| *** jfluhmann has quit IRC | 00:02 | |
| *** jfluhmann has joined #openstack | 00:14 | |
| *** joearnold has joined #openstack | 00:23 | |
| *** jfluhmann has quit IRC | 00:24 | |
| *** jfluhmann has joined #openstack | 00:36 | |
| *** joearnold has quit IRC | 00:52 | |
| *** zul has joined #openstack | 01:08 | |
| *** karmabot has quit IRC | 01:12 | |
| *** _0x44 has quit IRC | 01:13 | |
| *** pvo has quit IRC | 01:13 | |
| *** _0x44 has joined #openstack | 01:13 | |
| *** ChanServ sets mode: +v _0x44 | 01:13 | |
| *** karmabot has joined #openstack | 01:13 | |
| *** pvo has joined #openstack | 01:14 | |
| *** ChanServ sets mode: +v pvo | 01:14 | |
| *** _skrusty has quit IRC | 01:31 | |
| *** alekibango has quit IRC | 01:31 | |
| *** MikeZ has joined #openstack | 01:40 | |
| MikeZ | Hi, New user here. I've used google, and looked around the wiki. Could sombody please point me to where I can get a the ramdisk, kernel, machine image for Ubuntu LTS? I've only launched images.tgz and vmbuilder doesn't seem to work for me | 01:44 |
|---|---|---|
| *** _skrusty has joined #openstack | 01:44 | |
| MikeZ | Id really appreciate knowing where to get or create the ramdisk, kernel, and machine image for ubuntu. A URL would be great, or even the search to perform at google. | 01:56 |
| *** Firaga has joined #openstack | 01:58 | |
| Firaga | ahoy | 01:59 |
| Firaga | i need assistance regarding creating an image for openstack compute... | 01:59 |
| Firaga | anyone? please... | 01:59 |
| creiht | MikeZ, Firaga: have you looked at wiki.openstack.org? | 02:00 |
| creiht | sorry, I haven't done anything on the nova side, so can't help much | 02:00 |
| MikeZ | Yes, I can't find any information other than download images.tgz, or check out this guys repo which doesn't tell me anything | 02:00 |
| creiht | But several people should be on by tomorrow that can answer your questions | 02:01 |
| Firaga | correct... | 02:01 |
| Firaga | i see then... | 02:01 |
| creiht | Sorry I can't help you further | 02:01 |
| Firaga | yeah, hard working on holidays... | 02:01 |
| MikeZ | creiht: Thanks anyway | 02:01 |
| Firaga | anyway HAPPY NEW YEAR Guys... | 02:02 |
| creiht | hehe | 02:02 |
| creiht | indeed | 02:02 |
| MikeZ | Firaga: Happy New year to you | 02:02 |
| Firaga | i'll be back for the same question, maybe... aloha... | 02:03 |
| MikeZ | Firaga: If you figure it out, I would love to know | 02:04 |
| *** alekibango has joined #openstack | 02:10 | |
| Firaga | no problem... | 02:13 |
| Firaga | how would i let you know? | 02:13 |
| Firaga | i dont use facebook | 02:13 |
| Firaga | hehehe | 02:13 |
| MikeZ | If you don't intend to come back here, I'll send you my email | 02:20 |
| Firaga | nah no need for that, i'll be back... asta le vista baby | 02:25 |
| *** Guest26562 is now known as matiu | 02:29 | |
| *** matiu has joined #openstack | 02:29 | |
| *** greenisus has quit IRC | 02:44 | |
| *** greenisus has joined #openstack | 02:47 | |
| *** Firaga has quit IRC | 02:47 | |
| *** greenisus has quit IRC | 02:55 | |
| *** damon__ has joined #openstack | 03:00 | |
| *** sophiap has quit IRC | 03:01 | |
| *** sophiap has joined #openstack | 03:01 | |
| *** damon__ has quit IRC | 03:07 | |
| *** krish has joined #openstack | 03:16 | |
| *** damon__ has joined #openstack | 03:21 | |
| *** matiu_ has joined #openstack | 03:37 | |
| *** matiu has quit IRC | 03:41 | |
| *** matiu_ is now known as matiu | 03:41 | |
| *** krish has left #openstack | 03:55 | |
| *** sophiap has quit IRC | 04:01 | |
| *** sophiap has joined #openstack | 04:18 | |
| *** zaitcev has quit IRC | 04:20 | |
| *** jfluhmann has quit IRC | 04:29 | |
| *** sophiap has quit IRC | 04:31 | |
| *** sophiap has joined #openstack | 04:33 | |
| *** joearnold has joined #openstack | 04:48 | |
| *** kashyapc has joined #openstack | 04:50 | |
| *** joearnold has quit IRC | 05:23 | |
| *** sophiap has quit IRC | 05:26 | |
| *** hggdh has quit IRC | 05:27 | |
| *** aimon has quit IRC | 05:57 | |
| *** aimon has joined #openstack | 05:57 | |
| *** hadrian has quit IRC | 06:05 | |
| *** damon__ has quit IRC | 06:16 | |
| *** chmouel_ is now known as chmouel | 06:20 | |
| eday | exit | 06:35 |
| eday | focus fail | 06:35 |
| soren | Happy new year, #openstack. | 06:44 |
| soren | eday: 'morning. | 06:44 |
| eday | soren: evening! :) | 06:45 |
| *** ramkrsna has joined #openstack | 06:47 | |
| *** joearnold has joined #openstack | 06:50 | |
| *** guigui has joined #openstack | 06:54 | |
| *** aimon_ has joined #openstack | 07:05 | |
| *** aimon has quit IRC | 07:07 | |
| *** aimon_ is now known as aimon | 07:07 | |
| *** miclorb_ has quit IRC | 07:09 | |
| *** joearnold has quit IRC | 07:17 | |
| * ttx waves | 07:19 | |
| ttx | MikeZ: http://uec-images.ubuntu.com/releases/10.04/release/ | 07:21 |
| ttx | (scroll down to tarballs) | 07:21 |
| *** MarkAtwood has joined #openstack | 07:40 | |
| *** larstobi has joined #openstack | 07:47 | |
| *** opengeard_ has joined #openstack | 07:53 | |
| *** opengeard has quit IRC | 07:54 | |
| *** brd_from_italy has joined #openstack | 08:01 | |
| *** MikeZ1 has joined #openstack | 08:01 | |
| *** MikeZ has quit IRC | 08:04 | |
| *** rcc has joined #openstack | 08:06 | |
| *** allsystemsarego has joined #openstack | 08:13 | |
| * soren screams a little bit | 08:23 | |
| ttx | soren: happy new year to you too :) | 08:23 |
| soren | o/ | 08:24 |
| * soren is cleaning up the iptables-firewall branch after the eventlet merge.. painful. | 08:24 | |
| ttx | hah | 08:24 |
| * ttx reads HTML email on the ML, it's painful too. | 08:25 | |
| * soren takes a break | 08:26 | |
| *** miclorb has joined #openstack | 08:28 | |
| *** ibarrera has joined #openstack | 08:33 | |
| *** DigitalFlux has joined #openstack | 08:36 | |
| DigitalFlux | Hi Guys | 08:39 |
| soren | o/ | 08:42 |
| *** calavera has joined #openstack | 08:44 | |
| *** sassyn has quit IRC | 08:52 | |
| DigitalFlux | Hi soren | 09:00 |
| DigitalFlux | I just added the Nova PPA repo to a lucid 64 installation | 09:00 |
| DigitalFlux | i cannot seem to find the packages when i search for them | 09:01 |
| DigitalFlux | I added the same repo to a maverick 64 installation | 09:01 |
| soren | Which PPA? | 09:01 |
| DigitalFlux | Nova-core/ppa | 09:01 |
| soren | Who or what told you to use that one? | 09:01 |
| DigitalFlux | add-apt-repository ppa:nova-core/ppa | 09:01 |
| DigitalFlux | who -> some installation doc | 09:02 |
| DigitalFlux | checking .. | 09:02 |
| DigitalFlux | http://nova.openstack.org/adminguide/multi.node.install.html | 09:02 |
| soren | You're supposed to use nova-core/trunk. If there's still docs telling you to use nova-core/ppa I want to fix that. | 09:02 |
| soren | Darn it. | 09:02 |
| soren | DigitalFlux: Thanks! | 09:02 |
| DigitalFlux | soren: Thank you :) | 09:02 |
| * soren pulls a bit of hair out | 09:02 | |
| DigitalFlux | So do we have any updates related to the installation docs other than this one ? | 09:03 |
| soren | I've been away for a bit.. | 09:04 |
| soren | ttx: Maybe you can help out? | 09:04 |
| DigitalFlux | Also I'm doing some pilot project to test out what we can do with OpenStack, what release do you recommend for this ? lucid or maverick ? | 09:05 |
| soren | Maverick. | 09:05 |
| DigitalFlux | soren: OK cool | 09:06 |
| soren | https://code.launchpad.net/~soren/nova/ppa-references/+merge/45019 filed to fix PPA references. | 09:09 |
| soren | DigitalFlux: Thanks for the pointers! I really appreciate it. | 09:09 |
| DigitalFlux | soren: Than you really, you're the one who helped not me :-) | 09:10 |
| DigitalFlux | soren: BTW, do you know anybody who is working on that puppet module in the contrib dir ? | 09:10 |
| soren | DigitalFlux: It came from the Anso guys, so you could talk to one of them (vishy or xtoddx, for instance). They're on US West coast time, though, so don't expect much of a response at this hour :) | 09:12 |
| DigitalFlux | soren: Great, Thanks man | 09:13 |
| soren | Sure thing. | 09:14 |
| ttx | No update that I know of. I have to refactor them to weed out junk | 09:16 |
| DigitalFlux | ttx: Indeed, I see some stuff that needs some refactoring, Augeas/inline can be used instead of perl, shell scripts can be replaced too i think .. | 09:18 |
| DigitalFlux | ttx: I would happy to collaborate on this .. | 09:18 |
| ttx | DigitalFlux: I meant, reorganizing the install documents themselves | 09:19 |
| *** befreax has joined #openstack | 09:19 | |
| ttx | DigitalFlux: where do you see perl ? | 09:19 |
| DigitalFlux | ttx: let me check | 09:19 |
| ttx | (all the shell scripts I see are working around distro packaging issues... they should just not be needed) | 09:20 |
| DigitalFlux | ttx: seems like that perl snipper is inside the kern_module.pp, which follows that kernel modules patterns from PuppetLabs | 09:21 |
| DigitalFlux | Anyway, i don't think i fully understood how to use it right away when i opened site.pp | 09:21 |
| DigitalFlux | ttx: that needs some documentation i guess | 09:21 |
| ttx | oh, I see, the puppet stuff | 09:22 |
| DigitalFlux | ttx: it somehow can be modularized a little bit so that people don't get confused with all the extras like Gangalia ..etc.. | 09:22 |
| soren | DigitalFlux: That would be great! | 09:40 |
| soren | DigitalFlux: The current stuff is just what NASA is using (I think). | 09:40 |
| soren | DigitalFlux: It was put there because there was no particular reason to keep it hidden, but I don't think much effort was put into making it super useful to others. That was left as an exercise to someone else. | 09:41 |
| DigitalFlux | soren: OK great, i will check what i can do about it, but will indeed get back here to check with the rest of the community how we can make it plug'n'play and flexible in the same time | 09:41 |
| soren | Any and all help is appreciated. | 09:42 |
| soren | Phew. iptables-firewall branch merged with trunk. | 09:57 |
| soren | That was no fun at all. | 09:58 |
| alekibango | soren: congrats to merge | 10:02 |
| alekibango | and happy new year everyone ! | 10:02 |
| *** opengeard_ has quit IRC | 10:12 | |
| *** opengeard has joined #openstack | 10:14 | |
| *** opengeard has quit IRC | 10:15 | |
| *** opengeard has joined #openstack | 10:16 | |
| *** arthurc has joined #openstack | 10:20 | |
| *** miclorb has quit IRC | 10:25 | |
| *** miclorb_ has joined #openstack | 10:30 | |
| soren | alekibango: Happy new year to you too :) | 10:40 |
| *** miclorb_ has quit IRC | 10:42 | |
| *** matiu has quit IRC | 10:44 | |
| ttx | soren: could you set the hyper-V BMP status to "needs fixing" so that it doesn't show up on the "pending review" list ? | 11:13 |
| * ttx lunches | 11:24 | |
| alekibango | ttx: gj editing RunningNova wiki... | 11:57 |
| DigitalFlux | Oh , seems like the Nova packages in the standard Maverick's repos are kind of old | 12:11 |
| DigitalFlux | still uses multiple config file .. | 12:12 |
| alekibango | use the repo soren gave you for best results :) | 12:12 |
| DigitalFlux | alekibango: Excellet, Thanks | 12:12 |
| *** ctennis has quit IRC | 12:37 | |
| *** pvo has quit IRC | 12:37 | |
| *** [ack] has quit IRC | 12:37 | |
| *** dagger has quit IRC | 12:37 | |
| *** Xenith has quit IRC | 12:37 | |
| *** rossij has quit IRC | 12:37 | |
| *** odyi has quit IRC | 12:37 | |
| *** timrc has quit IRC | 12:37 | |
| *** openstackhudson has quit IRC | 12:37 | |
| *** jero has quit IRC | 12:37 | |
| *** PiotrSikora has quit IRC | 12:37 | |
| *** stewart has quit IRC | 12:37 | |
| *** ArdRigh has joined #openstack | 12:43 | |
| alekibango | i was not here few days, was the nasa web interface released already? | 12:46 |
| DigitalFlux | nova-manage network create cidr num_networks total_ips | 12:49 |
| DigitalFlux | network does not match any options : | 12:49 |
| DigitalFlux | :S | 12:49 |
| DigitalFlux | The command was mentioned in http://wiki.openstack.org/NovaInstall/MultipleServer | 12:50 |
| DigitalFlux | I remember i used nova-manager network before, it is not included with the new packages or what ? | 12:51 |
| ttx | alekibango: no | 12:51 |
| *** ramkrsna has quit IRC | 12:53 | |
| soren | DigitalFlux: I don't think I understand your question. | 12:53 |
| *** pvo has joined #openstack | 12:53 | |
| *** [ack] has joined #openstack | 12:53 | |
| *** stewart has joined #openstack | 12:53 | |
| *** dagger has joined #openstack | 12:53 | |
| *** Xenith has joined #openstack | 12:53 | |
| *** rossij has joined #openstack | 12:53 | |
| *** odyi has joined #openstack | 12:53 | |
| *** timrc has joined #openstack | 12:53 | |
| *** openstackhudson has joined #openstack | 12:53 | |
| *** jero has joined #openstack | 12:53 | |
| *** PiotrSikora has joined #openstack | 12:53 | |
| *** zelazny.freenode.net sets mode: +v pvo | 12:53 | |
| *** kashyapc has quit IRC | 12:53 | |
| DigitalFlux | I did the command -> nova-manage network create cidr num_networks total_ips (replacing the IPs of course) following the doc | 12:54 |
| DigitalFlux | it threw an error: network does not match any options : (list of options) | 12:54 |
| soren | What's the exact options you used? | 12:56 |
| DigitalFlux | nova-manage network create 10.0.255.0/24 4 512 | 12:57 |
| soren | What does "which nova-manage" say? | 12:58 |
| soren | Oh, and that's not going to work. | 12:59 |
| soren | 4*512 == 2048 addresses. They're not going to fit in a /24. | 12:59 |
| *** arthurc has quit IRC | 13:00 | |
| DigitalFlux | I did suspect that :) | 13:00 |
| DigitalFlux | what was i thinking .. | 13:01 |
| soren | I don't know if that's your problem, though. | 13:02 |
| soren | ...but if not, it will be once this other problem is sorted. | 13:02 |
| soren | :) | 13:02 |
| DigitalFlux | Yeah seems like that | 13:03 |
| sandywalsh | o/ | 13:08 |
| ttx | sandywalsh: o/ | 13:10 |
| *** kashyapc has joined #openstack | 13:10 | |
| alekibango | i hope that is not meant to be roman salute | 13:10 |
| *** ppetraki has joined #openstack | 13:10 | |
| *** arthurc has joined #openstack | 13:10 | |
| sandywalsh | Hi ttx, read your api email. I'll talk with _cerberus_ about it this morning. | 13:11 |
| ttx | alekibango: that would be D/ | 13:11 |
| ttx | sandywalsh: ack | 13:11 |
| alekibango | D? never seen that | 13:11 |
| ttx | alekibango: good for you :) | 13:12 |
| alekibango | :) | 13:12 |
| *** hadrian has joined #openstack | 13:12 | |
| DigitalFlux | OK, what about this one -> sudo /usr/bin/python /usr/bin/nova-manage project zip admin admin creds/nova.zip | 13:14 |
| DigitalFlux | what are those admins ? :), I have one user already .. | 13:14 |
| DigitalFlux | AH i see project_id user_id | 13:15 |
| alekibango | yes | 13:15 |
| alekibango | DigitalFlux: are u reading RunningNova wiki now? | 13:15 |
| * soren is having trouble getting his brain back into gear after the holidays | 13:16 | |
| DigitalFlux | and it threw another expection :) | 13:16 |
| DigitalFlux | alekibango: no, i did that before, however i totally forgot what i did before :-) | 13:16 |
| alekibango | soren: dont worry, all others are in similar state | 13:16 |
| alekibango | http://wiki.openstack.org/RunningNova <- | 13:17 |
| alekibango | that will help you | 13:17 |
| alekibango | if not, ask again | 13:17 |
| alekibango | and we will correct the doc | 13:17 |
| DigitalFlux | No such file or directory '/usr/lib/pymodules/python2.6/nova/auth/novarc.template' | 13:18 |
| DigitalFlux | alekibango: Thanks man, really helpful | 13:18 |
| alekibango | np | 13:18 |
| DigitalFlux | That file is not there indeed | 13:18 |
| alekibango | i have seen this error month ago or so, but i cannot remember what i did | 13:18 |
| * DigitalFlux checking the code .. | 13:19 | |
| *** ArdRigh has quit IRC | 13:24 | |
| DigitalFlux | I will grap it from launchpad and see if it works .. | 13:32 |
| DigitalFlux | OK, it's complaining about another file now .. CA | 13:34 |
| *** ctennis has joined #openstack | 13:38 | |
| *** westmaas has joined #openstack | 13:46 | |
| soren | How did you install this? | 13:52 |
| soren | And how are you attemptint to run it? | 13:52 |
| soren | DigitalFlux: ^ | 13:52 |
| *** westmaas has quit IRC | 13:53 | |
| *** hggdh has joined #openstack | 14:01 | |
| *** westmaas has joined #openstack | 14:02 | |
| soren | Err... | 14:04 |
| soren | eday: What's going on here? https://code.launchpad.net/~jk0/nova/diagnostics-per-instance/+merge/44872 | 14:04 |
| *** gaveen has joined #openstack | 14:04 | |
| *** jdarcy has joined #openstack | 14:07 | |
| DigitalFlux | soren: using ppa:nova-core/trunk | 14:08 |
| DigitalFlux | soren: then installed the packages .. | 14:08 |
| soren | DigitalFlux: And where are you seeing these errors? | 14:09 |
| DigitalFlux | soren: while executing nova-manager project zip ... | 14:10 |
| soren | I don't really understand why that would happen. It works fine for me. | 14:18 |
| soren | How did you install the packages? | 14:19 |
| soren | ttx: Do you know what this is about: https://code.launchpad.net/~jk0/nova/diagnostics-per-instance/+merge/44872 ? | 14:20 |
| soren | ttx: All the retrying to tests shenanigans I mean. | 14:20 |
| soren | .-1 s/ to / of / | 14:20 |
| * ttx looks | 14:20 | |
| ttx | soren: you mean why did hudson apparently retry it 3 times in a row ? | 14:22 |
| soren | DigitalFlux: What does "which nova-manage" say? | 14:22 |
| soren | ttx: No, I mean why did Eric ask Hudson to retry it 3 times in a row. | 14:22 |
| ttx | soren: no, I don't know what it's about, I missed all the action too. | 14:23 |
| soren | :( | 14:23 |
| befreax | soren: just did a fresh install and I'm running in the same issues as DigitalFlux...nova-manage is in /usr/local/bin | 14:26 |
| soren | The packages don't install stuff in /usr/local/bin. | 14:27 |
| befreax | interesting... | 14:27 |
| soren | So if you have stuff in /usr/local/bin, you put it there by some other means. | 14:27 |
| befreax | will check - was a fresh installed ubuntu | 14:28 |
| soren | befreax: Maverick? | 14:28 |
| soren | (Not that it should matter, really) | 14:28 |
| befreax | jap | 14:28 |
| befreax | probably I ran one command too many...will check what did go wrong | 14:29 |
| soren | Stuff should only land in /usr/local if you've downloaded the tarball and done a "python setup.py install" or something. | 14:29 |
| *** EdwinGrubbs has joined #openstack | 14:29 | |
| befreax | did a python setup.py install... | 14:30 |
| ttx | nova-core: please set https://code.launchpad.net/~chiradeep/nova/msft-hyper-v-support/+merge/43724 status to "needs fixing" so that it doesn't show up on the "pending review" list (we already know it needs an eventlet refresh) | 14:31 |
| *** Edwin__ has joined #openstack | 14:31 | |
| dabo | eday: Don't know if you saw this on Friday: http://paste.openstack.org/show/362/ | 14:31 |
| soren | befreax: There's your problem. | 14:32 |
| dabo | I got it running this on your compute-api-cleanup branch: ./bin/nova-manage --flagfile=../nova.conf project zip xstest admin | 14:32 |
| soren | ttx: Set it to "work in progress". There's no such mp state as "needs fixing". | 14:32 |
| ttx | soren: ok :) I couldn't browse them to name it appropriately | 14:33 |
| soren | dabo: I explained in a comment on the mp why that happens. | 14:34 |
| befreax | soren: ah :-/ BTW. is there any detailed documentation on API development? | 14:34 |
| *** EdwinGrubbs has quit IRC | 14:34 | |
| soren | befreax: Not that I can think of. | 14:35 |
| befreax | too bad :-/ will dig further through code :-) | 14:35 |
| soren | Feel free to write some :) | 14:35 |
| *** gondoi has joined #openstack | 14:36 | |
| befreax | I feared that statement :-) will do for OCCI first ;-) | 14:36 |
| dabo | soren: I understand the problem; I'm just not sure what needs to be fixed. The diff only shows lines where the import was removed, not added. | 14:37 |
| sandywalsh | hey guys, what is the criteria for something going in the network service or not? | 14:40 |
| sandywalsh | I'm seeing all kinds of calls in compute, api & virt related to network stuff | 14:40 |
| sandywalsh | I know that some operations I need will be guest related and that implies network calling virt ... is that ok? Does that seem like a duplication of the efforts of compute? | 14:41 |
| sandywalsh | perhaps virt should be a service as well? | 14:41 |
| sandywalsh | or is virt just a shared library across all services? | 14:42 |
| sandywalsh | (should this question go to ozone mailing list?) | 14:43 |
| *** arthurc_ has joined #openstack | 14:43 | |
| * sandywalsh over | 14:43 | |
| *** arthurc has quit IRC | 14:44 | |
| ttx | sandywalsh: I think that's a good discussion for the *openstack* ML, with maybe more precise examples of operations you'd need | 14:47 |
| sandywalsh | whoops :) openstack | 14:47 |
| sandywalsh | thanks ttx | 14:47 |
| ttx | I've been asking myself a few questions about compute making network calls as well :) | 14:47 |
| sandywalsh | and api allocating mac addresses on instances | 14:48 |
| ttx | looks like the separation of nodes and modules is not a perfect overlap | 14:48 |
| sandywalsh | yup. k, I'm going to do a full pass on the flow today and try and write something up eod | 14:49 |
| *** arthurc_ has quit IRC | 14:52 | |
| *** trin_cz has joined #openstack | 14:52 | |
| *** arthurc has joined #openstack | 14:52 | |
| *** dendrobates is now known as dendro-afk | 14:55 | |
| trin_cz | Hi everyone | 14:56 |
| trin_cz | I've started playing with openstack on opensuse. Does anyone of you see something, which could prevent that? | 14:57 |
| *** stewart has quit IRC | 15:00 | |
| wedjat | hi everyone ! | 15:02 |
| wedjat | for what exactly is memcached used by swift ? for authentication only right ? | 15:02 |
| *** spectorclan has joined #openstack | 15:06 | |
| *** dendro-afk is now known as dendrobates | 15:07 | |
| DigitalFlux | solved by a re-install | 15:07 |
| ttx | trin_cz: no | 15:07 |
| *** jtimberman has quit IRC | 15:07 | |
| trin_cz | ttx: thanks | 15:08 |
| *** uvirtbot` is now known as uvirtvot | 15:09 | |
| *** uvirtvot is now known as uvirtbot | 15:09 | |
| *** sophiap has joined #openstack | 15:09 | |
| *** EdwinGrubbs has joined #openstack | 15:14 | |
| *** EdwinGrubbs has joined #openstack | 15:14 | |
| *** Edwin__ has quit IRC | 15:14 | |
| notmyname | wedjat: memcache is used in swift for auth credentials and also container and account existence | 15:15 |
| *** stewart has joined #openstack | 15:21 | |
| wedjat | notmyname: ok, so objects can be cached ? | 15:27 |
| *** iRTermite has joined #openstack | 15:29 | |
| *** dendrobates is now known as dendro-afk | 15:38 | |
| *** Edwin__ has joined #openstack | 15:39 | |
| *** EdwinGrubbs has quit IRC | 15:42 | |
| *** Edwin__ has quit IRC | 15:46 | |
| *** EdwinGrubbs has joined #openstack | 15:46 | |
| *** EdwinGrubbs has joined #openstack | 15:46 | |
| soren | do any of you guys grok openvswitch? | 15:48 |
| notmyname | wedjat: no, objects aren't cached in swift | 15:49 |
| dubs | soren: i think antonym has some experience with it | 15:49 |
| *** sophiap_ has joined #openstack | 15:49 | |
| *** sophiap has quit IRC | 15:50 | |
| *** sophiap_ is now known as sophiap | 15:50 | |
| *** reldan has joined #openstack | 15:50 | |
| soren | dubs: Ok, thanks. | 15:53 |
| * soren pauses for a couple of hours for dinner and such | 15:53 | |
| wedjat | notmyname: ok, thanks ! :) | 15:54 |
| notmyname | wedjat: it could be possible to cache objects (maybe with some extra middleware), but you would have to be careful because swift can store objects larger than your cache | 15:55 |
| *** Edwin__ has joined #openstack | 15:55 | |
| creiht | wedjat: yeah if you wanted to cache objects, you would probably want to layer something on top of swift for that | 15:57 |
| creiht | perhaps something like varnish | 15:58 |
| notmyname | creiht: some sort of varnish middleware.....yeah, that | 15:58 |
| *** EdwinGrubbs has quit IRC | 15:59 | |
| *** jimbaker has joined #openstack | 15:59 | |
| *** jimbaker has quit IRC | 16:00 | |
| *** jimbaker has joined #openstack | 16:01 | |
| *** dendro-afk is now known as dendrobates | 16:01 | |
| wedjat | creiht, notmyname: yeah, would be very interesting ! i would like to offer some service to customers where i could be able to make available the same object for a great number of customers | 16:03 |
| wedjat | i will try that ^^ | 16:04 |
| creiht | wedjat: yeah at that point you are pretty much offering a CDN like capability on top of swift | 16:04 |
| jk0 | soren: re that hudson test, there was a weird bug in my branch that we didn't catch right away | 16:04 |
| notmyname | wedjat: essentially, the "extra layer of caching on top of swift" is what Rackspace does with our CDN provider. the CDN is a globally distributed object cache | 16:04 |
| creiht | which shouldn't be that difficult | 16:04 |
| jk0 | soren: it took a few tries to figure it out | 16:04 |
| creiht | hehe | 16:04 |
| *** guigui has quit IRC | 16:05 | |
| *** zaitcev has joined #openstack | 16:10 | |
| *** dubsquared1 has joined #openstack | 16:12 | |
| *** DigitalFlux has quit IRC | 16:13 | |
| wedjat | creiht, notmyname: ok, so swift is used by Rackspace now ? | 16:13 |
| alekibango | wedjat: yes it is | 16:13 |
| notmyname | yes. swift is cloud files open-sourced | 16:13 |
| wedjat | very impressive ! :) | 16:15 |
| *** dendrobates is now known as dendro-afk | 16:16 | |
| *** jtimberman has joined #openstack | 16:18 | |
| *** calavera has quit IRC | 16:21 | |
| *** littleidea has joined #openstack | 16:24 | |
| jdarcy | Aren't there some differences around authentication? | 16:30 |
| *** dendro-afk is now known as dendrobates | 16:30 | |
| *** arreyder has joined #openstack | 16:31 | |
| ttx | vishy / anso folks: who is working on https://blueprints.launchpad.net/nova/+spec/snapshot-instance ? | 16:32 |
| creiht | jdarcy: in which way? | 16:32 |
| ttx | it's assigned to "Anso Labs"... and has "Unknown" implementation status. | 16:32 |
| creiht | There are some very small differences that are very specific to RS, and don't add any value outside of rackspace | 16:32 |
| *** kashyapc has quit IRC | 16:33 | |
| *** EdwinGrubbs has joined #openstack | 16:36 | |
| *** EdwinGrubbs has joined #openstack | 16:36 | |
| *** Edwin__ has quit IRC | 16:36 | |
| jdarcy | creiht: Yeah, that's what I figured. I would've been surprised if the auth system used within RSC was *exactly* the same as what's in Swift. Everyone's going to have some small differences. | 16:39 |
| *** schisamo has joined #openstack | 16:46 | |
| *** kashyapc has joined #openstack | 16:50 | |
| creiht | jdarcy: yeah, the dev auth stuff does emulate some RS type things which makes testing a little easier for us | 16:53 |
| Ryan_Lane | anyone mind doing a review for me? :) https://code.launchpad.net/~rlane/nova/lp681164/+merge/44930 | 17:00 |
| *** brd_from_italy has quit IRC | 17:00 | |
| *** Edwin__ has joined #openstack | 17:01 | |
| *** EdwinGrubbs has quit IRC | 17:01 | |
| *** trin_cz has quit IRC | 17:15 | |
| *** Edwin__ has quit IRC | 17:21 | |
| *** EdwinGrubbs has joined #openstack | 17:24 | |
| *** rcc has quit IRC | 17:25 | |
| *** reldan has quit IRC | 17:26 | |
| *** Edwin__ has joined #openstack | 17:27 | |
| *** EdwinGrubbs has quit IRC | 17:30 | |
| *** EdwinGrubbs has joined #openstack | 17:35 | |
| *** EdwinGrubbs has joined #openstack | 17:35 | |
| *** larstobi has quit IRC | 17:35 | |
| *** ibarrera has quit IRC | 17:36 | |
| *** Edwin__ has quit IRC | 17:36 | |
| *** Edwin__ has joined #openstack | 17:44 | |
| *** EdwinGrubbs has quit IRC | 17:44 | |
| *** reldan has joined #openstack | 17:49 | |
| *** EdwinGrubbs has joined #openstack | 17:55 | |
| *** Edwin__ has quit IRC | 17:55 | |
| *** daleolds has joined #openstack | 18:00 | |
| *** joearnold has joined #openstack | 18:02 | |
| openstackhudson | Project nova build #350: SUCCESS in 1 min 17 sec: http://hudson.openstack.org/job/nova/350/ | 18:14 |
| *** Edwin__ has joined #openstack | 18:14 | |
| *** maplebed has joined #openstack | 18:15 | |
| *** EdwinGrubbs has quit IRC | 18:17 | |
| *** maplebed has quit IRC | 18:19 | |
| *** maplebed has joined #openstack | 18:20 | |
| *** dirakx has joined #openstack | 18:20 | |
| *** WonTu has joined #openstack | 18:25 | |
| *** WonTu has left #openstack | 18:25 | |
| *** nelson__ has quit IRC | 18:25 | |
| *** nelson__ has joined #openstack | 18:26 | |
| *** EdwinGrubbs has joined #openstack | 18:27 | |
| *** Edwin__ has quit IRC | 18:27 | |
| *** jc_smith has joined #openstack | 18:30 | |
| *** johnpur has joined #openstack | 18:32 | |
| *** ChanServ sets mode: +v johnpur | 18:32 | |
| nelson__ | annegentle: hi. | 18:32 |
| *** trin_cz has joined #openstack | 18:34 | |
| *** MikeZ1 has quit IRC | 18:35 | |
| * nelson__ looks around for someone to review his patch and spots creiht. https://code.launchpad.net/~nelson-crynwr/swift/refactor/+merge/44964 | 18:35 | |
| creiht | hehe | 18:36 |
| * creiht is still catching up from beeing on 2 weeks vacation :) | 18:37 | |
| annegentle | nelson__: hey | 18:38 |
| * nelson__ looks around for another victim and spots annegentle | 18:38 | |
| nelson__ | annegentle: hey, so I just (late last week) noticed that howto_installmultinode.html or at least the source for it is in trunk | 18:39 |
| creiht | nelson__: http://swift.openstack.org/howto_installmultinode.html | 18:39 |
| nelson__ | If I suggest changes to the source file (ja, that's the one) via bzr, is that the best way to get them to you? | 18:40 |
| sandywalsh | how do calls in SimpleScheduler get bounced to the appropriate service? I don't see any code that does the relay. | 18:41 |
| *** befreax has quit IRC | 18:42 | |
| nelson__ | annegentle: because ... I was noticing that those instructions presume Ubuntu ... so I was thinking of making the instructions work a little better. | 18:46 |
| annegentle | nelson__: ah, great! Sorry I didn't make that clear. :) | 18:47 |
| annegentle | nelson__: clarity as to where the docs live :) | 18:47 |
| nelson__ | Okay, I'll do that. | 18:48 |
| nelson__ | annegentle: and I'm guessing that you'll get an email about it when I propose the merge, just like you got the one I proposed to swift-ring-builder. correct? | 18:48 |
| sandywalsh | I should be a little more clear. I see run_instance stuffs a message on the scheduler queue, but I don't see how that gets routed to the compute queue. | 18:50 |
| *** Edwin__ has joined #openstack | 18:51 | |
| sandywalsh | it appears I'm using the Chance scheduler | 18:51 |
| annegentle | nelson__: the merge proposals go to core teams, of which I'm not a member, but you can add me as a requested reviewer I believe | 18:51 |
| *** EdwinGrubbs has quit IRC | 18:52 | |
| annegentle | nelson__: you can ping me anywhich way you want, really. I'm happy to review doc! That first sentence may not be quoted out of context. Hee. :) | 18:52 |
| nelson__ | annegentle: :) You didn't expect a sexist remark, now did you?? | 18:53 |
| creiht | hah | 18:53 |
| creiht | nelson__: yeah if it is doc related, we will make sure annegentle takes a look at it | 18:53 |
| creiht | and thanks for the suggestions in advance | 18:53 |
| creiht | nelson__: and I did see your merge request for the ring builder stuff, but we are all a bit busy with internal stuff at the moment | 18:54 |
| creiht | so it may be a bit before anyone looks at it | 18:54 |
| *** shermanboyd has joined #openstack | 18:55 | |
| creiht | ttx: speaking of which, I highly doubt I will be able to get the swift S3 compatibility done for Bexar | 18:55 |
| annegentle | nelson__: no, no. :) | 18:55 |
| *** reldan has quit IRC | 18:55 | |
| nelson__ | Okay. It's "just" a documentation patch, natch it's not a priority. (/me looks at annegentle for moral support). | 18:56 |
| shermanboyd | Hi, I just discovered openstack and I'm wondering if it's production ready for iaas | 18:56 |
| annegentle | nelson__: you've got it! | 18:56 |
| nelson__ | ( creiht: I'm just teasing. I understand about priorities. ) | 18:56 |
| devcamcar | shermanboyd: nasa is running nebula in production using openstack: http://nebula.nasa.gov/ | 18:57 |
| Ryan_Lane | shermanboyd: it really depends on what features you consider making it production ready | 18:58 |
| shermanboyd | let's see | 18:58 |
| *** rlucio has joined #openstack | 18:59 | |
| shermanboyd | does it provide for network isolation | 18:59 |
| shermanboyd | and security | 18:59 |
| *** MikeZ has joined #openstack | 19:00 | |
| shermanboyd | how does it compare to something like enomaly | 19:00 |
| *** yosh_ is now known as yosh | 19:01 | |
| *** MikeZ1 has joined #openstack | 19:01 | |
| shermanboyd | the openstack website does a great job of telling me how great openstack is and how lot's organizations are behind it, but it doesn't do a great job explaining what it is | 19:03 |
| *** MikeZ has quit IRC | 19:04 | |
| annegentle | shermanboyd: you'd probably appreciate more detail at nova.openstack.org, like http://nova.openstack.org/adminguide/managing.projects.html talks about how projects provide the separate networks and so on | 19:04 |
| Ryan_Lane | shermanboyd: network and isolation is available via security groups | 19:04 |
| Ryan_Lane | err | 19:04 |
| annegentle | shermanboyd: nova=OpenStack Compute, and swift=OpenStack Object Storage | 19:04 |
| Ryan_Lane | network isolation and security | 19:05 |
| notmyname | shermanboyd: Rackspace is running swift in production and NASA is running nova in production, both at large scale | 19:05 |
| shermanboyd | is openstack more of a toolkit or library? | 19:06 |
| annegentle | shermanboyd: more on security groups for network traffic: http://nova.openstack.org/nova.concepts.html#concept-security-groups | 19:06 |
| alekibango | shermanboyd: its stack :) | 19:06 |
| shermanboyd | annegentle: thank you | 19:06 |
| alekibango | i would say its cloud operating system | 19:07 |
| shermanboyd | alekibango: like rails for clouds | 19:07 |
| notmyname | shermanboyd: toolkit, I suppose. think of it as a software package (like apache or hadoop or rails or whatever) that you install in a DC rather than on one machine | 19:07 |
| notmyname | although, for testing, you can install it on one machine | 19:08 |
| alekibango | shermanboyd: there are libs you can use to call openstack api | 19:08 |
| notmyname | and openstack is a collection of different pieces. nova and swift and whatever comes next | 19:08 |
| alekibango | notmyname: and glance :) which connects those 2 | 19:08 |
| notmyname | sorry, didn't mean to leave anything out ;-) | 19:09 |
| rlucio | i had a merge fail on an approved branch.. now that i've fixed the issue do i need to resubmit the merge proposal or anything? | 19:09 |
| shermanboyd | so if I want to start my own iaas based on openstack I need to be prepared to do a bit of programming | 19:09 |
| alekibango | rlucio: after push i would resubmit merge proposal | 19:10 |
| *** jdurgin has joined #openstack | 19:10 | |
| rlucio | alekibango: ok, thanks | 19:10 |
| alekibango | i am not 1000% sure, i did that only once heh | 19:10 |
| rlucio | oh | 19:10 |
| alekibango | but it will not harm | 19:10 |
| rlucio | true | 19:11 |
| alekibango | which is your merge request? | 19:11 |
| rlucio | https://code.launchpad.net/~rlucio/nova/lp690314/+merge/45064 | 19:11 |
| alekibango | ah this one | 19:12 |
| rlucio | yep :) | 19:12 |
| annegentle | shermanboyd: have a healthy sense of stick-to-it and a patience for oft-changing code, I'd say. | 19:12 |
| notmyname | shermanboyd: probably, but it depends on your needs. you probably shouldn't have to change the core code, but you may have to write something to interact with your auth/ident system or any billing system you may have or even just some operational deployment tools | 19:13 |
| annegentle | shermanboyd: starting your own IaaS with any toolkit probably requires programming (right, what notmyname said) | 19:13 |
| shermanboyd | annegentle: true, some more than others | 19:13 |
| shermanboyd | what about hardware architecture; is there one machine that is a central controller or is it distributed. Is there a document that shows a typical deployment? | 19:17 |
| notmyname | shermanboyd: think of it as "you get what you pay for". if you put in a little work, you will get a system that will scale a little bit. put in a lot of work (tuning parameters, selecting hardware, etc) you will get a system that will scale to massive levels | 19:18 |
| notmyname | there is no central element to any part of openstack. everything is distributed (horizontally scalable) | 19:19 |
| shermanboyd | notmyname: that's nice | 19:19 |
| ttx | creiht: ok | 19:19 |
| notmyname | shermanboyd: check out wiki.openstack.org | 19:19 |
| notmyname | especially the developer documentation sections | 19:19 |
| *** kashyapc has quit IRC | 19:25 | |
| creiht | ttx: well I may have to take that back | 19:28 |
| creiht | looks like someone actually did some work on the S3 stuff while I was on vacation | 19:29 |
| creiht | I'll have to take a look at it, re-evaluate, and then let you know | 19:29 |
| *** bingoo has joined #openstack | 19:30 | |
| *** bingoo has quit IRC | 19:31 | |
| *** bingoo has joined #openstack | 19:31 | |
| soren | antonym: Word on the street says you've worked with openvswitch? | 19:35 |
| *** jbryce has joined #openstack | 19:39 | |
| openstackhudson | Project nova build #351: SUCCESS in 1 min 20 sec: http://hudson.openstack.org/job/nova/351/ | 19:44 |
| *** opengeard has quit IRC | 19:46 | |
| *** opengeard has joined #openstack | 19:47 | |
| *** opengeard has joined #openstack | 19:48 | |
| openstackhudson | Project nova build #352: SUCCESS in 1 min 21 sec: http://hudson.openstack.org/job/nova/352/ | 19:49 |
| soren | eday: re https://code.launchpad.net/~eday/nova/instance-uuids/+merge/43301.. | 19:51 |
| soren | eday: Which component generates the id? | 19:51 |
| *** sirp- has joined #openstack | 19:51 | |
| soren | eday: API server? Scheduler? | 19:51 |
| eday | soren: right now API server, but the goal is to be able to do it anywhere | 19:52 |
| eday | soren: anything consuming nova.compute/API | 19:52 |
| *** bingoo has quit IRC | 19:52 | |
| soren | Why would a network node create a new instance? | 19:53 |
| eday | soren: they wouldn't | 19:53 |
| *** bingooo has joined #openstack | 19:53 | |
| soren | Then you've lost me :) | 19:53 |
| soren | Why is it a goal to be able to do it anywhere? | 19:54 |
| openstackhudson | Project nova build #353: SUCCESS in 1 min 20 sec: http://hudson.openstack.org/job/nova/353/ | 19:54 |
| eday | I enver said anything about network nodes creating instances? | 19:54 |
| eday | right now ec2 or openstack API, lets keep it simple | 19:54 |
| soren | No, but surely they won't actually be created "anywhere". | 19:54 |
| eday | but in the future the compute worker may be the thing generating the ID for the request | 19:55 |
| soren | I really do hope that we'll eventually actually settle on something. | 19:55 |
| soren | Anyways, my point is this: | 19:56 |
| eday | so do I, but it seems everytime we have a direction something/someone enters the discussion and it's all rehashed | 19:56 |
| *** bingooo has quit IRC | 19:56 | |
| soren | What I'd like is to have a way to have a key space that is manageable. | 19:57 |
| soren | I don't think 128 bit is manageable. | 19:57 |
| soren | I also don't like collisions. | 19:57 |
| soren | To make those ends meet, we have a number of options. | 19:57 |
| *** gasbakid has joined #openstack | 19:57 | |
| soren | We can have a central thingamajig that hands them out. It can know for sure that it never gives out the same one twice. | 19:57 |
| soren | This sucks because it's a great bit SPOF. | 19:58 |
| soren | We can split up the keyspace somehow, so that each component that might give out ID's is granted a certain prefix of a smaller keyspace. | 19:58 |
| annegentle | any (two) nova core members want to be super productive in 2011 and approve this merge proposal? install script and doc updates :) https://code.launchpad.net/~annegentle/nova/newscript/+merge/45067 | 19:59 |
| soren | So, with a 40 bit key space, the first 16 bits could denote the component that hands it out. | 19:59 |
| soren | That would avoid collisions, but would require them all to know about each other at all times, or the set would have to be very static. | 20:00 |
| soren | Or! | 20:00 |
| eday | so you're saying every process gets their own prefix? sounds like a deployment nightmare | 20:00 |
| soren | Exactly. | 20:00 |
| soren | It sucks. | 20:00 |
| eday | you've pushed the central delgation (of the prefix) up to another level, puppet or human | 20:00 |
| soren | Exactly. | 20:01 |
| soren | And that sucks. | 20:01 |
| soren | So, instead.. | 20:01 |
| soren | The problem with a SPOF is that we'd be screwed if it died. | 20:01 |
| eday | so, I've had this discussion many times, at least 3 in the context of Nova (one at the last design summit), and we usually settle on just using 128 keys so no central coordination is required | 20:01 |
| soren | What if we accepted a SPOF, but made it so that we weren't screwed if it disappeared for a second (or a couple of days)? | 20:02 |
| eday | UUIDs, and pvo's idea of ipv6 address, seem to work well | 20:02 |
| soren | We could have a central authority that would hand out ID's in huge batches. | 20:02 |
| devcamcar | annegentle: i'll take a look at it now :) | 20:02 |
| eday | batched reservations would be ok, but then we're running another service just to hand out IDs, which is also a headache | 20:02 |
| *** sirp- has left #openstack | 20:03 | |
| soren | When a new component comes along that needs to be able to hand out IDs, it could request a million of them. | 20:03 |
| soren | Once it's down to a couple hundred thousand, it could replenish its supply. | 20:03 |
| dubsquared1 | Soren: Are you the proper authority to ask how close the trunk source is with the nova-* packages? | 20:03 |
| soren | That way, we can lose the SPOF for quite a while without loss of availabilty. | 20:03 |
| eday | I don't see the problem with 128 bit IDs.. most userrs won't even see them, and if they do, it will be a link (will usually show display name of instance not ID) | 20:03 |
| soren | dubsquared1: I am. | 20:04 |
| soren | eday: Every single user sees them. | 20:04 |
| *** gasbakid has quit IRC | 20:04 | |
| dubsquared1 | when you have a moment, could we chat? i dont mean to interrupt ;) | 20:04 |
| soren | eday: All the time. That's how you refer to your instances in the EC2 world. | 20:04 |
| eday | soren: we're not in the ec2 world for long | 20:04 |
| soren | dubsquared1: I can (usually) multiplex conversations. | 20:04 |
| soren | dubsquared1: What's the problem? | 20:04 |
| annegentle | devcamcar: thanks a bunch | 20:04 |
| *** gasbakid has joined #openstack | 20:05 | |
| soren | eday: What does that mean? | 20:05 |
| eday | ec2 is not our focus API, openstack is | 20:05 |
| soren | eday: That's a poor reason to screw EC2 users over if we can avoid it. | 20:05 |
| eday | and in my rackspace UI, I see my user-assigned named for all ops, ie, "dns", not some random int | 20:05 |
| dubsquared1 | soren: I recall seeing a 'versioning' merge discussed on the bug tracker, do you know if this has been implemented. Secondly, how often are the packages updated to included the latest fixes that are proposed? | 20:06 |
| eday | even with ec2, I don't think we're screwing anyone, it's just a bit longer of an ID now | 20:06 |
| dubsquared1 | I wonder due to the types of VMs I can use to test, raw disk vs kernel/ramdisk | 20:06 |
| eday | that's what copy/paste is for | 20:06 |
| soren | dubsquared1: Every time stuff gets merged into trunk, we build new packages. | 20:06 |
| devcamcar | annegentle: is there a reason you removed the us gov copyright? usually we end up with both the us gov copyright as well as the openstack one | 20:06 |
| soren | eday: 77dwoeu4 is a lot easier on the eyes than 02216400-1775-11e0-9463-001f1638bd59. | 20:07 |
| eday | soren: of course, but just as easy to copy/paste | 20:07 |
| soren | eday: People using the EC2 API see these ID's *all* the time. | 20:07 |
| eday | perhaps I'm a minority, but I wouldn't really think twice about it.. perhaps a "oh, look, it's longer", and them move on | 20:08 |
| soren | eday: I got *really* annoyed for a few days when the numeric->ec2 conversion got lost (a couple of months ago). | 20:08 |
| soren | It's an eye sore. | 20:08 |
| annegentle | devcamcar: in this case the source is from a Racker, not Anso... so the OpenStack one seemed appropriate, what do you think? | 20:09 |
| soren | dubsquared1: I'm not sure what you're referring to wrt to "versioning"? | 20:09 |
| eday | soren: well, not sure what to say... I see this as the easiest path forward, as did others at the summit (I *thought* we had some kind of consensus there, guess not) | 20:09 |
| dubsquared1 | soren: just being able to see what version of nova you are running | 20:10 |
| soren | dubsquared1: I'm cool with 128 internal ID's (if we really must have a flat key space). | 20:10 |
| soren | Whoops. | 20:10 |
| devcamcar | annegentle: not a big deal, just nitting. i assume that a huge doc like this has contributors from both | 20:10 |
| soren | eday: I'm cool with 128 internal ID's (if we really must have a flat key space). | 20:10 |
| dubsquared1 | soren: me too! | 20:10 |
| soren | eday: It's just that I actually have to look at them all the time. | 20:10 |
| annegentle | devcamcar: nits are welcome, always :) | 20:10 |
| soren | dubsquared1: That's really separate from the packaging effort, don't you think? | 20:11 |
| *** joearnold has quit IRC | 20:11 | |
| dubsquared1 | soren: i didn't know where along it was :D like i said, i just barely recall seeing it... | 20:11 |
| soren | dubsquared1: I'd love to see a patch that lets us somehow query the exact version of Nova you're interacting with. | 20:12 |
| soren | eday: So they're not just internal. They're exposed to the user (i.e. me). | 20:12 |
| soren | dubsquared1: I'm just don't think it should be specific to the packaging. | 20:14 |
| eday | soren: yup, but we can have new tools reference display/server name, which is what the RS tools do | 20:14 |
| eday | at least for the UI presentation, it still would send the UUID under the hood | 20:14 |
| eday | and ambiguities can be resolved at the UI presentation layer | 20:14 |
| alekibango | soren: how query the version? using api? | 20:14 |
| eday | but I know this doesn't help existing EC2 tools at all | 20:14 |
| dubsquared1 | soren: alright, good to know…curious | 20:14 |
| soren | alekibango: Maybe. | 20:14 |
| eday | soren: think of it this way, it will get folks used to seeing 128 bit numbers for IPv6 | 20:15 |
| alekibango | should nova work when disconnected from internet? it fails to connect to 8.8.8.8 (ouch!) | 20:16 |
| *** Ryan_Lane is now known as Ryan_Lane|food | 20:17 | |
| soren | eday: I just want to give everyone a good experience with Nova. Including myself (and consumers of the EC2 API). | 20:18 |
| devcamcar | alekibango: ouch! i run it commonly in my dev env, for instance on an airplane or whatever with no internet and haven't had those problems. i'm a few revisions behind trunk on that env though | 20:18 |
| soren | eday: http://paste.openstack.org/show/363/ | 20:18 |
| soren | alekibango: It doesn't connect to it. | 20:18 |
| alekibango | i was trying to run it while on mountains... | 20:18 |
| soren | alekibango: It just really likes to have a route to it. | 20:18 |
| *** dirakx1 has joined #openstack | 20:19 | |
| soren | alekibango: It uses it to try to find its external IP. | 20:19 |
| *** dirakx has quit IRC | 20:19 | |
| soren | alekibango: If you don't have an external IP, you're sort of screwed. | 20:19 |
| alekibango | this should be configurable using a flag? | 20:19 |
| alekibango | what if i want to run a cloud in my closed closet? | 20:20 |
| alekibango | :) | 20:20 |
| eday | soren: well, if we want to put in extra work to make the ec2 tools look pretty, fine, but I was really hoping to not have to write or introduce a dependency for an ID service and introduce a SPOF (even if batched). | 20:20 |
| soren | alekibango: Patches happily accepted. | 20:23 |
| soren | alekibango: It used to be that you actually had to be able to reach www.google.com on port 80 for it to not fail. | 20:23 |
| soren | So you'd not only do a DNS lookup for www.google.com, but also open a TCP connection to it. | 20:24 |
| alekibango | you know how some people feel about google :) | 20:24 |
| creiht | alekibango: add an etc/hosts entry to point 8.8.8.8 to localhost :) | 20:24 |
| soren | You just need a route. | 20:24 |
| alekibango | i added dummy interface | 20:24 |
| alekibango | with 8.8.8.8 -- actually it was alias on eth1 | 20:25 |
| soren | ip route add 8.8.8.8 dev lo | 20:25 |
| soren | Or whatever. | 20:25 |
| soren | alekibango: As long as you don't do it on a production system, that's fine. | 20:25 |
| soren | eday: It's about providing an experience that's not much worse than what you get on the real EC2. | 20:27 |
| soren | eday: I'm even ignoring the fact that there's almost certainly tools out there that expect the ID's they get back from amazon to fit in a CHAR(16) or whatnot. | 20:28 |
| alekibango | but it didnt help me much. so i teached my sons some ice skating :) | 20:28 |
| eday | soren: and that's what the openstack tools can provide, we shouldn't compromise the correct technical path (IMO) for a display issue with what will be legacy tools for the project | 20:28 |
| soren | Perhaps I missed a memo somewhere, but are we supposed to ignore the *massive* ecosystem around EC2 and the countless tools that have been written to interact with that very API? | 20:30 |
| eday | I'm not suggesting we ignore it, but at some point (soon, probably) the feature set will outgrow the ec2 capabilities, and we're going to need to use the new openstack API/tools to access any of those features | 20:31 |
| soren | Our stuff can be as technically correct and lovely as we want it to be, but if people can't use their existing tools with it, we've failed IMO. | 20:32 |
| soren | Lots of people are perfectly happy with the stuff EC2 provides. | 20:32 |
| soren | ...and have tools that interact with EC2. I would hate to destroy their migration path to a freeer cloud. | 20:33 |
| alekibango | we need both... | 20:33 |
| eday | I'm not suggesting we drop support or anything here | 20:33 |
| soren | You sort of are. | 20:33 |
| eday | *sigh* not at all | 20:33 |
| soren | Not intentionally. | 20:33 |
| alekibango | soren: is it only display issue? | 20:33 |
| eday | I'm suggesting the IDs get larger | 20:33 |
| soren | Yes. | 20:33 |
| eday | and if that exceeds some limits, we file a bug against the tool to increase it | 20:34 |
| soren | Why would people bother? | 20:34 |
| eday | because you want to use nova with the old tools, as you just suggested? | 20:35 |
| alekibango | i would gladly use python-cloudfiles or other openstack api compatible tools instead of ec2, but i agree we need to make it easy to migrate in/out | 20:35 |
| soren | I know why *I* would, but not why anyone else would. | 20:35 |
| soren | Lots of people are perfectly happy with EC2. The API and the service. | 20:35 |
| eday | I'm very well aware of this | 20:35 |
| eday | but we can't change it, so we need to focus on our own API | 20:36 |
| soren | So what is their motivation to change their tools to accomodate a change in a piece of software foreign to E"? | 20:36 |
| soren | E" == EC2. | 20:36 |
| dubsquared1 | python-cloudservers bindings are great, if they were the Rackspace default tool (and was advertised as such) folks would assume the RS api was the same as ec2's IMO | 20:36 |
| soren | (If you type poorly enough) | 20:36 |
| eday | soren: why does anyone but services from anyone else besides amazon? or setup a private cloud based on ec2? | 20:37 |
| eday | I thought we were triyng to build an open, distributed architecture, which ec2 will not be part of for long because they are not open | 20:37 |
| soren | Nevertheless, we're providing an EC2 API. | 20:38 |
| soren | Sort of. | 20:38 |
| soren | Except it's different. | 20:38 |
| soren | Possibly (even likely) in a way that's incompatible with a number of tools the are currently used to consume the EC2 api. | 20:38 |
| eday | which will be used even less as we add more features to the API we can control | 20:39 |
| soren | Just because a *user* of a piece of software wants to use something other than EC2 doesn't mean that the *author* of the the software cares. At all. | 20:39 |
| eday | ec2 should be a "best effort" API, we shouldn't let it compromise the features/architecture we want | 20:40 |
| JordanRinke | eday: +1 | 20:40 |
| eday | rackspace and other non-ec2 cloud providers are doing quite well without the ec2 api | 20:41 |
| soren | I want nova to elegantly accommodate multiple API's. | 20:41 |
| soren | And the EC2 API has a *massive* ecosystem. | 20:41 |
| alekibango | for cloud migration, libcloud / deltacloud? | 20:42 |
| soren | Likely orders of magnitude larger than any of the others. | 20:42 |
| eday | I know, but it will never support everything we need it to because we can't control it | 20:42 |
| soren | eday: I don't follow your argument at all. | 20:42 |
| eday | openstack adds feature X, and amazon doesn't (and maybe never will) | 20:43 |
| soren | eday: You're arguing why we can't change the API, yet you're half insisting that we do it anyway (by using a (much) differentlly sized ID space). | 20:43 |
| eday | user of openstack wants feature X, so they use the openstack/nova tools, not some ec2 tools that uses limited functionality | 20:43 |
| eday | we already have some features like shared IP groups that don't map to ex2 | 20:43 |
| dendrobates | both of you guys are right, ec2 has a massive ecosystem, and we don't want to cede the api. The cloud ecosystem is still a small fraction of what it will be in several years. | 20:44 |
| soren | eday: I don't follow your point at all. | 20:44 |
| soren | eday: Yes, there's stuff we want to do that the EC2 API doesn't/won't expose. | 20:44 |
| eday | soren: my argument is we shouldn't be bound by what ec2 tools can/can't support | 20:44 |
| soren | eday: That's a poor reason to change the backend in a way that very likely will break one of our frontends. | 20:45 |
| eday | we can do the best we can, but at the end of the day, we should do what is correct for the architecture and OpenStack API | 20:45 |
| dubsquared1 | Creating an OS API, that allows us to do everything that is developed for OS, and if folks are happy with a 'dumbed' down version of an API, allow them to use the 'ec2' API that is a best effort/as the community wants dev. On that note, where does Euca fall into this debate? Aren't they coming on strong with the euca2ools, and does that play into the ec2 API in it's entirety? | 20:45 |
| eday | soren: ok, then we don't build a truly distributed system and introduce a SPOF? that's also against our original design tenets | 20:45 |
| soren | Any feature that limits our main goals must be optional | 20:46 |
| eday | supporting ec2 tools is a poor reason to make that compromise | 20:46 |
| soren | Is another one of our basic design tenets. | 20:46 |
| eday | (also, many tools should still work fine with the longer IDs) | 20:47 |
| *** maplebed has quit IRC | 20:47 | |
| soren | I've never said otherwise. | 20:47 |
| alekibango | i love those tenets and yes u both are right... | 20:47 |
| eday | yes you did, you said some tools may have an arbitrary 16 char limit on the IDs | 20:47 |
| soren | eday: Yes. | 20:48 |
| eday | and to that I said we/someone can submit a patch if thats the case | 20:48 |
| soren | eday: I never said that all consumers of the EC2 API had such a limitation. | 20:48 |
| eday | I didn't say you did :) | 20:49 |
| soren | eday: I'm saying that I'm confident that some do. Conversely, some don't. | 20:49 |
| soren | 20:47 <+eday> (also, many tools should still work fine with the longer IDs) | 20:49 |
| soren | 20:47 <+soren> I've never said otherwise. | 20:49 |
| soren | 20:47 <+eday> yes you did, you said some tools may have an arbitrary 16 char limit on the IDs | 20:49 |
| soren | There. | 20:49 |
| soren | "yes you did" | 20:49 |
| *** sirp- has joined #openstack | 20:49 | |
| soren | ?!? | 20:49 |
| eday | *sigh*, some tools, not all | 20:49 |
| soren | Exactly! | 20:49 |
| eday | n/m this is pointless | 20:49 |
| *** eday has left #openstack | 20:49 | |
| *** westmaas has quit IRC | 20:58 | |
| jk0 | kudos to whoever cleaned up the unit test output :) | 21:01 |
| *** dfg_ has joined #openstack | 21:01 | |
| jk0 | so, thank you termie ^ | 21:02 |
| *** dfg_ has quit IRC | 21:03 | |
| *** joearnold has joined #openstack | 21:06 | |
| *** jfluhmann has joined #openstack | 21:14 | |
| *** ctennis has quit IRC | 21:16 | |
| *** hggdh has quit IRC | 21:27 | |
| *** hggdh has joined #openstack | 21:27 | |
| *** ctennis has joined #openstack | 21:29 | |
| *** ctennis has joined #openstack | 21:29 | |
| *** Ryan_Lane|food is now known as Ryan_Lane | 21:32 | |
| openstackhudson | Project nova build #354: SUCCESS in 1 min 23 sec: http://hudson.openstack.org/job/nova/354/ | 21:34 |
| * soren goes to bed | 21:36 | |
| *** Daviey has quit IRC | 21:38 | |
| *** littleidea has quit IRC | 21:38 | |
| *** littleidea has joined #openstack | 21:39 | |
| *** Daviey has joined #openstack | 21:40 | |
| *** fabiand has quit IRC | 21:44 | |
| *** reldan has joined #openstack | 21:47 | |
| *** joearnold has quit IRC | 22:00 | |
| *** sassyn has joined #openstack | 22:04 | |
| *** ppetraki has quit IRC | 22:04 | |
| *** jdarcy has quit IRC | 22:09 | |
| *** arthurc has quit IRC | 22:10 | |
| *** _skrusty has quit IRC | 22:20 | |
| *** mfoemmel has quit IRC | 22:20 | |
| *** gholt has quit IRC | 22:20 | |
| *** mfoemmel has joined #openstack | 22:20 | |
| *** dwight_ has quit IRC | 22:21 | |
| *** gholt has joined #openstack | 22:21 | |
| *** allsystemsarego has quit IRC | 22:21 | |
| *** dwight_ has joined #openstack | 22:22 | |
| *** KnightHa1ker has quit IRC | 22:28 | |
| *** KnightHacker has joined #openstack | 22:28 | |
| *** _skrusty has joined #openstack | 22:32 | |
| *** nelson__ has quit IRC | 22:36 | |
| *** uvirtbot has quit IRC | 22:36 | |
| *** dovetaildan has quit IRC | 22:36 | |
| *** DeAnza has quit IRC | 22:36 | |
| *** spectorclan has quit IRC | 22:37 | |
| *** uvirtbot has joined #openstack | 22:41 | |
| *** dovetaildan has joined #openstack | 22:41 | |
| *** DeAnza has joined #openstack | 22:44 | |
| *** sassyn has quit IRC | 22:45 | |
| *** nelson__ has joined #openstack | 22:48 | |
| *** joearnold has joined #openstack | 22:50 | |
| annegentle | ok, re-request on a new proposal since I couldn't get conflicts resolved on t'other branch. https://code.launchpad.net/~annegentle/nova/fixnewscript/+merge/45085 | 23:03 |
| creiht | hehe | 23:03 |
| *** littleidea has quit IRC | 23:04 | |
| *** littleidea_ has joined #openstack | 23:06 | |
| vishy | holy scrollback batman | 23:09 |
| *** ccustine has joined #openstack | 23:10 | |
| vishy | is anyone still here | 23:10 |
| Ryan_Lane | yep :) | 23:12 |
| vishy | holy crap | 23:12 |
| Ryan_Lane | freenode has been having serious problems latelty | 23:12 |
| vishy | so eday proposed all of his branches | 23:15 |
| vishy | s/proposed/unproposed | 23:15 |
| creiht | vishy: I think he and soren had a little quarrel | 23:16 |
| *** gasbakid has quit IRC | 23:17 | |
| creiht | not sure if that has anything to do with it | 23:17 |
| creiht | but he seemed pretty ticked | 23:17 |
| *** KnightHacker has quit IRC | 23:21 | |
| *** KnightHacker has joined #openstack | 23:21 | |
| *** Cybo has joined #openstack | 23:24 | |
| uvirtbot | New bug: #697019 in nova "Recover from a lost data store connection" [Undecided,In progress] https://launchpad.net/bugs/697019 | 23:26 |
| vishy | yes apparently so | 23:29 |
| jk0 | vishy: would you mind looking at https://code.launchpad.net/~jk0/nova/lp688545/+merge/45080 once more for approval? | 23:31 |
| *** holoway_ has quit IRC | 23:33 | |
| *** holoway has joined #openstack | 23:33 | |
| vishy | soren, eday: we can rescue the euca tools pretty easily. We can require only enough of the UUID to disambiguate for an individual project. so people could just type out the first 6-8 characgers of the uuid instead of the whole thing. We could also allow people to use ami-<display_name> so they can nick the instances themselves. | 23:34 |
| *** dirakx1 has quit IRC | 23:34 | |
| jk0 | vishy: thanks dude | 23:35 |
| *** filler has quit IRC | 23:35 | |
| vishy | jk0: np | 23:35 |
| *** filler has joined #openstack | 23:35 | |
| vishy | did i miss anything else important during vacation. I lost some of the scrollback and i don't really want to go digging through the logs | 23:35 |
| openstackhudson | Project nova-tarmac build #49,307: FAILURE in 14 sec: http://hudson.openstack.org/job/nova-tarmac/49307/ | 23:37 |
| openstackhudson | Yippie, build fixed! | 23:46 |
| openstackhudson | Project nova-tarmac build #49,308: FIXED in 3 min 53 sec: http://hudson.openstack.org/job/nova-tarmac/49308/ | 23:46 |
| *** dubsquared1 has left #openstack | 23:47 | |
| openstackhudson | Project nova build #355: SUCCESS in 1 min 18 sec: http://hudson.openstack.org/job/nova/355/ | 23:49 |
| *** nelson__ has quit IRC | 23:58 | |
| *** nelson__ has joined #openstack | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!