| *** zenmatt has quit IRC | 00:01 | |
| *** adjohn has joined #openstack | 00:13 | |
| *** sophiap has joined #openstack | 00:19 | |
| *** sophiap has quit IRC | 00:25 | |
| *** RobertLJ has joined #openstack | 00:27 | |
| *** jtdowney has joined #openstack | 00:49 | |
| *** codejunkie has quit IRC | 00:53 | |
| *** zenmatt has joined #openstack | 00:59 | |
| *** zenmatt has quit IRC | 01:17 | |
| *** zenmatt has joined #openstack | 01:27 | |
| *** abecc has quit IRC | 01:30 | |
| *** zenmatt has quit IRC | 01:40 | |
| *** sirp1 has quit IRC | 01:51 | |
| *** mrayzenoss has joined #openstack | 02:03 | |
| *** mrayzenoss is now known as mray | 02:03 | |
| *** mray has joined #openstack | 02:03 | |
| *** mray has quit IRC | 02:16 | |
| *** mrayzenoss has joined #openstack | 02:17 | |
| *** mrayzenoss is now known as mray | 02:17 | |
| *** mray has joined #openstack | 02:17 | |
| *** zenmatt has joined #openstack | 02:39 | |
| *** mray has quit IRC | 02:42 | |
| *** metoikos has quit IRC | 02:43 | |
| *** codejunkie has joined #openstack | 02:59 | |
| *** silassewell has quit IRC | 03:02 | |
| *** RobertLJ has quit IRC | 03:10 | |
| *** jtdowney has quit IRC | 03:14 | |
| *** dmd17 has joined #openstack | 03:27 | |
| *** sirp1 has joined #openstack | 03:41 | |
| *** dmd17 has quit IRC | 04:29 | |
| *** adjohn has quit IRC | 04:30 | |
| *** BartVB has joined #openstack | 04:31 | |
| *** dmd17 has joined #openstack | 04:33 | |
| *** adjohn has joined #openstack | 04:33 | |
| *** jsgotangco has joined #openstack | 04:35 | |
| *** veszig has quit IRC | 04:44 | |
| *** jsgotangco has quit IRC | 04:49 | |
| *** jamiew has joined #openstack | 04:50 | |
| *** veszig has joined #openstack | 04:56 | |
| *** rdw__ has quit IRC | 05:04 | |
| *** sophiap has joined #openstack | 05:34 | |
| *** KnuckleSangwich has quit IRC | 05:41 | |
| *** joey_ has joined #openstack | 05:45 | |
| *** rajijoom has joined #openstack | 05:51 | |
| *** joey_ has quit IRC | 05:54 | |
| *** emt has joined #openstack | 06:05 | |
| emt | \quit | 06:05 |
|---|---|---|
| emt | \quit | 06:05 |
| msparks | other way | 06:06 |
| *** emt has quit IRC | 06:06 | |
| *** rajijoom has quit IRC | 06:07 | |
| *** metoikos has joined #openstack | 06:13 | |
| *** rajijoom has joined #openstack | 06:24 | |
| *** jdmaturen has quit IRC | 06:36 | |
| *** allsystemsarego has joined #openstack | 06:39 | |
| *** kashyapc has joined #openstack | 06:45 | |
| *** bytee has left #openstack | 06:45 | |
| *** adjohn has quit IRC | 06:55 | |
| *** dmd17 has quit IRC | 06:58 | |
| *** fortune has joined #openstack | 07:07 | |
| *** fortune has left #openstack | 07:07 | |
| *** joshuamckenty has joined #openstack | 07:09 | |
| *** kashyapc has quit IRC | 07:10 | |
| *** rajijoom has quit IRC | 07:12 | |
| *** rajijoom has joined #openstack | 07:16 | |
| *** stewart has quit IRC | 07:18 | |
| *** stewart has joined #openstack | 07:19 | |
| *** rajijoom has quit IRC | 07:21 | |
| *** stewart has quit IRC | 07:24 | |
| *** calavera has joined #openstack | 07:30 | |
| *** rajijoom has joined #openstack | 07:36 | |
| *** stewart has joined #openstack | 07:36 | |
| *** joshuamckenty has quit IRC | 07:55 | |
| *** joshuamckenty has joined #openstack | 07:55 | |
| *** jmckenty has joined #openstack | 08:01 | |
| *** joshuamckenty has quit IRC | 08:02 | |
| *** jmckenty is now known as joshuamckenty | 08:02 | |
| *** miclorb_ has quit IRC | 08:15 | |
| *** miclorb has joined #openstack | 08:15 | |
| *** miclorb has quit IRC | 08:20 | |
| *** sophiap has quit IRC | 08:29 | |
| *** miclorb_ has joined #openstack | 08:40 | |
| *** kashyapc has joined #openstack | 09:27 | |
| *** fortune has joined #openstack | 10:16 | |
| *** CommunityString has joined #openstack | 10:17 | |
| *** CommunityString has left #openstack | 10:17 | |
| *** jonesy_ has quit IRC | 10:36 | |
| *** joshuamckenty has quit IRC | 10:47 | |
| *** miclorb_ has quit IRC | 10:51 | |
| *** miclorb has joined #openstack | 10:51 | |
| *** joshuamckenty has joined #openstack | 10:53 | |
| *** rajijoom has quit IRC | 10:54 | |
| *** miclorb has quit IRC | 10:56 | |
| *** joshuamckenty has quit IRC | 10:56 | |
| *** joshuamckenty has joined #openstack | 10:59 | |
| *** joshuamckenty has quit IRC | 11:01 | |
| *** joshuamckenty has joined #openstack | 11:01 | |
| chmouel | I wonder if this patch on 2.6.35 http://is.gd/dY7AQ will make things different W.R.T flushing and sqlite for swift | 11:20 |
| *** joshuamckenty has quit IRC | 11:21 | |
| *** ctennis has quit IRC | 11:26 | |
| *** jonesy_ has joined #openstack | 11:33 | |
| *** carolinad28 has joined #openstack | 11:37 | |
| *** [1]omar has joined #openstack | 11:37 | |
| *** Garo_ has quit IRC | 11:38 | |
| *** omar has quit IRC | 11:38 | |
| *** viewlogic has quit IRC | 11:38 | |
| *** imeyer has quit IRC | 11:38 | |
| *** glange has quit IRC | 11:38 | |
| *** termie has quit IRC | 11:38 | |
| *** [1]omar is now known as omar | 11:38 | |
| *** glange has joined #openstack | 11:40 | |
| *** imeyer has joined #openstack | 11:40 | |
| *** viewlogic has joined #openstack | 11:40 | |
| soren | chmouel: That's just a link to the main linux-2.6 repo gitweb :) | 11:42 |
| *** termie has joined #openstack | 11:42 | |
| *** gundlach has joined #openstack | 11:46 | |
| *** jmckenty has joined #openstack | 11:46 | |
| sureshgv | jmckenty: hey | 11:46 |
| jmckenty | hi | 11:47 |
| sureshgv | morning1 | 11:47 |
| sureshgv | morning!* | 11:47 |
| jmckenty | just trying to get a bip proxy set up so I can follow what's happening while I'm in Europe | 11:47 |
| sureshgv | ohh | 11:47 |
| *** Garo_ has joined #openstack | 11:47 | |
| sureshgv | what are u checking? | 11:48 |
| jmckenty | nothing at the moment, I had to get bip installed and running | 11:49 |
| jmckenty | it's good now | 11:49 |
| *** jmckenty has quit IRC | 11:49 | |
| *** joshuamckenty has joined #openstack | 11:50 | |
| joshuamckenty | ah, excellent | 11:50 |
| joshuamckenty | I have bip, and scrollback | 11:50 |
| sureshgv | he he | 11:51 |
| sureshgv | :) | 11:51 |
| sureshgv | cool | 11:51 |
| * soren got fed up with bip at some point. | 11:54 | |
| *** joshuamckenty has quit IRC | 11:54 | |
| soren | mtaylor: Did you get a chance to look at starting a redis-server for tests and all that? | 11:55 |
| *** joshuamckenty has joined #openstack | 11:55 | |
| termie | a bird just flew in my window and stole one of my cherries from a bowl | 11:56 |
| termie | and then freaked out, smashed into the closed part of the window on the way out and smashed the cherry | 11:57 |
| termie | and then made it out the other part of the window | 11:58 |
| soren | pics or it didn't happen. | 11:58 |
| soren | :p | 11:58 |
| soren | What kind of bird? A big scary one? | 11:58 |
| soren | Aw. | 11:58 |
| *** joshuamckenty has joined #openstack | 12:02 | |
| chmouel | oops i mean http://bit.ly/bdzutw | 12:02 |
| joshuamckenty | anyone have opinions or feedback on http://wiki.openstack.org/RemnantServers | 12:03 |
| joshuamckenty | hmm, I guess it's quite early in the US, still | 12:03 |
| soren | joshuamckenty: I like the idea. | 12:06 |
| *** justinsheehy has quit IRC | 12:06 | |
| soren | joshuamckenty: Other than that, there's not much in there yet to have strong opinions on :) | 12:06 |
| redbo | chmouel: you want us to run swift's servers in laptop mode? :) | 12:07 |
| *** silassewell has joined #openstack | 12:07 | |
| chmouel | yeah we probably don't want to have the full laptop-mode features enabled :) | 12:08 |
| soren | joshuamckenty: I puts some restrictions on our discussions about some other stuff.. For instance, some of the talks we had about scheduling didn't necessarily require a separate scheduler. | 12:08 |
| joshuamckenty | ah, k | 12:09 |
| soren | joshuamckenty: so scheduling would just be defined in terms of an algorithm and/or an interface. | 12:09 |
| *** ctennis has joined #openstack | 12:10 | |
| *** ctennis has joined #openstack | 12:10 | |
| soren | joshuamckenty: E.g. a request would come in, and the scheduling mechanism might just be "broadcast it to all nodes and whichever one bothers to respond, wins". | 12:10 |
| soren | joshuamckenty: That's a contrived example, of course :) | 12:10 |
| joshuamckenty | sure, but feasible | 12:10 |
| soren | joshuamckenty: To some scale, certainly. | 12:10 |
| *** justinsheehy has joined #openstack | 12:12 | |
| *** joshuamckenty has quit IRC | 12:12 | |
| soren | joshuamckenty: Were you expecting that.. | 12:12 |
| soren | oh. | 12:12 |
| *** joshuamckenty has joined #openstack | 12:13 | |
| soren | joshuamckenty: Were you expecting that attempting to launch these instances would be rejected if there weren't enough capacity or would you expect it to be queued. | 12:13 |
| soren | ? | 12:13 |
| joshuamckenty | for Remnants? | 12:13 |
| soren | Yes. | 12:13 |
| joshuamckenty | I think of them as a permanent queue | 12:13 |
| joshuamckenty | e.g., they'll take up capacity whenever it's available | 12:14 |
| *** codejunkie has quit IRC | 12:14 | |
| joshuamckenty | and when the instances get terminated (to run dedicated instances), the request will continue to sit around | 12:14 |
| soren | Oh, so they'll get restarted once there's capacity again? | 12:14 |
| soren | Interesting. | 12:14 |
| joshuamckenty | right | 12:14 |
| *** jsgotangco has joined #openstack | 12:14 | |
| joshuamckenty | it fits the NASA use case really well, since they always have big jobs that need as much horsepower as possible, but aren't urgent | 12:15 |
| soren | I understand it might work for NASA, but for Rackspace, I'm less sure. | 12:15 |
| soren | Unless, of course, there's some sort of rotation in the scheduling. | 12:15 |
| soren | Yeah, that would probably solve things. | 12:16 |
| soren | I'm just thinking the queue very quickly would grow to a point where new request just never would get serviced. For AWS this is simpler, because their whole instance paradigm defines and ensures rotation. | 12:18 |
| soren | Hmm.. I guess instances could even get killed to service other requests from this queue (rather than only rotate when "dedicated instances" are to be run). | 12:19 |
| * soren is just thinking out loud | 12:20 | |
| *** jsgotangco has quit IRC | 12:21 | |
| *** tonywolf has joined #openstack | 12:23 | |
| *** Podilarius has left #openstack | 12:25 | |
| joshuamckenty | ah, good thoughts | 12:25 |
| joshuamckenty | sorry, am also leading a training session on TDD in Zurich | 12:25 |
| * joshuamckenty weeps that RL intrudes so brutally on happy IRC | 12:25 | |
| redbo | I don't think that's a queue anymore | 12:25 |
| joshuamckenty | we could make remnant requests expire after a few days unless renewed | 12:25 |
| notmyname | so my high-cpu, low-priority job would have to be rescheduled every few days unless I renewed the request (assuming it didn't complete yet)? | 12:27 |
| joshuamckenty | it's a concept | 12:30 |
| joshuamckenty | it would force sharing of the system | 12:30 |
| joshuamckenty | and your batch job management system could handle renewing? | 12:30 |
| redbo | Do hypervisors make it possible to create a VM that has little to no dedicated CPU time, and just gets whatever is left over? It'd be interesting to have like a hadoop VM on every machine, and it can use whatever slack is in the cluster. | 12:32 |
| notmyname | sure, but it seems (to me, at least) an extra and perhaps unnecessary requirement to put on the end user | 12:32 |
| ttx | soren: in an ISP setting, isn't fluctuating price the way you force sharing of the system ? | 12:41 |
| termie | mtaylor: any chance i can bother you to dig into my mailing list issue further? | 12:45 |
| soren | ttx: Normally, yes. | 12:53 |
| soren | ttx: However, since these are even cheaper than regular instances, I prophecise that people will schedule enough of them to saturate your capacity. | 12:54 |
| soren | I've probably got 3-4 of the smallest instances Rackspace provides running that I could probably just turn off if I'd bother using an hour to consolidate their jobs, but because they're so cheap, I don't care. | 12:54 |
| ttx | soren: how is AWS "instance paradigm" solving that ? | 12:55 |
| soren | Making them even cheaper will exacerbate the problem. | 12:55 |
| soren | ttx: When your instance dies, it's gone. | 12:55 |
| soren | ttx: So whoever is next in queue gets their turn. | 12:55 |
| ttx | ah! so it's the "restart" ability that kills it. | 12:55 |
| soren | ttx: Yeah. | 12:56 |
| soren | Maybe the API to start them could say how long you wanted them to run for. | 12:56 |
| *** pvo has joined #openstack | 12:56 | |
| *** ChanServ sets mode: +v pvo | 12:56 | |
| soren | Total. | 12:56 |
| soren | And they could get the first 10 minutes for free for bootstrapping purposes or whatnot. | 12:57 |
| *** maximilien has joined #openstack | 12:58 | |
| soren | So keeping them running indefinitely would actually require effort. | 12:58 |
| soren | Minimal effort, but still effort. | 12:58 |
| ttx | soren: I'm sure you can plug some economics that would solve it, at least in the ISP scenario. That doesn't solve the "fairness" in a lab setting though. | 12:59 |
| soren | ttx: An auction of sorts, perhaps? | 13:00 |
| soren | "I want 4 cores for 100 hours, and I'm willing to pay $7 for it." | 13:01 |
| soren | Only when there aren't anyone else around willing to pay more, you get your stuff scheduled. | 13:01 |
| ttx | that's kinda how spot instances work | 13:02 |
| soren | Ah :) | 13:02 |
| soren | Go figure. | 13:02 |
| ttx | (I think) :) | 13:02 |
| ttx | but tha's an interesting problem. | 13:03 |
| ttx | I wonder if there is a "one model fits all" solution, or if it should be pluggable. | 13:03 |
| soren | Very likely. | 13:04 |
| soren | (The latter, that is) | 13:04 |
| *** rajijoom has joined #openstack | 13:10 | |
| soren | What's interesting here, I think, is that this data does not seem to have a natural home other than a global data store. | 13:12 |
| *** kashyapc has quit IRC | 13:14 | |
| *** jsmith-away is now known as jsmith | 13:14 | |
| soren | ..something I think I may have hypothecised at some point wouldn't happen :) | 13:17 |
| *** abecc has joined #openstack | 13:19 | |
| _0x44 | soren: Before Destro joined RS, Jared and I were doing a lot of research/planning for the new cloud-servers arch, and kept coming back to needing one. Everytime we'd get rid of it, one of us would say something like that and end up needing it again. | 13:19 |
| pvo | global datastore or regional? | 13:22 |
| pvo | global to a group of clusters? or global as in the dictionary sense? | 13:23 |
| _0x44 | In Jared's and my design, it was eventually consistent and global across regions. | 13:24 |
| pvo | yea, I've been having that discussion as well.... though we can't seem to find the need to have the state of one instance needed in another region | 13:24 |
| pvo | other than reporting or whatnot. | 13:25 |
| soren | pvo: we're talking about http://wiki.openstack.org/RemnantServers | 13:25 |
| soren | pvo: These are instance futures, essentially. | 13:25 |
| soren | pvo: they don't live anywhere, /yet/. | 13:25 |
| pvo | right, I've talked with jesse about them | 13:25 |
| pvo | I know what they are and why they want them | 13:25 |
| pvo | but we're talking about location of data across regions and finding out state about those, right? | 13:26 |
| _0x44 | Is there a reason you wouldn't want instance futures against multiple regions? | 13:26 |
| _0x44 | The "global datastore" in this case, could just be a queue of instance futures from which each region could pull | 13:27 |
| *** gundlach has quit IRC | 13:29 | |
| pvo | different regions will have different costs, right? | 13:30 |
| pvo | lon power is going to be higher than power in say, Tucson | 13:30 |
| _0x44 | Sure, but it's not in Texas v. Chicago | 13:32 |
| pvo | than is RS specific. | 13:32 |
| _0x44 | And the way to prevent that from happening is to have the remnant request on our side specify a maximum price the customer is willing to pay. | 13:32 |
| soren | Different regions /can/ have different costs, yes. | 13:32 |
| soren | They might happen to be the same, but needn't be. | 13:33 |
| _0x44 | If a region's remnant costs exceed that, it could ignore the request. | 13:33 |
| _0x44 | And the cost of a remnant would probably increase on our side as a region/cluster filled, because there would be less available space for remnants. | 13:34 |
| mcgurrin | this looks kind of lie the AWS model which I think is actually auto biding based | 13:40 |
| *** anm_ has joined #openstack | 13:41 | |
| *** Phil_FL has joined #openstack | 13:41 | |
| mcgurrin | that might be the way to go, what additional cost is there to have an extra VM running in spare capacity? I think there is none, just have people bid for extra space like with ebay autobidding with some very small minimum price | 13:41 |
| *** gundlach has joined #openstack | 13:44 | |
| *** Podilarius has joined #openstack | 13:53 | |
| *** silassewell has quit IRC | 13:57 | |
| *** sophiap has joined #openstack | 13:58 | |
| *** pvo has quit IRC | 13:58 | |
| *** pcrews has joined #openstack | 13:59 | |
| *** pvo has joined #openstack | 14:02 | |
| *** pvo has joined #openstack | 14:02 | |
| *** ChanServ sets mode: +v pvo | 14:02 | |
| *** jsmith is now known as jsmith-away | 14:03 | |
| *** jsmith-away is now known as jsmith | 14:06 | |
| *** aliguori has joined #openstack | 14:09 | |
| *** justinsheehy has quit IRC | 14:11 | |
| *** jdarcy has joined #openstack | 14:21 | |
| *** pvo has quit IRC | 14:25 | |
| *** jaypipes has joined #openstack | 14:25 | |
| *** justinsheehy has joined #openstack | 14:27 | |
| *** mrayzenoss has joined #openstack | 14:30 | |
| *** mrayzenoss is now known as mray | 14:30 | |
| *** mray has joined #openstack | 14:30 | |
| joshuamckenty | I like the term instance futures | 14:31 |
| joshuamckenty | I think it's better than "Remnants" | 14:31 |
| *** armmer has quit IRC | 14:33 | |
| thom | can there be an instance futures warehouse? (http://wiki.lspace.org/wiki/Pork_Futures_Warehouse) | 14:33 |
| *** armmer has joined #openstack | 14:35 | |
| *** pvo has joined #openstack | 14:36 | |
| *** ChanServ sets mode: +v pvo | 14:36 | |
| *** jbryce has joined #openstack | 14:40 | |
| *** sirp1 has quit IRC | 14:41 | |
| *** dendrobates is now known as dendro-afk | 14:41 | |
| *** tonywolf has quit IRC | 14:42 | |
| *** mray has quit IRC | 14:43 | |
| *** sirp1 has joined #openstack | 14:43 | |
| *** matteo74 has quit IRC | 14:44 | |
| *** gnrfan has joined #openstack | 14:50 | |
| *** pcrews is now known as pcrews_errands | 14:53 | |
| *** matteo74 has joined #openstack | 14:56 | |
| *** codejunkie has joined #openstack | 14:58 | |
| *** armmer has quit IRC | 15:01 | |
| *** armmer has joined #openstack | 15:02 | |
| *** mrayzenoss has joined #openstack | 15:03 | |
| *** mrayzenoss is now known as mray | 15:03 | |
| *** tr3buchet has joined #openstack | 15:06 | |
| *** mray is now known as mrayzenoss | 15:10 | |
| *** mrayzenoss has joined #openstack | 15:10 | |
| *** mrayzenoss is now known as mray | 15:10 | |
| *** justinsheehy has quit IRC | 15:23 | |
| *** jamiew has quit IRC | 15:28 | |
| *** mcgurrin has quit IRC | 15:29 | |
| *** mcgurrin has joined #openstack | 15:30 | |
| *** mcgurrin has quit IRC | 15:30 | |
| *** mcgurrin has joined #openstack | 15:31 | |
| *** mcgurrin has left #openstack | 15:31 | |
| *** justinsheehy has joined #openstack | 15:36 | |
| *** calavera has quit IRC | 15:46 | |
| *** mrayzenoss has joined #openstack | 15:46 | |
| *** mrayzenoss has joined #openstack | 15:46 | |
| *** mray has quit IRC | 15:49 | |
| *** mrayzenoss is now known as mray | 15:49 | |
| *** rajijoom has quit IRC | 15:50 | |
| *** justinsheehy has quit IRC | 15:51 | |
| *** pcrews_errands is now known as pcrews | 15:52 | |
| *** jbryce has quit IRC | 15:58 | |
| *** jbryce has joined #openstack | 16:01 | |
| *** dendro-afk is now known as dendrobates | 16:05 | |
| *** kashyapc has joined #openstack | 16:05 | |
| *** jbarratt has joined #openstack | 16:16 | |
| *** jdmaturen has joined #openstack | 16:18 | |
| *** jdmaturen has quit IRC | 16:19 | |
| *** jdmaturen has joined #openstack | 16:19 | |
| *** sophiap has quit IRC | 16:25 | |
| *** jsmith is now known as jsmith-away | 16:30 | |
| *** kashyapc has quit IRC | 16:33 | |
| *** cwilson has joined #openstack | 16:33 | |
| *** justinsheehy has joined #openstack | 16:34 | |
| mtaylor | termie: yes. there is a ticket open on it. lemme check what the latest status on it is | 16:38 |
| mtaylor | soren: I have done nothing with start a redis server for tests | 16:39 |
| *** dendrobates is now known as dendro-afk | 16:43 | |
| alekibango | hello. i am reading about nova and i am curious, is there some effort to create web interface for the public api servers? | 16:43 |
| alekibango | and 2) what are guest agents good for? do i need to install something on guest? | 16:44 |
| _0x44 | alekibango: We use the guest agent for live configuration of the VM, they're not required in nova | 16:45 |
| alekibango | aha, so its possible to manage guests somehow ? | 16:46 |
| _0x44 | In nova's current state, guest VM configuration only happens on instantiation. I think we're trying to get live configuration via the agents by our first release, and then non-live reconfiguration via guestfs/whatever later. | 16:49 |
| alekibango | ok thanks for explanation | 16:51 |
| *** jsmith-away is now known as jsmith | 16:51 | |
| _0x44 | Sure, no problem. :) | 16:52 |
| alekibango | there are some projects trying to do configuration management, but you prolly know them already -- for example cfengine | 16:54 |
| alekibango | so ... is there some effort to create web interface for the public api servers? | 16:55 |
| _0x44 | I think greenisus is working on a web interface | 16:55 |
| alekibango | hmm i would like to help | 16:55 |
| _0x44 | I know he's working on an ipad/iphone/android app | 16:55 |
| alekibango | i would love web management for public clouds... | 16:56 |
| _0x44 | He's not on right now, but send him an email. | 16:56 |
| alekibango | thanks, i will be online almost 24/7 | 16:56 |
| _0x44 | I just privmsged his email | 16:56 |
| alekibango | ok, ty :) | 16:57 |
| alekibango | when do you think nova will be stable enough to start a public service build upon it? :) | 16:58 |
| alekibango | built* | 16:58 |
| alekibango | september? | 16:58 |
| gundlach | vish1: I followed the instructions at http://wiki.openstack.org/InstallationNova20100729 on Ubuntu and all went well except that nova-compute required libvirt (unless I said --connection-type=fake). should i add installation of libvirt to the instructions? | 16:59 |
| gundlach | or should i remove the note saying that --connection-type=fake is for OSX only? | 16:59 |
| _0x44 | gundlach: if so, you need to compile libvirt from source and patch it not to crash if it can't get dmi | 17:00 |
| _0x44 | alekibango: Our first release will be in October | 17:00 |
| _0x44 | alekibango: I don't know what the project manager is saying about it's stability at that point. | 17:00 |
| _0x44 | alekibango: But swift is in production at Rackspace now | 17:01 |
| alekibango | ic | 17:01 |
| vish1 | gundlach: there are a few additions needed for actually running the system in non-fake mode | 17:03 |
| vish1 | libvirt is one | 17:03 |
| mtaylor | termie: https://answers.edge.launchpad.net/launchpad/+question/119549 | 17:03 |
| vish1 | although the current lucid version is kind of old | 17:04 |
| gundlach | vish1: great -- then it sounds like i should either add libvirt to the instructions, or remove this comment: | 17:04 |
| gundlach | tools/with_venv.sh bin/nova-compute --verbose --nodaemon --connection_type=fake #fake is for OSX only | 17:04 |
| vish1 | good point maybe remove the comment for now | 17:04 |
| gundlach | kthx | 17:04 |
| vish1 | we need to add how to install the rest of the dependencies for ubuntu | 17:04 |
| gundlach | the dependencies req'd to run in non-fake mode, you mean? | 17:05 |
| vish1 | yes | 17:05 |
| *** joshuamckenty has quit IRC | 17:05 | |
| vish1 | to actually run a system | 17:05 |
| vish1 | there are other steps like getting a working image | 17:05 |
| vish1 | setting up metadata url | 17:05 |
| vish1 | etc | 17:05 |
| gundlach | those are config steps as opposed to installation, right? so hudson wouldn't need to do those in order to verify our install process works and tests pass? | 17:06 |
| vish1 | correct... | 17:06 |
| vish1 | unit tests pass before the last steps | 17:07 |
| vish1 | integration testing/smoketests will require configuration and such | 17:07 |
| gundlach | oh yeah, 1 unit test did fail when i did the install, i forgot to mention that. | 17:07 |
| *** kallistec has quit IRC | 17:07 | |
| vish1 | which test? | 17:07 |
| gundlach | [FAIL]: nova.tests.process_unittest.ProcessTestCase.test_max_processes | 17:08 |
| gundlach | Traceback (most recent call last): | 17:08 |
| gundlach | Failure: twisted.trial.unittest.FailTest: [Failure instance: Traceback: <class 'twisted.trial.unittest.FailTest'>: not equal: | 17:08 |
| gundlach | a = 'd4' | 17:08 |
| gundlach | b = 'd3' | 17:08 |
| gundlach | haven't dug in to see what it means yet -- perhaps you know | 17:08 |
| *** silassewell has joined #openstack | 17:09 | |
| gundlach | termie noted a possible race condition in tests... i'll rerun | 17:10 |
| gundlach | yep, passed that time, so i assume it was indeed a race. | 17:10 |
| vish1 | yes it is | 17:14 |
| vish1 | be nice to have that test not randomly fail :) | 17:14 |
| comstud | vish, i'm looking at the scheduler section on the etherpad... | 17:17 |
| comstud | and i have a stupid question... what is a cast vs a call? | 17:18 |
| _cerberus_ | comstud: cast is an asynchronous request, call is synchronous | 17:19 |
| comstud | ok, that's what I thought | 17:19 |
| comstud | just checking | 17:19 |
| comstud | thnx | 17:19 |
| _cerberus_ | comstud: np | 17:20 |
| *** jamiew has joined #openstack | 17:20 | |
| *** davidX- has joined #openstack | 17:24 | |
| *** anotherjesse has joined #openstack | 17:26 | |
| *** kallistec has joined #openstack | 17:35 | |
| *** pvo has quit IRC | 17:38 | |
| *** jonesy_ has quit IRC | 17:53 | |
| *** maximilien has quit IRC | 17:55 | |
| *** carolinad28 has quit IRC | 18:00 | |
| *** jonesy_ has joined #openstack | 18:02 | |
| *** carolinad28 has joined #openstack | 18:03 | |
| *** anm_ has quit IRC | 18:15 | |
| *** binaryWarrior has joined #openstack | 18:19 | |
| *** joshuamckenty has joined #openstack | 18:28 | |
| *** andreh has joined #openstack | 18:31 | |
| *** andreh has quit IRC | 18:34 | |
| *** carolinad28 has quit IRC | 18:41 | |
| *** pvo has joined #openstack | 18:45 | |
| *** pvo has joined #openstack | 18:45 | |
| *** ChanServ sets mode: +v pvo | 18:45 | |
| *** carolinad28 has joined #openstack | 18:46 | |
| *** silassewell has quit IRC | 18:46 | |
| *** cwilson has left #openstack | 18:55 | |
| *** cwilson has joined #openstack | 18:56 | |
| *** rdw has joined #openstack | 19:00 | |
| *** syah has quit IRC | 19:13 | |
| *** JimCurry has joined #openstack | 19:18 | |
| *** theron has joined #openstack | 19:24 | |
| *** JimCurry has quit IRC | 19:26 | |
| *** aliguori has quit IRC | 19:43 | |
| jamiew | quiet channel :) excited about this announcement last week, I have been experimenting building a small private cloud for my company with Eucalyptus | 19:52 |
| creiht | too quiet... | 19:53 |
| creiht | :) | 19:53 |
| * jsmith stirs things up a bit | 19:54 | |
| alekibango | jamiew: its only illusion. this channel is really responsive, when something is up... | 19:57 |
| alekibango | jamiew: did you like eucalyptus? | 19:57 |
| jamiew | it was fine when I thought there were no other options ;) | 19:57 |
| alekibango | imho java is pure suckage... and imho walrus is not really scalable | 19:57 |
| jamiew | yeah I am very glad OpenStack is python | 19:58 |
| * alekibango loves python (see python love story) | 19:58 | |
| * creiht is actually quite surprised at the lack of language hate | 19:58 | |
| alekibango | creiht: most pythoners hate java as much as we hate pascal | 19:58 |
| creiht | hehe | 19:59 |
| *** aliguori has joined #openstack | 19:59 | |
| ianweller | jsmith: stop that | 19:59 |
| jsmith | ianweller: Oh yeah -- that's your job! | 19:59 |
| ianweller | damn right | 19:59 |
| davidX- | alekibango: i had a very strange experience with it | 19:59 |
| alekibango | davidX-: with eucalyptus? | 19:59 |
| davidX- | java would bomb and was told i had to reboot each node after java bombed on the "cloud controller" node (UEC) for everything to behave properly | 20:00 |
| davidX- | i stepped away after that.. | 20:00 |
| jsmith | ianweller: Speaking of which -- did you have a chance to review that package? | 20:00 |
| davidX- | i never really looked into why that is | 20:00 |
| mtaylor | oh yuck. show's how much I pay attention - didn't really realize that eucalyptus was java | 20:00 |
| ianweller | jsmith: haven't yet, hoping to tomorrow morning | 20:01 |
| * ianweller has been running around in circles all weekend | 20:01 | |
| alekibango | ianweller: use teamwork to break those circles | 20:03 |
| codejunkie | just curious, are there any SOA distributed components within open stack? if so, how are the operations kept atomic? | 20:05 |
| *** maximilien_ has joined #openstack | 20:06 | |
| redbo | that word... it burns! | 20:06 |
| creiht | haha | 20:06 |
| alekibango | davidX-: i looked also at niftyname, cloud.com and other cloud technologies. i was about to write my own. But it looks like openstack saved me from doing that... lol | 20:06 |
| jamiew | alekibango: did you look into OpenNebula at all? | 20:07 |
| jamiew | the survey of what's out there has been kind of tricky | 20:07 |
| alekibango | just one eye :D | 20:07 |
| *** maximilien_ has quit IRC | 20:08 | |
| creiht | codejunkie: I'm not SOA certified, so I'm afraid I can not answer your question :) | 20:08 |
| davidX- | alekibango: cloud.com doesn't seem too bad | 20:08 |
| alekibango | jamiew: i like the feel of extendability of openstack | 20:08 |
| alekibango | the code looks more like my own | 20:09 |
| jamiew | yeah ditto. And I like Rackspace. And NASA :) | 20:09 |
| alekibango | easier to get through | 20:09 |
| alekibango | well i dont like nasa much :) they have really strange logo... but they do nice software, thats sure... | 20:09 |
| jamiew | they got spaceships man!! spaceships!! | 20:10 |
| alekibango | i have few spaceships on my desk right now | 20:10 |
| alekibango | so what | 20:10 |
| alekibango | :D | 20:10 |
| jamiew | haha | 20:10 |
| codejunkie | me either, but i get the jist. just wondering how transactions are handled in openstack across distributed components, if there in fact distributed components. | 20:10 |
| notmyname | codejunkie: are you referring to swift (storage) or nova (compute)? | 20:11 |
| notmyname | or in general? | 20:11 |
| codejunkie | was just wondering in general | 20:11 |
| alekibango | jamiew: about cloud.com - if they would opensource all the code, i would go with them. but theirs community edition is not good enough for me | 20:12 |
| jamiew | ah yes, the ol Enterprise Edition >:| | 20:13 |
| alekibango | ... i like the web interface - i need similar for openstack | 20:13 |
| alekibango | davidX-: that line about cloud was meant for you | 20:13 |
| *** [1]omar has joined #openstack | 20:14 | |
| davidX- | well, i just figured out today is ex vmops folks | 20:14 |
| davidX- | :D | 20:14 |
| alekibango | enterprise edition is good for spaceships - not here on earth | 20:14 |
| creiht | codejunkie: I think the easiest way to answer your question (at least for swift) is that we went out of our way to not have distributed transactions | 20:15 |
| *** allsystemsarego has quit IRC | 20:15 | |
| notmyname | I think NASA even has a spaceship named "enterprise" | 20:16 |
| codejunkie | i see, thanks :) | 20:16 |
| notmyname | s/has/had | 20:16 |
| jsmith | They did -- it's now in the Smithsonian | 20:17 |
| alekibango | cloud technology is getting commoditized. smart move with openstack :D | 20:17 |
| davidX- | it's interesting and an impressive move for rackspace. | 20:18 |
| *** omar has quit IRC | 20:18 | |
| *** [1]omar is now known as omar | 20:18 | |
| alekibango | i think if they will not make bad moves, it can help a lot. | 20:18 |
| *** rokoroko has joined #openstack | 20:18 | |
| *** silassewell has joined #openstack | 20:19 | |
| alekibango | combining openstack with http://deltacloud.org/ could give web management of public clouds, right? this might be interesting | 20:20 |
| cw | notmyname: OV-101 was named Enterprise after a whole slew of trekkies had a nergasm and pushed for the name | 20:20 |
| *** rokoroko has quit IRC | 20:22 | |
| notmyname | IIRC, the first nuclear sub was named Nautilus (from Verne's 20000 Leagues Under the Sea). I think it's cool to name stuff for science fiction once that stuff becomes reality | 20:22 |
| *** rnewson has joined #openstack | 20:23 | |
| alekibango | i was reading that book when i was 8 or 10 | 20:23 |
| creiht | http://en.wikipedia.org/wiki/USS_Nautilus_(SSN-571) | 20:24 |
| alekibango | but most modern scifi works are scary when you see what can and will be done.. | 20:27 |
| *** justinsheehy has quit IRC | 20:28 | |
| alekibango | manbearpig was fiction. now spidergoat is reality | 20:28 |
| _0x44 | alekibango: Spidergoat has been a reality for a while now. | 20:29 |
| alekibango | monstanto patents genes in food (including pigs) and wants to take on the biosphere by GM creatures | 20:29 |
| alekibango | singularity will be here soon | 20:29 |
| alekibango | along with hive mentality of many people | 20:29 |
| alekibango | well no need to read scifi anymore :D | 20:30 |
| alekibango | its happening | 20:30 |
| alekibango | so fast that they cannot write enough good books about it | 20:30 |
| *** omar has quit IRC | 20:30 | |
| mray | The future is already here - it is just unevenly distributed. | 20:31 |
| mray | like Gibson said | 20:31 |
| alekibango | yes nice quote | 20:31 |
| *** miclorb_ has joined #openstack | 20:31 | |
| notmyname | I'm not sure if alekibango's scifi prophecies or codejunkie's SOAs are more scary | 20:31 |
| creiht | lol | 20:31 |
| alekibango | i am not trying to scare you :D | 20:31 |
| notmyname | nah, no worries | 20:32 |
| alekibango | what is SOA? | 20:33 |
| alekibango | aha, got it | 20:34 |
| notmyname | society of actuaries, according to google | 20:34 |
| _0x44 | You can't trust them either, they've got tables for everything. | 20:34 |
| _0x44 | Including that last statement, oh damn, they're onto me. | 20:35 |
| notmyname | but enterprise-y stuff for us. (maybe we could put it on a spaceship!) | 20:35 |
| alekibango | google censors a lot... :D | 20:35 |
| uvirtbot | New bug: #612722 in swift "stats-report doesn't do a good job of error reporting" [Undecided,New] https://launchpad.net/bugs/612722 | 20:45 |
| *** jonesy_ has quit IRC | 20:47 | |
| *** pvo_ has joined #openstack | 20:48 | |
| *** pvo_ has joined #openstack | 20:48 | |
| *** ChanServ sets mode: +v pvo_ | 20:48 | |
| *** pvo_ has quit IRC | 20:48 | |
| *** BartVB has quit IRC | 20:49 | |
| *** carolinad28 has quit IRC | 20:49 | |
| *** pvo_ has joined #openstack | 20:49 | |
| *** ChanServ sets mode: +v pvo_ | 20:49 | |
| *** pvo has quit IRC | 20:52 | |
| *** pvo_ is now known as pvo | 20:52 | |
| *** tdeath has joined #openstack | 20:53 | |
| *** justinsheehy has joined #openstack | 20:53 | |
| *** jdarcy has quit IRC | 20:57 | |
| *** pvo has quit IRC | 21:00 | |
| *** justinsheehy_ has joined #openstack | 21:00 | |
| *** justinsheehy has quit IRC | 21:01 | |
| *** justinsheehy_ is now known as justinsheehy | 21:01 | |
| *** ctennis has quit IRC | 21:03 | |
| *** carolinad28 has joined #openstack | 21:03 | |
| *** codejunkie has quit IRC | 21:12 | |
| *** justinsheehy has quit IRC | 21:15 | |
| *** carolinad28 has quit IRC | 21:15 | |
| *** ctennis has joined #openstack | 21:18 | |
| *** ctennis has joined #openstack | 21:18 | |
| *** theron has quit IRC | 21:19 | |
| *** maximilien has joined #openstack | 21:20 | |
| *** justinsheehy has joined #openstack | 21:21 | |
| *** aliguori has quit IRC | 21:24 | |
| *** aliguori has joined #openstack | 21:25 | |
| *** gustavomzw has joined #openstack | 21:26 | |
| *** justinsheehy has quit IRC | 21:26 | |
| *** maximilien has quit IRC | 21:28 | |
| *** maximilien has joined #openstack | 21:32 | |
| *** pvo has joined #openstack | 21:35 | |
| *** pvo has joined #openstack | 21:35 | |
| *** ChanServ sets mode: +v pvo | 21:35 | |
| *** jtdowney has joined #openstack | 21:39 | |
| *** cwilson has quit IRC | 21:40 | |
| *** justinsheehy has joined #openstack | 21:41 | |
| *** jc_smith has joined #openstack | 21:45 | |
| *** pvo has quit IRC | 21:50 | |
| *** gustavomzw has quit IRC | 21:57 | |
| *** miclorb_ has quit IRC | 22:04 | |
| *** Podilarius has left #openstack | 22:04 | |
| *** miclorb has joined #openstack | 22:04 | |
| jc_smith | does nova support windows2k3/2k8 guests? | 22:08 |
| *** miclorb has quit IRC | 22:09 | |
| _0x44 | jc_smith: Not currently. | 22:11 |
| anotherjesse | jc_smith: eventually - there are technical & license issues which meant we (ansolabs) didn't prioritize it yet | 22:11 |
| anotherjesse | _0x44: you worked on the windows guest agent right? | 22:11 |
| _0x44 | anotherjesse: No, I'm working on the linux agent. | 22:11 |
| _0x44 | I just presented them together. | 22:11 |
| anotherjesse | jc_smith: so rackspace has a windows agent that works with xen - which nova will support soon-ish | 22:12 |
| jc_smith | what does the agent do? gather statistics and run commands? | 22:12 |
| _0x44 | jc_smith: Just configuration | 22:13 |
| anotherjesse | it is a solution to the technical problem of how you set passwords, ... | 22:13 |
| jc_smith | some of the systems I've worked on in the past had to inject sysprep files with license codes and passwords, resize NTFS, drop in appropriate paravirt drivers, etc | 22:13 |
| _0x44 | And KMS activation | 22:13 |
| anotherjesse | unlike linux where you can shove a ssh key into .ssh/authorized_keys | 22:13 |
| jc_smith | I see, so the password set it does in the agent | 22:13 |
| jc_smith | yeah, I did that with sysprep, but an agent works | 22:14 |
| jc_smith | is AMI the only source format for VM images? | 22:14 |
| anotherjesse | there is a patch to use raw disks | 22:15 |
| anotherjesse | and AMI is a very simplistic format - it is a set of raw partition images - disk, ramdisk, kernel | 22:15 |
| jc_smith | partimage-ng has an interesting format that is file system aware, its like one step above raw | 22:15 |
| anotherjesse | looking forward to https://code.launchpad.net/~justin-fathomdb/nova/raw-disk-images landing | 22:16 |
| *** dendro-afk is now known as dendrobates | 22:18 | |
| jc_smith | does the windows agent communicate over the network or through serial port? | 22:21 |
| dendrobates | jc_smith: right now xenstore | 22:22 |
| _0x44 | jc_smith: It communicates over the xenstore | 22:22 |
| dendrobates | _0x44: haha beat you. | 22:22 |
| * _0x44 shakes a fist. | 22:22 | |
| jc_smith | hence why things need to change with kvm, ah :) | 22:23 |
| _0x44 | I suggested a contentious, though hypervisor agnostic way of having the agents communicate with the host | 22:23 |
| _0x44 | Just throw another interface on the host bridge | 22:24 |
| jc_smith | we had some crazy ideas about running ppp through serial port to the host | 22:24 |
| dendrobates | jc_smith: I am hopeful we can use vmchannel on kvm, but I;ve never used it before | 22:24 |
| *** pvo has joined #openstack | 22:25 | |
| *** ChanServ sets mode: +v pvo | 22:25 | |
| jc_smith | I wish I had some more lab machines so I could get more hands on with nova | 22:25 |
| jc_smith | I think I will soon | 22:26 |
| anotherjesse | it should work inside of virtualization if you use simple/flat networking and qemu | 22:26 |
| jc_smith | I'm reading the docs, but its hard to really retain things until you seem them in action doing useful things | 22:26 |
| jc_smith | I have it on a vm, thats true I could use qemu | 22:26 |
| anotherjesse | most of the time we don't even use real virtualization - since most of the work is about the meta-data | 22:27 |
| anotherjesse | during development that is | 22:28 |
| jc_smith | is there a way to make a new image from an existing VM, or snapshot? | 22:28 |
| anotherjesse | you can use bundling from inside the guest | 22:29 |
| jc_smith | I see | 22:30 |
| jc_smith | Thanks Jesse | 22:31 |
| *** ctennis_ has joined #openstack | 22:32 | |
| anotherjesse | there are ways to do it from outside the guest as well | 22:32 |
| *** ctennis has quit IRC | 22:34 | |
| *** ctennis_ is now known as ctennis | 22:34 | |
| *** pvo has quit IRC | 22:38 | |
| *** jamiew has quit IRC | 22:40 | |
| *** jamiew has joined #openstack | 22:41 | |
| *** abecc has quit IRC | 22:52 | |
| *** jonesy_ has joined #openstack | 22:53 | |
| *** gundlach has quit IRC | 22:54 | |
| *** mray has quit IRC | 22:55 | |
| *** jtdowney has quit IRC | 23:14 | |
| *** miclorb_ has joined #openstack | 23:14 | |
| *** binaryWarrior has quit IRC | 23:31 | |
| *** jsmith is now known as jsmith-away | 23:33 | |
| *** hornbeck has joined #openstack | 23:39 | |
| *** maximilien has quit IRC | 23:50 | |
| *** codejunkie has joined #openstack | 23:52 | |
| *** pvo has joined #openstack | 23:54 | |
| *** pvo has joined #openstack | 23:54 | |
| *** ChanServ sets mode: +v pvo | 23:54 | |
| *** maximilien has joined #openstack | 23:55 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!