK_man | alekibango: atm == at the moment ? | 18:42 |
| alekibango | yes, we just started writing user/admin/deployment docs | 18:42 |
| alekibango | and your questions might help actually to write what is needed most | 18:42 |
| *** jc_smith has quit IRC | 18:42 | |
| BK_man | I hope that you answered my first question about live migration :) | 18:43 |
| alekibango | no, but i will. imho it is not supporting it atm, but i am not sure.. so i wait for someone smarter :) | 18:43 |
| alekibango | i am not able to be good spokeperson for nova :) | 18:44 |
| BK_man | any kind of VM storage migration? | 18:44 |
| dendrobates | live migration requires shared storage, which is implementation specific. We will eventually provide migration tools | 18:44 |
| alekibango | BK_man: i believe that is planned, but not for the coming austin release | 18:44 |
| BK_man | (kind of VMware - from one datastore to another while VM is running) | 18:44 |
| BK_man | alekibango: ok | 18:44 |
| alekibango | there definitelly are people who would like to have it | 18:45 |
| *** ctennis has quit IRC | 18:45 | |
| alekibango | BK_man: prolly sheepdog might be used by some | 18:45 |
| alekibango | but there are even more possibilities | 18:45 |
| alekibango | you know, openstack is pretty young now | 18:45 |
| alekibango | BK_man: even glance (swift integration with nova) will not be there for austin release... | 18:46 |
| *** neogenix has joined #openstack | 18:46 | |
| *** daleolds has joined #openstack | 18:47 | |
| alekibango | BK_man: what i value most on openstack: 1) open, adaptable, scalable 2) smart people 3) smart companies 4) mood, spirit 5) python, amqp, sqlalchemy | 18:48 |
| alekibango | i think this list is in order as i value those | 18:48 |
| *** hazmat has quit IRC | 18:49 | |
| alekibango | not to mention 6) design tenets :) but that comes from 2 and 3 | 18:49 |
| *** BK_man has quit IRC | 18:49 | |
| *** BK_man has joined #openstack | 18:49 | |
| BK_man | too bad. My ADSL link goes down and up :( | 18:50 |
| alekibango | np | 18:50 |
| alekibango | dendrobates: you can quote me :) i approve this message :) | 18:50 |
| BK_man | last question was: member:dendrobates: correct. That's lead us to another question - what kind of storage attachments are now supported (in addition to local storage)? | 18:50 |
| BK_man | NFS, some kind of distributed FS, iSCSI? | 18:50 |
| alekibango | BK_man: see glance subproject | 18:51 |
| neogenix | Sheepdog ;) | 18:51 |
| BK_man | alekibango: subproject of swift? | 18:51 |
| creiht | swift will not be used for that type of storage | 18:51 |
| alekibango | BK_man: few linsk for you | 18:51 |
| alekibango | https://code.launchpad.net/nova | 18:52 |
| dendrobates | anything your host os supports as block store. i.e iscsi, raw disk, nfs... this is outside of openstack compute. | 18:52 |
| alekibango | BK_man: i like for example glusterfs, until ceph will be mature :) | 18:52 |
| BK_man | dendrobates: so, it MUST be block storage (not SMB, NFS, etc) and be available on hosts. | 18:53 |
| alekibango | BK_man: no need for blocks :) just filesystem is ok | 18:54 |
| alekibango | like lustre | 18:54 |
| alekibango | ceph or wahtever | 18:54 |
| dendrobates | anything your os sees as a unix filesystem is OK | 18:55 |
| dendrobates | smb might not work correctly | 18:55 |
| dendrobates | but it might | 18:55 |
| alekibango | i would not dare to use smb :) | 18:55 |
| BK_man | alekibango: ok. So, if I understand correctly, swift is NOT used as storage for VMs images | 18:56 |
| alekibango | no, its there for images (with glance) | 18:56 |
| * BK_man VM images != images ? | 18:56 | |
| alekibango | (see glance in blueprints) | 18:56 |
| BK_man | alekibango: ok | 18:56 |
| alekibango | well not much valuable info there... :) | 18:57 |
| alekibango | jaypipes: could you please better describe what glance is in the blueprint? | 18:57 |
| alekibango | "image registry and storage service" is not enough for me :) | 18:58 |
| alekibango | it could mean different things | 18:58 |
| BK_man | yep, same for me | 18:58 |
| alekibango | some use cases would be great | 18:58 |
| alekibango | BK_man: jaypipes might be afk, but he will help, you need just wait :) | 18:59 |
| BK_man | another question: is there a "VMs bundle" (like VMware vApp) - several VMs with networking settings attached, dependencies for powering these VMs on, etc? | 18:59 |
| BK_man | this is requires for PaaS implementation | 19:00 |
| alekibango | afaik not yet | 19:00 |
| BK_man | alekibango: ok | 19:00 |
| dendrobates | we are just 10 days from release so most of the devs are busy with last minute coding. We will be more responsive after the release. | 19:00 |
| alekibango | but is not that hard to do imho | 19:00 |
| alekibango | dendrobates: i know, trying to help as much as i can | 19:00 |
| alekibango | :) | 19:00 |
| *** sirp1 has joined #openstack | 19:01 | |
| BK_man | dendrobates: sorry for disturbing, I also have my own deadlines for my paper :( | 19:01 |
| BK_man | alekibango: thanks a lot! your answers are invaluable | 19:01 |
| alekibango | btw tonight is a meeting | 19:02 |
| dendrobates | np, I just don;t want people thinking we are being rude. | 19:02 |
| *** ctennis has joined #openstack | 19:02 | |
| *** ctennis has joined #openstack | 19:02 | |
| BK_man | alekibango: I can't attend, it will be 1AM in Moscow and I need to stand up at 6AM to go to job | 19:02 |
| alekibango | BK_man: see? they are trying hard to be nice ;) thats the spirit i listed in reasons for openstack | 19:03 |
| BK_man | alekibango: :) | 19:03 |
| BK_man | alekibango: what are user authentication methods in nova? | 19:04 |
| BK_man | alekibango: VMware can do Active Directory & OpenLDAP | 19:04 |
| *** gaveen has joined #openstack | 19:04 | |
| *** gaveen has joined #openstack | 19:04 | |
| alekibango | there is ldap or emulation of ldap (users in database) | 19:04 |
| BK_man | alekibango: ok | 19:05 |
| alekibango | emulation is default and is easier for me :) | 19:05 |
| rlucio | hey guys, i am looking at the volume manager, i am just noticing that you can no longer specify a physical volume to use with aoe | 19:05 |
| creiht | It shouldn't be difficult to plug in other mehtods of authentication | 19:05 |
| rlucio | why was that functionality removed? | 19:05 |
| alekibango | BK_man: yes, nova is extremely adoptable | 19:06 |
| alekibango | adaptible? | 19:06 |
| rlucio | (removed in orm merge, rev 275) | 19:06 |
| alekibango | lol english is not my primary language, i sometimes invent my own words | 19:06 |
| alekibango | rlucio: i would ask the one who did the merged branch | 19:06 |
| rlucio | ah yea, i was just thinking that myself | 19:07 |
| jonwood | alekibango: adaptable | 19:07 |
| pvo | BK_man: re: your glance question. We need a place to store the base images for Nova. We intend on using Swift to store these images, but glance will be the system that will receive the request, then check if you are authorized for this image, then push the image to the host. Or some derivative. | 19:07 |
| alekibango | jonwood: ty :) | 19:07 |
| alekibango | pvo: now questions: waht is an image | 19:08 |
| BK_man | pvo: ok | 19:08 |
| rlucio | vishy: did you take out the physical volume support from nova-volume (rev 275) on purpose? or did that functionality move somewhere else? | 19:08 |
| alekibango | 2) can it be used (in future?) for storing live images over network, so we will be able to migrate instances? | 19:08 |
| jonwood | alekibango: That would be a template filesystem for a virtual machine usually. | 19:08 |
| pvo | alekibango: an image (for me at least) is a snapshot of a vm. The file(s) that make up a distro | 19:08 |
| alekibango | so disk image, iso image, etc | 19:08 |
| pvo | alekibango: yes, but it may not be in iso format. | 19:09 |
| pvo | but the idea is similar | 19:09 |
| alekibango | ha, which formats then? | 19:09 |
| alekibango | thats what i was asking :) | 19:09 |
| pvo | its up to you and your hypervisor | 19:09 |
| pvo | alekibango: there are lots of tradeoffs. | 19:09 |
| alekibango | you mean 2) ? wouldn't it be nice to have sheepdog like functionality using swift? why? | 19:10 |
| pvo | We are looking to use Xen and vhds which give us the ability to do snapshots and coalescing at the vdi layer | 19:10 |
| pvo | sheepdog is built for KVM at the moment which does me no good | 19:10 |
| pvo | but it might for you | 19:10 |
| alekibango | pvo: i dont mind using kvm, as it is slowly taking lead :) | 19:11 |
| pvo | alekibango: sure, I recognize that. | 19:11 |
| alekibango | still i know that xen could be using resources better a bit | 19:11 |
| pvo | alekibango: We run a pretty large xen server installation | 19:12 |
| alekibango | ic | 19:12 |
| alekibango | so - do you think its possibleto make sheepdog like storage using swift? | 19:12 |
| *** cdub has joined #openstack | 19:12 | |
| pvo | alekibango: no, thats not really what it does | 19:12 |
| alekibango | why ? can you please explain a bit? | 19:13 |
| alekibango | its not for high performance? | 19:13 |
| pvo | alekibango: I think of swift as a bunch of shelves I can put stuff on. For example, I have lots of things to store that I might need access to every so often, but not real time read/write. | 19:13 |
| alekibango | there could be some (fuse, heh) driver which might see some swift space as filesystem, right? | 19:14 |
| alekibango | so you say that would be slow.... | 19:14 |
| pvo | so in that case, I can take images/snapshots/iso and store them there for a long time then retrieve them as needed. | 19:14 |
| *** rlucio has quit IRC | 19:14 | |
| alekibango | that swift is not taht good for realtime changes | 19:14 |
| pvo | I *think* there is a fuse system for swift, but I wouldn't use it. I would expect it to be pretty slow. | 19:14 |
| alekibango | ok, thats reasoning i can accept :) thanks a lot | 19:15 |
| _cerberus_ | alekibango: One of the devs actually did make a FUSE implementation a while back http://github.com/redbo/cloudfuse/ | 19:15 |
| alekibango | lots of people are struggling with understanding this | 19:15 |
| pvo | swift == S3 | 19:15 |
| pvo | in concept | 19:15 |
| pvo | not api | 19:15 |
| *** hazmat has joined #openstack | 19:15 | |
| alekibango | ok pvo, thanks | 19:16 |
| alekibango | i will not drain your valuable time:) | 19:16 |
| uvirtbot | New bug: #659431 in swift "default backlog option needs to be added to the sample config files" [Undecided,New] https://launchpad.net/bugs/659431 | 19:16 |
| *** mtaylor has quit IRC | 19:17 | |
| *** cdub has quit IRC | 19:18 | |
| creiht | alekibango: swift isn't designed to do block storage, and the way we handle consistency in listings would make it difficult to work as block storage | 19:18 |
| jdarcy | I've actually thought about adapting VoldFS to use Swift as a (potential) back end. | 19:21 |
| jdarcy | Is there some sort of conditional-put support, like with If-None-Match or such? | 19:21 |
| creiht | I would not rule out that one could take some of the swift concepts and build block storage with it | 19:22 |
| creiht | jdarcy: no conditional puts yet | 19:22 |
| creiht | though that is something we have thought about | 19:23 |
| jdarcy | That makes it harder. Adding that would have to be part of the work. :) | 19:23 |
| gundlach | anyone around who can review https://code.launchpad.net/~gundlach/nova/64-bit-internal-ids/+merge/38262 for me? | 19:23 |
| johnpur | making swift a realistic block device would necessitate a good caching layer | 19:23 |
| sebastianstadil | Block storage (like EBS) will be part of Nova then, not Swift? | 19:24 |
| creiht | the other possibility is bolting something on top of swift that would be the hot layer and having swift be the cold storage, like johnpur is talking about | 19:24 |
| johnpur | block storage may be a separate project that complements nova | 19:25 |
| creiht | a lot about block storage is yet to be determined | 19:25 |
| *** abecc_ has joined #openstack | 19:25 | |
| *** abecc has quit IRC | 19:26 | |
| *** abecc_ is now known as abecc | 19:26 | |
| dendrobates | I personally see an EBS like solution that provides a pluggable backend that supports swift among other things. | 19:26 |
| creiht | unless some awesome open source block storage project magically appears :) | 19:26 |
| jdarcy | You mean better than Sheepdog or RBD? | 19:27 |
| *** hazmat has quit IRC | 19:27 | |
| *** jtimberman has quit IRC | 19:27 | |
| *** hisaki has quit IRC | 19:27 | |
| creiht | jdarcy: I imagine those will all be explored | 19:27 |
| jdarcy | Is block storage likely to be on the agenda at summit? | 19:27 |
| *** jtimberman has joined #openstack | 19:27 | |
| * creiht notes that this is all conjecture on my part | 19:28 | |
| *** Ryan_Lane|food is now known as Ryan_Lane | 19:28 | |
| dendrobates | we'll have to see how many blueprints get submitted | 19:28 |
| dendrobates | we are only going to discuss things that we think we can get done in the next 6 months | 19:28 |
| johnpur | block storage is very high on the priority list, it is likely to be a point of discussion. | 19:29 |
| johnpur | Rackspace wants a solution within the next 6 months and has folks looking at possible solutions currently | 19:30 |
| uvirtbot | New bug: #659445 in glance "Swift Backend URL parser breaks with Python 2.6.1" [Undecided,New] https://launchpad.net/bugs/659445 | 19:31 |
| *** irahgel has quit IRC | 19:31 | |
| sebastianstadil | I can imagine Rackspace wanting block storage sooner rather than later. It makes automating databases much better. Is Rackspace leaning towards creating a new project so far? | 19:34 |
| sebastianstadil | dendrobates: how many blueprints have been submitted? | 19:35 |
| *** gustavomzw has joined #openstack | 19:36 | |
| dendrobates | sebastianstadil: haven;t looked yet. We haven't made a call for submissions yet. We want people to focus on the release. | 19:36 |
| *** electroniceagle has joined #openstack | 19:37 | |
| *** electroniceagle has left #openstack | 19:38 | |
| sanjib | dendrobates: is there a decided protocol that projects spawned around openstack need to follow to ensure interoperability | 19:38 |
| pvo | sanjib: isn't that api compatibility? | 19:39 |
| dendrobates | sanjib: no, but once we stabilize, we will not break the API contract | 19:39 |
| pvo | sanjib: though I know there could be some other thing such as back end integration | 19:40 |
| *** allsystemsarego has quit IRC | 19:40 | |
| *** cdub has joined #openstack | 19:41 | |
| *** tobym has quit IRC | 19:41 | |
| *** tobym has joined #openstack | 19:41 | |
| *** jc_smith has joined #openstack | 19:43 | |
| sanjib | pvo:api compatibility is great, but there sure could be ideas that come around where backend integration ( some way of direct access to system assets ) would help performance | 19:44 |
| *** jc_smith_ has joined #openstack | 19:45 | |
| sanjib | one immediate think of is some kind of resource like cloudwatch | 19:45 |
| pvo | sanjib: back end integration should revolve around some api as well | 19:46 |
| *** gondoi has joined #openstack | 19:47 | |
| pvo | sanjib: possibly. Patches accepted. : ) | 19:47 |
| *** jc_smith has quit IRC | 19:48 | |
| *** jc_smith_ is now known as jc_smith | 19:48 | |
| *** ArdRigh has joined #openstack | 19:50 | |
| *** cdub has quit IRC | 19:53 | |
| *** pharkmillups has quit IRC | 19:54 | |
| *** aliguori has quit IRC | 19:55 | |
| *** mtaylor has joined #openstack | 19:55 | |
| *** ChanServ sets mode: +v mtaylor | 19:55 | |
| *** daleolds has quit IRC | 19:56 | |
| *** daleolds has joined #openstack | 19:57 | |
| *** hazmat has joined #openstack | 19:58 | |
| jonwood | Should nova-compute --help work using the maverick packages? | 20:00 |
| *** hazmat has quit IRC | 20:02 | |
| *** sirp1 has quit IRC | 20:02 | |
| *** ecarlin has quit IRC | 20:03 | |
| soren | jonwood: "should" is an interesting word. | 20:05 |
| jonwood | Heh, let me rephrase. | 20:06 |
| jonwood | Is it a known bug that it doesn't. | 20:06 |
| soren | Yes. | 20:06 |
| *** ecarlin has joined #openstack | 20:06 | |
| soren | Bug #647728 | 20:06 |
| uvirtbot | Launchpad bug 647728 in nova "--help is non-functional" [Low,Confirmed] https://launchpad.net/bugs/647728 | 20:06 |
| soren | jonwood: ^ | 20:07 |
| jonwood | Cool, just checking | 20:07 |
| soren | no problem at all. | 20:07 |
| *** aliguori has joined #openstack | 20:10 | |
| jonwood | Are command line flags stable enough to be documented? If so, I'll pull them out of the source, and put some pages up on the wiki. | 20:10 |
| zykes- | does openstack need to talk to a host directly | 20:12 |
| zykes- | or can it use a vCenter server f ex in vmwarE ? | 20:13 |
| neogenix | or SCVMM, or XenCenter(re) | 20:15 |
| alekibango | hi jdarcy :) | 20:16 |
| *** daleolds has quit IRC | 20:16 | |
| jdarcy | Hi alekibango. | 20:16 |
| alekibango | thansk to all for insights. i will try to document this somewhere :) | 20:16 |
| *** sneumann has quit IRC | 20:16 | |
| alekibango | zykes-: what do you mean openstack talk to a host directly? | 20:19 |
| alekibango | there is nova manager on every host... and cluster manager and cloud manager | 20:19 |
| alekibango | zykes-: see http://wiki.openstack.org/Overview | 20:20 |
| zykes- | zykes-: cause in vmware you tend to use vcenter for management | 20:21 |
| zykes- | and that doesn't excactly take a linux daemon on it :p | 20:21 |
| soren | Alright, nobody touch anything. | 20:21 |
| jonwood | alekibango: Not neccesarily. As I understand it in larger clusters you'll have nova-compute running on every host node, but only one (possibly HA) cluster and cloud manager. | 20:21 |
| alekibango | in nova, you talk to cluster controller | 20:21 |
| alekibango | and cluster controller will do it for you | 20:21 |
| soren | I'm merging my security groups branch and if someone gets stuff into trunk ahead of me that breaks the merge or the unit tests, I /will/ start striking names off my christmas card list. Be warned! | 20:22 |
| alekibango | jonwood: if you will have distributed machines, you will put them in a cluster - one or more on each place | 20:22 |
| alekibango | imagine having some servers in africa and some in china :) | 20:23 |
| zykes- | anyone knows ? | 20:23 |
| zykes- | that know a bit of vmware :p | 20:23 |
| alekibango | zykes-: i dont know a bit about vmware as i ignore them | 20:23 |
| alekibango | so i dont know what you are asking about | 20:23 |
| alekibango | maybe i already answered | 20:23 |
| jonwood | zykes-: I know what you mean by vCenter, but I have no idea if it can integrate with Nova. | 20:23 |
| jonwood | But if it has an API, it's probably possible to communicate with vCenter rather then the host nodes directly. | 20:24 |
| jonwood | How hard that would be is a different question ;) | 20:24 |
| alekibango | maybe vcenter = cloud controller | 20:24 |
| alekibango | (aka public api server) | 20:24 |
| zykes- | alekibango: you can control all the nodes from there yes | 20:25 |
| zykes- | it exposes the same "api" as the ESX boxes | 20:25 |
| alekibango | zykes-: look at http://wiki.openstack.org/Overview http://wiki.openstack.org/ArchitecturalOverview | 20:25 |
| alekibango | there si public api server | 20:25 |
| jonwood | zykes-: You could probably put it in as a backend for nova-compute, like they do with KVM and Xen at the moment. | 20:26 |
| alekibango | that uses amqp to talk to controllers | 20:26 |
| zykes- | k | 20:26 |
| alekibango | you can have more public api servers | 20:26 |
| alekibango | and you do not need to talk directly to cluster controllers or real servers as user afaik | 20:27 |
| jaypipes | gundlach: heya, if you're interested, would be great to get your feedback on https://code.launchpad.net/~jaypipes/glance/parallax-register-image/+merge/38258 | 20:27 |
| * jonwood goes to get dinner | 20:27 | |
| alekibango | you just talk to public api server | 20:27 |
| gundlach | jaypipes: i'm in the middle of that now, actually :) | 20:27 |
| *** tobym has left #openstack | 20:27 | |
| jaypipes | gundlach: ah, sweeeeeet :) | 20:27 |
| *** burris has joined #openstack | 20:27 | |
| soren | Security Groups merged! | 20:29 |
| soren | Man, that was no fun at all. | 20:29 |
| *** KanGouLya has quit IRC | 20:29 | |
| alekibango | soren: sounds like *fun* | 20:29 |
| alekibango | :) | 20:29 |
| * soren makes a rude gesture | 20:29 | |
| alekibango | soren: dont be angry with me, just joking :) | 20:30 |
| soren | :) | 20:31 |
| gundlach | jaypipes: approved with a couple of comments that could be addressed | 20:31 |
| jaypipes | gundlach: heya, thx :) | 20:31 |
| gundlach | np | 20:32 |
| jaypipes | soren: yeah! \o/ | 20:32 |
| *** joearnold has joined #openstack | 20:32 | |
| jaypipes | soren: only what, 10 times the charm? ;P | 20:32 |
| soren | jaypipes: srsly. 10 times merging with trunk, 5-6 times merging with other stuff.. That thing touched stuff *everywhere*, apparantly. | 20:33 |
| jaypipes | soren: yup. | 20:33 |
| alekibango | soren: and trunk is now working? | 20:35 |
| jaypipes | gundlach: for some reason I *always* forget about setdefault... | 20:36 |
| soren | alekibango: It wasn't before? | 20:37 |
| *** kevnfx has joined #openstack | 20:38 | |
| alekibango | soren: will try :) | 20:38 |
| *** jc_smith has quit IRC | 20:42 | |
| *** jbryce has joined #openstack | 20:44 | |
| alekibango | please someone help me list some noticeable nova austin release features.... | 20:44 |
| alekibango | for upcoming wiki page... | 20:45 |
| annegentle | alekibango: also refer to or update http://wiki.openstack.org/ReleaseNotes/Austin | 20:45 |
| alekibango | good one :) | 20:46 |
| dendrobates | annegentle: when we get to the documentation topic in the meeting, can I just turn it over to you? | 20:50 |
| annegentle | dendrobates: sure, no problem | 20:50 |
| alekibango | i will be also there :) | 20:50 |
| alekibango | (i put it on the list) | 20:50 |
| *** mtaylor has quit IRC | 20:51 | |
| dendrobates | alekibango: I saw that thanks. I love that a community member added an agenda item. I hope others will follow | 20:51 |
| alekibango | :) | 20:51 |
| *** khussein_ has joined #openstack | 20:53 | |
| *** mtaylor has joined #openstack | 20:54 | |
| *** ChanServ sets mode: +v mtaylor | 20:54 | |
| *** khussein has quit IRC | 20:55 | |
| *** khussein_ is now known as khussein | 20:55 | |
| alekibango | meeting starts soon | 20:58 |
| *** jc_smith has joined #openstack | 21:00 | |
| *** rlucio_ has joined #openstack | 21:01 | |
| *** adjohn has joined #openstack | 21:02 | |
| *** aliguori has quit IRC | 21:04 | |
| *** aliguori has joined #openstack | 21:04 | |
| *** aimon has quit IRC | 21:04 | |
| *** aimon has joined #openstack | 21:04 | |
| *** pharkmillups has joined #openstack | 21:06 | |
| *** rlucio_ has quit IRC | 21:06 | |
| *** DubLo7 has quit IRC | 21:10 | |
| *** miclorb has joined #openstack | 21:16 | |
| *** faccenda has quit IRC | 21:19 | |
| *** sandywalsh has joined #openstack | 21:19 | |
| *** faccenda has joined #openstack | 21:19 | |
| *** sirp1 has joined #openstack | 21:24 | |
| *** devcamcar has joined #openstack | 21:25 | |
| *** devcamcar has left #openstack | 21:26 | |
| *** jdarcy has quit IRC | 21:29 | |
| *** sneumann has joined #openstack | 21:30 | |
| *** sneumann has quit IRC | 21:37 | |
| *** daleolds has joined #openstack | 21:40 | |
| *** guynaor has left #openstack | 21:40 | |
| *** ppetraki has quit IRC | 21:43 | |
| *** faccenda has quit IRC | 21:43 | |
| *** spectorclan has quit IRC | 21:50 | |
| *** khussein has quit IRC | 21:51 | |
| *** AimanA has quit IRC | 21:52 | |
| dendrobates | good meeting people | 21:53 |
| gundlach | good meeting you too, dendrobates ;) | 21:53 |
| dendrobates | :) | 21:54 |
| *** DubLo7 has joined #openstack | 22:01 | |
| *** gondoi has quit IRC | 22:01 | |
| jonwood | The wiki claims that Nova needs to be started with --FAKE_subdomain=ec2 - is that strictly true, or can you set it use the openstack API instead? | 22:01 |
| *** irahgel has joined #openstack | 22:02 | |
| soren | jonwood: The latter. | 22:02 |
| *** Ryan_Lane has quit IRC | 22:02 | |
| jonwood | More to the point, is it even true that it needs running with --FAKE_subdomain. Could you just point subdomains at it, and have it work with both? | 22:04 |
| kevnfx | Is this more trunk breakage ... or just my installation? http://pastebin.com/cnvZ63hT | 22:04 |
| soren | jonwood: Yes. | 22:05 |
| * soren glances at gundlach | 22:05 | |
| jonwood | So the only reason it says to use FAKE_subdomain is because it's assumed people won't have the subdomains set up yet? | 22:06 |
| *** astrolin has joined #openstack | 22:06 | |
| soren | jonwood: Yes. | 22:07 |
| gundlach | jonwood: you can also set e.g. ec2api_subdomain to your main domain if you only plan to serve up the EC2 API, or osapi_subdomain if only serving up the OpenStack API. | 22:07 |
| *** Ryan_Lane has joined #openstack | 22:08 | |
| jonwood | That makes sense. Thanks. | 22:08 |
| soren | gundlach: kevnfx's pastebin above looks like something you'd know about. | 22:08 |
| gundlach | soren: yeah, taking a look now | 22:09 |
| gundlach | consarn it | 22:09 |
| *** irahgel has left #openstack | 22:09 | |
| gundlach | i keep relying on unit test coverage and i keep breaking trunk as a result. | 22:09 |
| gundlach | soren: input please, since you're around: | 22:15 |
| *** jakedahn has joined #openstack | 22:15 | |
| *** astrolin has quit IRC | 22:15 | |
| *** greenisus has joined #openstack | 22:15 | |
| gundlach | the complaint is because i'm using Binary type on the internal_id column, but also a unique constraint. | 22:15 |
| *** jakedahn has joined #openstack | 22:16 | |
| *** jbryce has quit IRC | 22:16 | |
| greenisus | does anyone know how to hit the swift auth API when running nova? | 22:16 |
| greenisus | trying to GET /v1.0, but i just get a 404 | 22:16 |
| gundlach | I could 1) remove the unique constraint, hoping that no two instances land the same random 64 bit ID, or 2) change to a 16-byte character in which to store the str() representation of the internal_id to simply serialize it. | 22:16 |
| soren | gundlach: 3) split the 64 bit int into two 32 bit ints? | 22:17 |
| gundlach | this is temporary code in a sense, because we're going to lose it in Bexar. I'm thus leaning toward 1). | 22:17 |
| soren | Not sure that's what I'd do, but it's an option. | 22:17 |
| gundlach | soren: yeah, that's another way to serialize the value, essentially. | 22:17 |
| *** dendrobates is now known as dendro-afk | 22:18 | |
| soren | gundlach: Remind me again why this is so important for Austin? | 22:18 |
| gundlach | soren: because you pointed out the statistics inherent in using a 32 bit integer -- a few thousand instances and we probably get a collision. | 22:18 |
| soren | gundlach: Do we really expect anyone to have run more than 4 billion VM's within the next 3 months? | 22:18 |
| *** dendro-afk is now known as dendrobates | 22:18 | |
| alekibango | btw why there is not wysiwyg editor on wiki pages? can someone please help there to make it happen? | 22:18 |
| gundlach | soren: nope, but more than a few thousand, perhaps | 22:18 |
| gundlach | this isn't autoinc, it's random | 22:19 |
| soren | gundlach: Sure. 64 bit doesn't solve that :) | 22:19 |
| soren | It just makes it less likely. Why don't we just detect the collision, generate a new id and move on. | 22:19 |
| soren | ? | 22:19 |
| soren | The chance of two collisions in a row is miniscule. | 22:20 |
| burris | you'd need like 65536 machines to have a better than even chance of a collision with an evenly distributed 32-bit hash | 22:20 |
| burris | it's approx the square root | 22:20 |
| gundlach | soren: yep, that can work too | 22:20 |
| gundlach | soren: it feels better to me to use something with a low chance of collision, rather than keep retrying until we get a free one (and have to hit the DB again each time we fail). | 22:21 |
| gundlach | soren: but it'll at least get rid of the current sqlalchemy issues :) | 22:21 |
| *** gaveen has quit IRC | 22:21 | |
| soren | gundlach: the way I see it, we either need to come up with a way to *ensure* we don't get collisions, or find a nice way to deal with them when they do happen. | 22:22 |
| soren | gundlach: I don't think there's a useful middle ground. | 22:22 |
| gundlach | ok, thanks for the input. i'll take a look at auto-retrying. | 22:22 |
| soren | 64 bit namespace doesn't ensure anything. Don't trust your PRNG :) | 22:22 |
| *** pvo_ has joined #openstack | 22:23 | |
| *** ChanServ sets mode: +v pvo_ | 22:23 | |
| *** rlucio has joined #openstack | 22:24 | |
| rlucio | vishy: you around? | 22:24 |
| jonwood | Can someone have a quick look at http://wiki.openstack.org/NovaApiFlags and http://wiki.openstack.org/NovaComputeFlags to make sure I've not got it completely wrong please? | 22:25 |
| *** jakedahn has quit IRC | 22:25 | |
| gundlach | soren: oh bloody heck. take a look at nova/db/sqlalchemy/api.py line 539. it looks like this check already existed, so 32 bit ints will suffice with no change. do you agree? | 22:26 |
| * soren looks | 22:26 | |
| soren | Erk. | 22:27 |
| soren | Well, yes. | 22:27 |
| *** jxta has joined #openstack | 22:27 | |
| soren | That's a horrible approach, though, but yes, you're right. | 22:27 |
| *** pvo has quit IRC | 22:27 | |
| *** pvo_ has quit IRC | 22:28 | |
| soren | That always costs a db roundtrip (and is racy!). Collission detection (as provided by unique constraints) would avoid the race and save a db roundtrip. | 22:29 |
| *** jonwood has quit IRC | 22:31 | |
| *** miclorb has quit IRC | 22:32 | |
| alekibango | jonwood: thanks much for helping. if its wrong i cant say | 22:32 |
| alekibango | i need to dig deeper to get the knowledge and confidence | 22:32 |
| gundlach | soren: https://code.launchpad.net/~gundlach/nova/32-bit-internal-ids/+merge/38287 if you don't mind... | 22:32 |
| *** ecarlin has quit IRC | 22:32 | |
| gundlach | on the plus side, at least i've wasted a good amount of time today... | 22:32 |
| soren | gundlach: Lucky. | 22:33 |
| *** gaveen has joined #openstack | 22:33 | |
| * soren twiddles thumbs, waiting for the diff to turn up | 22:33 | |
| alekibango | oops he left | 22:34 |
| gundlach | there it is | 22:34 |
| * soren makes the sound of a rubber stamp | 22:35 | |
| *** guynaor has joined #openstack | 22:35 | |
| gundlach | bureaucracy at work! | 22:35 |
| alekibango | is there a bot which can do \later tell someone something? | 22:36 |
| aimon | hi all | 22:36 |
| alekibango | aimon: welcome | 22:36 |
| aimon | just lrking for now | 22:36 |
| aimon | *lurking | 22:36 |
| aimon | but was reading the thread | 22:36 |
| aimon | characters = '01234567890abcdefghijklmnopqrstuvwxyz' | 22:36 |
| gundlach | aimon: you noticed that too huh? | 22:36 |
| aimon | why not include Cap letters and increase your randomness | 22:36 |
| gundlach | :) | 22:36 |
| aimon | then no worries about 64 bit length | 22:37 |
| gundlach | aimon: we're actually talking about an integer field, not a base-36 character field. | 22:37 |
| aimon | easy solution to decrease likelyhood of collisoipns on unique ids | 22:37 |
| aimon | ahh | 22:37 |
| aimon | I was gathering random uid collisions from that code chunk in the diff | 22:37 |
| alekibango | just use uuidgen :) | 22:39 |
| * soren heads bedwards | 22:39 | |
| gundlach | soren: thanks for feedback and good night | 22:39 |
| *** DubLo7 has quit IRC | 22:39 | |
| aimon | alekibango.. what is that.. a semi random generator that avoids repeats? | 22:39 |
| *** drizzle has joined #openstack | 22:40 | |
| gundlach | aimon: ahh. that code chunk is in an "else" used for all uids *except* those that we're worried about (the integer kind). | 22:40 |
| alekibango | aimon: just joking but yes | 22:40 |
| aimon | :) | 22:40 |
| aimon | gundlach.. ah ok.. | 22:40 |
| aimon | I was in right mind but with wrong code pience ;) | 22:40 |
| *** kevnfx has quit IRC | 22:42 | |
| gundlach | kevnfx: sorry for the trouble; pull trunk and hopefully you'll now be all set. | 22:42 |
| *** adjohn has quit IRC | 22:44 | |
| *** alekibango has quit IRC | 22:44 | |
| aimon | gundlach thx for taking a moment to explain. I am trying to cath up as much as I can to current state of things | 22:46 |
| aimon | I am excited about recent progress | 22:46 |
| *** burris has quit IRC | 22:47 | |
| *** Cybodog has quit IRC | 22:50 | |
| *** littleidea has quit IRC | 22:50 | |
| *** Cybodog has joined #openstack | 22:52 | |
| *** aliguori has quit IRC | 22:58 | |
| *** Cybodog has quit IRC | 23:02 | |
| drizzle | is the openstack meeting tonight in person or in irc? | 23:03 |
| uvirtbot | New bug: #659571 in nova "nova-volume support for physical volumes was removed" [Undecided,New] https://launchpad.net/bugs/659571 | 23:06 |
| *** pvo has joined #openstack | 23:07 | |
| *** pvo has joined #openstack | 23:07 | |
| *** ChanServ sets mode: +v pvo | 23:07 | |
| *** johnpur has quit IRC | 23:09 | |
| *** pharkmillups has quit IRC | 23:09 | |
| *** greenisus has quit IRC | 23:13 | |
| *** pvo has quit IRC | 23:13 | |
| *** pvo has joined #openstack | 23:13 | |
| *** ChanServ sets mode: +v pvo | 23:13 | |
| *** aimon_ has joined #openstack | 23:14 | |
| *** aimon_ has quit IRC | 23:15 | |
| *** burris has joined #openstack | 23:15 | |
| *** aimon_ has joined #openstack | 23:16 | |
| *** aimon has quit IRC | 23:17 | |
| *** aimon_ is now known as aimon | 23:17 | |
| *** neogenix has quit IRC | 23:17 | |
| *** johnbergoon has quit IRC | 23:25 | |
| *** johnbergoon has joined #openstack | 23:25 | |
| *** gundlach has quit IRC | 23:26 | |
| *** jxta has quit IRC | 23:40 | |
| *** pvo has quit IRC | 23:40 | |
| *** dizz has joined #openstack | 23:44 | |
| *** dizz|away has quit IRC | 23:45 | |
| *** rlucio has quit IRC | 23:45 | |
| *** DubLo7 has joined #openstack | 23:45 | |
| *** dizz has quit IRC | 23:48 | |
| *** jakedahn has joined #openstack | 23:50 | |
| *** jakedahn has joined #openstack | 23:51 | |
| *** Ryan_Lane has quit IRC | 23:52 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!