| Jaesang | security group didn't apply flat network mode? | 00:00 |
|---|---|---|
| *** reed has joined #openstack | 00:03 | |
| *** rustam has quit IRC | 00:06 | |
| *** retr0h has quit IRC | 00:06 | |
| *** trent has joined #openstack | 00:07 | |
| *** v0id has quit IRC | 00:10 | |
| *** nerdstein has joined #openstack | 00:15 | |
| *** juddm has quit IRC | 00:19 | |
| uvirtbot | New bug: #860158 in nova "Release deb pkg for nova missing dependency on dbus" [Undecided,New] https://launchpad.net/bugs/860158 | 00:21 |
| *** osier has joined #openstack | 00:22 | |
| *** MarkAtwood has joined #openstack | 00:22 | |
| *** dragondm has quit IRC | 00:26 | |
| *** ziyadb has joined #openstack | 00:29 | |
| *** llang629_ has joined #openstack | 00:29 | |
| *** jollyxen has quit IRC | 00:34 | |
| hacim | i'm doing a test swift setup, and I've just installed and configured the proxy node and added storage devices to a node, then I created a second node and did the swift-ring-builder commands. but when I do a rebalance I get a python traceback | 00:34 |
| hacim | i've been working off of http://docs.openstack.org/diablo/openstack-object-storage/admin/content/installing-and-configuring-the-proxy-node.html | 00:35 |
| *** dirakx has joined #openstack | 00:36 | |
| *** vladimir3p has quit IRC | 00:37 | |
| *** dysinger has quit IRC | 00:38 | |
| *** cereal_bars has quit IRC | 00:38 | |
| *** llang629_ has left #openstack | 00:40 | |
| *** ohnoimdead has quit IRC | 00:43 | |
| *** robbiew has joined #openstack | 00:43 | |
| *** ziyadb has quit IRC | 00:50 | |
| *** clauden_ has joined #openstack | 00:53 | |
| *** dysinger has joined #openstack | 00:55 | |
| notmyname | hacim: so you only have 2 nodes in your ring? add a third, and the error should go away | 00:58 |
| *** reed has quit IRC | 01:00 | |
| *** mkinitrd has quit IRC | 01:02 | |
| *** woleium has quit IRC | 01:03 | |
| hacim | notmyname: ok. its not clear to me if I am doing the swift-ring-builder command on the proxy for all storage on every node? | 01:04 |
| notmyname | hacim: you can build the ring(s) offline wherever you want. you end up with one file per ring, and you can distribute it to all of your servers however you see fit. the ring building process does not need access to the cluster | 01:05 |
| hacim | notmyname: thanks! that clarifies a lot, i'm still trying to figure out if I should create multiple zones at this point or not, but I'll build up another node and see if that helps | 01:06 |
| notmyname | yes, you should | 01:06 |
| notmyname | hacim: you should have <number of replicas> as a bare minimum for your zone count | 01:07 |
| *** atkisc has joined #openstack | 01:07 | |
| notmyname | preferably, you should have a couple more to support handoff nodes | 01:07 |
| notmyname | so, with the default of 3 replicas, start with 5 zones | 01:07 |
| notmyname | (and a zone could be only one server) | 01:08 |
| notmyname | (or only one storage volume, really) | 01:08 |
| hacim | notmyname: i'm just trying to get a hang of things and test it out, so i'm trying to setup as few machines as possible | 01:09 |
| notmyname | hacim: have you looked at the all-in-one? | 01:09 |
| hacim | i have not, where can I find that? | 01:09 |
| notmyname | http://swift.openstack.org/development_saio.html | 01:10 |
| notmyname | hacim: I have to warn you about the swift-all-in-one (SAIO), though | 01:10 |
| notmyname | hacim: it's kinda like unix man pages. it's perfectly clear when you already know everything that's going on | 01:10 |
| hacim | hehe | 01:11 |
| notmyname | hacim: the point of the SAIO is to have a dev environment that simulates a cluster. so it fudges some things to make it appear that you actually have 4 drives in 4 zones | 01:11 |
| hacim | yeah, i was thinking if I set it up manually then I would understand it better | 01:11 |
| notmyname | hacim: the fact that the saio fundges some stuff (like using loopback devices for storage volumes and running multiple copies of a service on one machine) does make it a little unlike a real cluster, and that could be confusing | 01:13 |
| *** hadrian_ has joined #openstack | 01:13 | |
| notmyname | unfortunately, since I've been living and breathing swift for 2 years, I can't really tell you where the confusing parts are | 01:13 |
| hacim | understood! | 01:13 |
| notmyname | best to not be shy and hop in here and ask :-) | 01:13 |
| *** hadrian_ has quit IRC | 01:14 | |
| hacim | ok, i'll have some time to fire up another vm tomorrow and see what I can get, thanks again | 01:14 |
| *** ziyadb has joined #openstack | 01:14 | |
| * hacim sticks around here for the time being | 01:14 | |
| notmyname | hacim: will you by any chance be at the openstack summit/conference next week? | 01:14 |
| *** kaz has quit IRC | 01:14 | |
| *** stanchan has quit IRC | 01:15 | |
| hacim | notmyname: i just read about that - its in boston? | 01:16 |
| notmyname | yes | 01:16 |
| hacim | hm. i'm in nyc, its possible | 01:16 |
| *** kaz_ has joined #openstack | 01:16 | |
| notmyname | the summit (dev portion) is sold out and registration is closed, but the conference at the end of the week still has open seats, I think | 01:16 |
| *** hadrian has quit IRC | 01:16 | |
| hacim | i have hardly installed it or wrapped my head around things still, so a summit/conference might be way over my head | 01:16 |
| notmyname | ah. no worries. the bulk of the week will be devoted to deep dives into the code and designing new features | 01:17 |
| atkisc | \ | 01:17 |
| notmyname | the other important parts of the conference are on the business side of things for company collaboration | 01:18 |
| *** livemoon has joined #openstack | 01:22 | |
| Jaesang | Can I use security group when using Flat Networkmanager? | 01:23 |
| Jaesang | In Flatnetworkmanager, When I launched an instance, every traffic being sent. | 01:24 |
| Jaesang | In manual, Default is every traffic discard. | 01:24 |
| *** lborda has joined #openstack | 01:24 | |
| *** obino has quit IRC | 01:27 | |
| *** jdurgin has quit IRC | 01:29 | |
| *** arus7 has quit IRC | 01:31 | |
| *** aurigus has quit IRC | 01:33 | |
| *** aurigus has joined #openstack | 01:33 | |
| *** huslage has quit IRC | 01:34 | |
| *** robbiew has quit IRC | 01:35 | |
| *** jfluhmann has joined #openstack | 01:37 | |
| *** krow has joined #openstack | 01:37 | |
| livemoon | hi, what is meaning of "routing_source" in nova.conf? | 01:40 |
| baynaa | livemoon: check here http://wiki.openstack.org/FlagsGrouping | 01:41 |
| *** ziyadb has quit IRC | 01:44 | |
| *** woleium has joined #openstack | 01:45 | |
| *** ziyadb has joined #openstack | 01:45 | |
| livemoon | baynaa: thanks | 01:49 |
| *** adjohn has quit IRC | 01:50 | |
| *** ziyadb has quit IRC | 01:55 | |
| *** eallrich has joined #openstack | 02:03 | |
| *** krow has quit IRC | 02:04 | |
| *** miclorb_ has quit IRC | 02:05 | |
| *** dysinger has quit IRC | 02:06 | |
| *** odyi has quit IRC | 02:07 | |
| *** odyi has joined #openstack | 02:08 | |
| *** odyi has joined #openstack | 02:08 | |
| *** lborda has quit IRC | 02:12 | |
| *** nyeates has joined #openstack | 02:20 | |
| *** mkinitrd has joined #openstack | 02:23 | |
| *** joonwon has joined #openstack | 02:32 | |
| *** joshua_d has quit IRC | 02:32 | |
| *** dysinger has joined #openstack | 02:44 | |
| livemoon | Hi, I use nova-manage network delete command, but the entries are still exist in table fixed_ips in mysql. | 02:48 |
| livemoon | Does it need me to delete entries manually in mysql? | 02:49 |
| *** osier has quit IRC | 02:51 | |
| *** matiu has joined #openstack | 02:57 | |
| hacim | notmyname: well, i couldn't wait. i created another vm and added it, but I'm still getting the rebalance python traceback | 02:59 |
| *** stuntmachine has quit IRC | 03:00 | |
| hacim | but I might be confused because I only have one zone, but three machines. should I be putting those all into their own unique zone? | 03:00 |
| *** reed has joined #openstack | 03:01 | |
| *** woleium has quit IRC | 03:07 | |
| atkisc | hi ,everybody | 03:09 |
| atkisc | i running dashboard, error: | 03:09 |
| atkisc | Unable to get usage info: Invalid service catalog service: nova | 03:09 |
| *** joonwon has quit IRC | 03:09 | |
| *** dysinger has quit IRC | 03:14 | |
| *** stanchan has joined #openstack | 03:17 | |
| *** katkee has joined #openstack | 03:20 | |
| livemoon | atkisc: are you yz/ | 03:21 |
| *** ziyadb has joined #openstack | 03:24 | |
| *** fayce has joined #openstack | 03:27 | |
| *** fayce has quit IRC | 03:28 | |
| *** osier has joined #openstack | 03:32 | |
| *** OutBackDingo has quit IRC | 03:33 | |
| *** obino has joined #openstack | 03:34 | |
| *** krow has joined #openstack | 03:34 | |
| *** miclorb_ has joined #openstack | 03:35 | |
| *** krow has joined #openstack | 03:36 | |
| *** odyi has quit IRC | 03:36 | |
| *** odyi has joined #openstack | 03:37 | |
| *** odyi has joined #openstack | 03:37 | |
| *** ziyadb has quit IRC | 03:38 | |
| *** bdwarr61 has joined #openstack | 03:38 | |
| *** OutBackDingo has joined #openstack | 03:38 | |
| *** krow has quit IRC | 03:40 | |
| notmyname | hacim: yes. all three in different zones | 03:40 |
| *** PeteDaGuru has left #openstack | 03:48 | |
| *** wariola has joined #openstack | 03:57 | |
| *** mkinitrd has quit IRC | 03:57 | |
| *** joonwon has joined #openstack | 04:01 | |
| *** matiu has quit IRC | 04:06 | |
| *** jakedahn has quit IRC | 04:07 | |
| *** jtanner has quit IRC | 04:08 | |
| *** neogenix has quit IRC | 04:11 | |
| *** miclorb_ has quit IRC | 04:15 | |
| *** halfss has joined #openstack | 04:18 | |
| *** trent has quit IRC | 04:20 | |
| *** osier has quit IRC | 04:26 | |
| *** marrusl has quit IRC | 04:33 | |
| *** nyeates has quit IRC | 04:33 | |
| *** jakedahn has joined #openstack | 04:34 | |
| *** zul has quit IRC | 04:36 | |
| *** adjohn has joined #openstack | 04:37 | |
| *** zul has joined #openstack | 04:40 | |
| *** jakedahn has quit IRC | 04:43 | |
| *** wariola has quit IRC | 04:52 | |
| *** miclorb_ has joined #openstack | 04:52 | |
| *** stanchan has quit IRC | 04:58 | |
| *** jollyxen has joined #openstack | 05:07 | |
| *** joonwon has quit IRC | 05:07 | |
| *** ldleworker has joined #openstack | 05:12 | |
| *** jdag has quit IRC | 05:15 | |
| *** stanchan has joined #openstack | 05:24 | |
| livemoon | Hi,does anyone meet this problem https://bugs.launchpad.net/nova/+bug/855971 | 05:27 |
| *** jedi4ever has quit IRC | 05:28 | |
| *** jollyxen has quit IRC | 05:33 | |
| *** ldleworker has quit IRC | 05:39 | |
| *** rocambole has joined #openstack | 05:42 | |
| *** bdwarr61 has left #openstack | 05:45 | |
| *** atkisc has left #openstack | 05:50 | |
| *** jollyxen has joined #openstack | 06:07 | |
| winston-d | jaypipes : around? | 06:09 |
| *** jedi4ever has joined #openstack | 06:10 | |
| *** jedi4ever has joined #openstack | 06:11 | |
| *** nerens has joined #openstack | 06:14 | |
| *** osier has joined #openstack | 06:18 | |
| *** shaibn has joined #openstack | 06:24 | |
| *** openpercept has joined #openstack | 06:25 | |
| *** adjohn has quit IRC | 06:25 | |
| *** nerens has quit IRC | 06:29 | |
| *** fayce has joined #openstack | 06:30 | |
| *** kaigan_ has joined #openstack | 06:33 | |
| *** rocambole has quit IRC | 06:48 | |
| *** Razique has joined #openstack | 06:51 | |
| *** rocambol1 has joined #openstack | 06:53 | |
| *** larstobi has quit IRC | 06:55 | |
| *** wariola has joined #openstack | 06:57 | |
| Razique | Hi all Cloudies! | 06:57 |
| *** koolhead17 has joined #openstack | 07:00 | |
| *** mies_ has quit IRC | 07:00 | |
| *** mies has joined #openstack | 07:00 | |
| *** shang has quit IRC | 07:00 | |
| ohadlevy | Razique: sorry, only mist today | 07:00 |
| Razique | hehe | 07:00 |
| *** koolhead17 has quit IRC | 07:00 | |
| *** koolhead17 has joined #openstack | 07:00 | |
| Razique | From which country ? :) | 07:00 |
| *** Razique has left #openstack | 07:01 | |
| *** mandela123 has joined #openstack | 07:01 | |
| *** Razique has joined #openstack | 07:01 | |
| mandela123 | hi | 07:02 |
| *** mies has quit IRC | 07:03 | |
| mandela123 | i am confused about the config file of glance in [filter:authtoken] | 07:03 |
| *** shiny_ has joined #openstack | 07:03 | |
| *** mies has joined #openstack | 07:03 | |
| mandela123 | should it be [filter:tokenauth] | 07:03 |
| mandela123 | because in the example directory of keystone | 07:03 |
| mandela123 | the demo program of echo | 07:03 |
| mandela123 | in keystone | 07:03 |
| *** shang has joined #openstack | 07:04 | |
| mandela123 | the echo.ini config file | 07:04 |
| mandela123 | it shows [filter:tokenauth] | 07:04 |
| *** mies has quit IRC | 07:04 | |
| mandela123 | does it matter ? | 07:04 |
| *** mies has joined #openstack | 07:05 | |
| *** mies_ has joined #openstack | 07:06 | |
| *** mies has quit IRC | 07:06 | |
| *** martines has quit IRC | 07:06 | |
| Razique | HI, you are right, I've briefly looked at the doc, but according to the latest commits | 07:09 |
| Razique | it's filter:authtoken | 07:09 |
| *** reidrac has joined #openstack | 07:09 | |
| *** martines has joined #openstack | 07:09 | |
| *** miclorb_ has quit IRC | 07:10 | |
| *** mies_ has quit IRC | 07:11 | |
| *** mies has joined #openstack | 07:11 | |
| *** dirkx__ has joined #openstack | 07:19 | |
| mandela123 | you mean it does net matter whether it is authtoken or tokenauth | 07:23 |
| *** katkee has quit IRC | 07:24 | |
| mandela123 | also, is keystone now must install with glance and swift proxy in the same machine? | 07:25 |
| mandela123 | when service to be authed remotely, keystone use a password to varify the service | 07:27 |
| *** mandela123 has quit IRC | 07:27 | |
| *** mancdaz has joined #openstack | 07:29 | |
| *** nerens has joined #openstack | 07:30 | |
| *** adjohn has joined #openstack | 07:33 | |
| *** jakedahn has joined #openstack | 07:36 | |
| *** larstobi has joined #openstack | 07:40 | |
| Razique | mandela123 : It does matter actually, you should use the right flag ; regarding the three components, they can be installed separately | 07:44 |
| Razique | Dunno if you need Glance here if you plan to install swift :) | 07:45 |
| *** dobber has joined #openstack | 07:47 | |
| *** atkisc has joined #openstack | 07:47 | |
| atkisc | hi , everybody ? | 07:47 |
| Razique | hey | 07:49 |
| *** MarcMorata has joined #openstack | 07:50 | |
| *** larstobi has quit IRC | 07:53 | |
| *** foexle has joined #openstack | 07:56 | |
| *** swift82 has joined #openstack | 07:58 | |
| *** javiF has joined #openstack | 07:58 | |
| *** javiF has quit IRC | 08:03 | |
| *** FabriceB has joined #openstack | 08:04 | |
| *** javiF has joined #openstack | 08:16 | |
| livemoon | Hi, I got "ndb15: Receive control failed (result -32)" error when I start an instance in compute node. | 08:16 |
| livemoon | Then how to reslove it. | 08:17 |
| *** osier has quit IRC | 08:19 | |
| *** newsbad_com has quit IRC | 08:21 | |
| *** newsbad_com has joined #openstack | 08:22 | |
| *** nacx has joined #openstack | 08:25 | |
| *** reidrac has quit IRC | 08:25 | |
| *** alekibango_ has joined #openstack | 08:26 | |
| Razique | livemoon: hi, I already seen that error, it's related as far as I recall to the DHCP server | 08:28 |
| Razique | what nova-manage service list shows ? | 08:28 |
| *** reidrac has joined #openstack | 08:30 | |
| *** alekibango_ is now known as alekibango | 08:31 | |
| *** fayce has quit IRC | 08:34 | |
| *** reidrac has quit IRC | 08:35 | |
| *** reidrac has joined #openstack | 08:35 | |
| *** osier has joined #openstack | 08:37 | |
| *** koolhead17 has quit IRC | 08:37 | |
| *** katkee has joined #openstack | 08:41 | |
| *** dirkx__ is now known as dirkx | 08:41 | |
| *** adjohn has quit IRC | 08:44 | |
| *** wariola has quit IRC | 08:45 | |
| *** osier has quit IRC | 08:45 | |
| *** wariola has joined #openstack | 08:46 | |
| *** rustam has joined #openstack | 08:47 | |
| *** dirkx has quit IRC | 08:47 | |
| *** Gerr1t has joined #openstack | 08:48 | |
| Gerr1t | morning! | 08:50 |
| *** darraghb has joined #openstack | 08:50 | |
| Gerr1t | quick question, we are planning on implementing the openstack platform for our cloud based solution.. | 08:51 |
| Gerr1t | we are looking into keystone for authentication, unfortunately there is not much documentation about it yet | 08:51 |
| *** miclorb_ has joined #openstack | 08:51 | |
| *** tryggvil_ has joined #openstack | 08:52 | |
| Gerr1t | the reason we want to use Keystone is because we would like to have a centralized authentication server that does auth for swift, glance and nova.. | 08:52 |
| Gerr1t | does anyone know a decent tutorial on how to implement it into all layers? | 08:52 |
| *** osier has joined #openstack | 08:54 | |
| *** javiF has quit IRC | 09:00 | |
| *** blueblade has joined #openstack | 09:03 | |
| *** stanchan has quit IRC | 09:07 | |
| uvirtbot | New bug: #860377 in openstack-qa "nova.rpc.impl_carrot" [Undecided,New] https://launchpad.net/bugs/860377 | 09:12 |
| *** po has joined #openstack | 09:13 | |
| Razique | Gerr1t: here http://docs.openstack.org/incubation/identity-dev-guide/identity-dev-guide.pdf | 09:14 |
| Gerr1t | Razique: thanks a bunch! | 09:14 |
| Razique | you welcome :) | 09:15 |
| *** nicolas2b has joined #openstack | 09:15 | |
| *** jollyxen has quit IRC | 09:16 | |
| *** irahgel has joined #openstack | 09:20 | |
| uvirtbot | New bug: #860386 in openstack-qa "nova.virt.driver" [Undecided,New] https://launchpad.net/bugs/860386 | 09:21 |
| atkisc | Razique: diablo release keystone and dashboard, good ? | 09:22 |
| Razique | atkisc: you asking me if I use these ? | 09:22 |
| *** wariola has quit IRC | 09:25 | |
| *** javiF has joined #openstack | 09:27 | |
| *** dirkx_ has joined #openstack | 09:27 | |
| *** wariola has joined #openstack | 09:28 | |
| *** mies has quit IRC | 09:34 | |
| *** mies has joined #openstack | 09:35 | |
| Gerr1t | Razique: we are also planning on doing 2 glance servers which will be loadbalanced and high available.. we are planning on using a mysql back-end, in that way we can master-master replicate the databases between the two glance nodes | 09:37 |
| Gerr1t | do you know if that is possible? | 09:37 |
| Gerr1t | and what happens with nova when he his connection is switched to another server (because of a failover, or whatever :) ) | 09:38 |
| Razique | Gerr1t: here is what I would do | 09:39 |
| Razique | install two Glance server, and use mysql replication, also a repository replication | 09:39 |
| Razique | regarding the ip : keep alive + heartbeat | 09:39 |
| Gerr1t | repository replication?.. what exactly do you mean?.. | 09:40 |
| Razique | so the two instances have a private IP, while keepalive will dynamicaly assign the "public ip" (the one nova will connect to) to one of the two glance servers | 09:40 |
| Razique | well actually the data is stored by default in /var/lib/glance/ | 09:40 |
| Gerr1t | ah , that is what you mean.. | 09:40 |
| Razique | so you need to make sure that dir. is consitent | 09:40 |
| Razique | you could share that repository also, but NFS perf are somehow bad | 09:41 |
| Gerr1t | in your example there is no loadbalancing, we are worried that active-passive setup will be a bottleneck when you have multiple nova nodes talking to the glance server.. | 09:41 |
| Razique | so, maybe another way of sharing (if you want to share the repo) | 09:41 |
| Razique | you could use haproxy then :) | 09:42 |
| Razique | haproxy does load-balancing based on algorithms you define | 09:42 |
| *** koolhead17 has joined #openstack | 09:42 | |
| Gerr1t | :).. LVS + Keepalive we are thinking about, because LVS does Tunneling, which means the loadbalancer does not handle returned connection.. | 09:42 |
| Razique | so you can have two instances of glance, with two privates IP, the haproxy server handles the "public ip" | 09:42 |
| Gerr1t | yeah, that is what we were thinking about :).. I will make a note about the repository .. will look into it :) | 09:43 |
| Razique | you could also yes, I found haproxy nice to administer, plus it has a nice status page | 09:43 |
| Gerr1t | thanks a bunch, again ;) | 09:43 |
| Razique | you welcome :) | 09:43 |
| *** javiF has quit IRC | 09:46 | |
| Razique | still, pay attention that the bottleneck, instead of waiting it from nova to glance, doesn't appear between the several glance's instances themselves | 09:47 |
| Razique | waiting -> expecting* | 09:47 |
| Razique | that would make the cluster even more useless | 09:47 |
| Gerr1t | sorry I dont understand that sentence.. | 09:47 |
| Gerr1t | you dont expect that nova->glance is not a bottleneck | 09:48 |
| Gerr1t | ? | 09:48 |
| Razique | mm I thought you wanted avoid any performances issues when nova gets images, so you want to use several glance instances | 09:48 |
| Gerr1t | yes , that is exactly what I mean :) | 09:49 |
| Razique | ok, so, suppose the Glance cluster will allows you to prevent any perf issues here, these issues could happen between the glance servers themselves if you plan to use a shared repository | 09:50 |
| Razique | SHARED-REPO <----- /// PERF ISSUES ----- > GLANCE 1/2/XXX | 09:51 |
| Gerr1t | ahhh, ok, yes you are right.. you mean that the disk I/O would become a bottleneck when using the same repo | 09:51 |
| Razique | could be the I/O, the network, the protocol used :) | 09:51 |
| Gerr1t | that's why we are using a multinode swift back-end with 2 proxy's, 3 zones with 2 8TB nodes.. | 09:51 |
| Razique | in fact any involved layer at that point has to be studied carefully, if you don't want to ruin your cluster perfs :) | 09:52 |
| Razique | hehe seems great, how many instances are we talking about here ? | 09:52 |
| Gerr1t | you are right :).. but that is something we will have to find out when we go live :).. | 09:52 |
| *** javiF has joined #openstack | 09:53 | |
| Razique | I did for a customer a study, not related at all to nova, but here are the elements I found important to consider : VLAN cause an important overhead in perf ; Network configuration, and disks | 09:54 |
| Gerr1t | so basically its 4 nova machines running XEN VM's, connecting to Glance (2 nodes loadbalances) for their data, glance fetches en posts it on the swift cluster, which are 2 proxy's with 3 storage zones with 2 nodes of 2TB each.. | 09:54 |
| *** fayce has joined #openstack | 09:54 | |
| *** pixelbeat has joined #openstack | 09:55 | |
| Razique | what is the conf. for the nova-machines ? | 09:55 |
| Gerr1t | yeah,, im actually trying to convince my managers to use 10gig NIC's and several 10 GB switches to make sure that network connnectivity is not an issue :) | 09:55 |
| Razique | from my (tiny) experience the network is never the first issue, I mean you quickly see if it saturates | 09:55 |
| uvirtbot | New bug: #860401 in openstack-qa "nova.compute.api" [Undecided,New] https://launchpad.net/bugs/860401 | 09:55 |
| Razique | the applications involved and extra features on some low-layers equipments are sometimes more responsible for perf. issues | 09:56 |
| Gerr1t | ok | 09:57 |
| Razique | for the same customer, his perf issues were … the CPUs | 09:57 |
| Razique | :D | 09:57 |
| Gerr1t | ah ;) | 09:57 |
| Razique | neither the virtualization, nor the firewall, and even less the network | 09:57 |
| Razique | that's why u need to be carefull | 09:58 |
| Razique | lucy time :p | 09:58 |
| Gerr1t | actually, that is nother point of "finding out".. im having a hard time trying to sort out what kind of hardware to use for the different nodes, swift storage, swift proxy, glance (api+reg).. | 09:58 |
| Razique | lunch | 09:58 |
| Razique | Intel CPUs :D | 09:58 |
| *** tryggvil_ has quit IRC | 09:58 | |
| Razique | be back in 30 mn s :) | 09:58 |
| Gerr1t | :D | 09:59 |
| Gerr1t | same here | 09:59 |
| *** halfss has quit IRC | 10:10 | |
| *** Razique has quit IRC | 10:13 | |
| livemoon | q | 10:15 |
| *** livemoon has quit IRC | 10:15 | |
| *** tryggvil_ has joined #openstack | 10:15 | |
| koolhead17 | how do we attach set of floating IP to a particular project? | 10:18 |
| *** nerdstein has quit IRC | 10:19 | |
| hugokuo | koolhead17 , hi | 10:21 |
| hugokuo | A techinical question about global system | 10:21 |
| koolhead17 | hugokuo: i need answer of my question first | 10:22 |
| *** shang has quit IRC | 10:27 | |
| *** wariola has quit IRC | 10:28 | |
| hugokuo | koolhead17 , thinking | 10:31 |
| hugokuo | nova-manage floating create ...... is there any new params for specify project ? | 10:33 |
| hugokuo | Question : does any one knows the work flow of nova-compute retrive a image from Swift ? does Swift send data to nova-compute directly or through Galcne ? | 10:35 |
| *** dirkx_ has quit IRC | 10:39 | |
| uvirtbot | New bug: #860437 in openstack-qa "nova.compute.instance_types" [Undecided,New] https://launchpad.net/bugs/860437 | 10:41 |
| koolhead17 | hugokuo: ./ | 10:44 |
| hugokuo | koolhead17 : :/ as I know | 10:44 |
| koolhead17 | hugokuo: ./00\. | 10:45 |
| hugokuo | while you setup floating_ip pools , all project can see that :< | 10:45 |
| *** katkee has quit IRC | 10:46 | |
| *** shang has joined #openstack | 10:48 | |
| *** nijaba has quit IRC | 10:51 | |
| *** nijaba has joined #openstack | 10:51 | |
| *** nijaba has joined #openstack | 10:51 | |
| *** javiF has quit IRC | 10:51 | |
| *** marrusl has joined #openstack | 10:52 | |
| *** dirkx_ has joined #openstack | 10:53 | |
| *** rods has joined #openstack | 10:54 | |
| Gerr1t | hugokuo: glance builds the image for nova from fragments found throughout the swift cluster | 10:55 |
| uvirtbot | New bug: #860445 in openstack-qa "nova.volume.api" [Undecided,New] https://launchpad.net/bugs/860445 | 10:56 |
| uvirtbot | New bug: #860447 in openstack-qa "nova.volume.driver" [Undecided,New] https://launchpad.net/bugs/860447 | 10:56 |
| *** Binbin has joined #openstack | 10:57 | |
| *** osier has quit IRC | 10:58 | |
| *** marrusl has quit IRC | 10:58 | |
| hugokuo | Gerr1t , so that means the data traffic must go through Glance ? | 10:59 |
| Gerr1t | exactly, glance writes the data files in pre-set batches to swift.. | 11:00 |
| hugokuo | !!! | 11:00 |
| openstack | hugokuo: Error: "!!" is not a valid command. | 11:00 |
| *** marrusl has joined #openstack | 11:01 | |
| hugokuo | gerr1t , thanks .... | 11:01 |
| uvirtbot | New bug: #860452 in openstack-qa "nova.volume.manager" [Undecided,New] https://launchpad.net/bugs/860452 | 11:01 |
| uvirtbot | New bug: #860453 in openstack-qa "nova.volume.volume_types" [Undecided,New] https://launchpad.net/bugs/860453 | 11:01 |
| hugokuo | so the workflow is ......................nova-compute ----> Glance ------>swift ----->image fragments ---->glance ----->nova-compute | 11:02 |
| Gerr1t | nova -> Glance -> Swift proxy -> Swift storage -> Swift proxy -> Glance -> Nova | 11:03 |
| hugokuo | thanks it's more detail XD , you r so kind | 11:03 |
| Gerr1t | no problem. | 11:03 |
| hugokuo | oh , do you ever try twin swift proxy with Loadbalancer ? | 11:04 |
| hugokuo | I found that Pound could only handle 600 connections :< | 11:04 |
| hugokuo | that's weakness | 11:04 |
| Gerr1t | yeah, we are planning to use a loadbalancer for our 2 proxy servers.. | 11:05 |
| Gerr1t | gonna use LVS-TUN and Keepalived for loadbalancing | 11:06 |
| atkisc | hi | 11:07 |
| *** javiF has joined #openstack | 11:07 | |
| atkisc | my dashboard error : | 11:07 |
| atkisc | Unable to get usage info: Invalid service catalog service: nova | 11:08 |
| viraptor | is there somewhere a complete list of configuration flags changes between diablo releases? | 11:08 |
| atkisc | For example ? | 11:09 |
| *** shang has quit IRC | 11:09 | |
| hugokuo | Gerr1t , thanks .............. | 11:10 |
| Gerr1t | hugokuo: no problem | 11:10 |
| *** Vek has quit IRC | 11:13 | |
| *** darraghb has quit IRC | 11:14 | |
| *** mkinitrd has joined #openstack | 11:14 | |
| Jaesang | Anyone success deploy nova flatDHCP using Xen?? | 11:14 |
| *** valeriy has joined #openstack | 11:15 | |
| FabriceB | How can I terminate invalid instance ? | 11:16 |
| FabriceB | I have some instance with wrong network configuration | 11:16 |
| FabriceB | euca-terminate-instances fail with a python stack | 11:17 |
| koolhead17 | FabriceB: please check the log. i think you might have missed some parameter | 11:17 |
| *** valeriy is now known as valeriy2 | 11:17 | |
| koolhead17 | valeriy2: hey | 11:18 |
| valeriy2 | hey | 11:18 |
| FabriceB | teh configured network for this instance is destroy | 11:18 |
| FabriceB | so i'm getting some TypeError: 'NoneType' object is unsubscriptable | 11:18 |
| *** shang has joined #openstack | 11:19 | |
| koolhead17 | FabriceB: so the instance has nomore network associated with it? how will you communicate then | 11:19 |
| FabriceB | I want a force delete | 11:19 |
| FabriceB | there are even not started | 11:19 |
| *** nerdstein has joined #openstack | 11:19 | |
| *** po has quit IRC | 11:20 | |
| koolhead17 | FabriceB: /var/lib/instances see if somewhere like that you have instace ID | 11:20 |
| koolhead17 | and other details remove it | 11:20 |
| koolhead17 | i have no other idea about forced delete. | 11:21 |
| koolhead17 | remove | 11:21 |
| FabriceB | there are even not in /var/lib/nova/instances | 11:21 |
| FabriceB | they are kept somewhere | 11:21 |
| *** darraghb has joined #openstack | 11:22 | |
| *** chemikadze has quit IRC | 11:27 | |
| *** chemikadze has joined #openstack | 11:30 | |
| *** miclorb_ has quit IRC | 11:32 | |
| *** zul has quit IRC | 11:36 | |
| *** zul has joined #openstack | 11:36 | |
| *** zul has quit IRC | 11:36 | |
| *** tangyaguang has joined #openstack | 11:40 | |
| *** MarkAtwood has quit IRC | 11:41 | |
| uvirtbot | New bug: #860473 in nova "Documentation not consistent for the authentication options" [Undecided,New] https://launchpad.net/bugs/860473 | 11:43 |
| *** atkisc has left #openstack | 11:45 | |
| *** ivoks has quit IRC | 11:46 | |
| *** Binbin has quit IRC | 11:48 | |
| *** Binbin has joined #openstack | 11:48 | |
| *** livemoon has joined #openstack | 11:48 | |
| alexn65 | Hi all! Cannot figure where is the problem with internet connect from VMs. Could it be switch/router problem? Nat table has two rules for IN and OUT real_IP_to_Virt_IP if I allocate real address, but possibly it change nothing. | 11:53 |
| *** jtanner has joined #openstack | 11:54 | |
| livemoon | Hi, VM cannot started on node but can on the controller | 11:54 |
| *** openpercept has quit IRC | 11:55 | |
| livemoon | vm cannot be ping but not start it os. It stop the BIOS status. | 11:55 |
| alexn65 | livemoon: what says /var/lib/nova/instances/_yourInstance/console.log? just try to look at it | 11:56 |
| alexn65 | livemoon: on second node of course | 11:57 |
| *** mfer has joined #openstack | 11:58 | |
| livemoon | nothing ,only one line about BOIS information | 12:00 |
| *** hugokuo has quit IRC | 12:01 | |
| livemoon | I think it cannot be get ip. | 12:01 |
| *** mkinitrd has quit IRC | 12:02 | |
| *** lts has joined #openstack | 12:03 | |
| alexn65 | livemoon: than what errors in /var/log/nova-compute.log on second node? | 12:03 |
| livemoon | alexn65: Do you meet it? | 12:03 |
| *** ahasenack has joined #openstack | 12:05 | |
| *** helfrez has joined #openstack | 12:08 | |
| *** PotHix has joined #openstack | 12:10 | |
| livemoon | no error | 12:10 |
| livemoon | alexn65: no error | 12:10 |
| alexn65 | livemoon: not really, but know some cases. Than I don` t know your, sorry | 12:11 |
| livemoon | alexn65: in nova-network.log, there is ndb15 error, maybe just type | 12:11 |
| *** javiF has quit IRC | 12:12 | |
| tangyaguang | 很多中国的啊 | 12:12 |
| *** Razique_ has joined #openstack | 12:13 | |
| livemoon | tanyaguang: yes, chinese | 12:13 |
| *** katkee has joined #openstack | 12:15 | |
| alexn65 | livemoon: cases are - 1)not enough resources or arch on second node. 2) error while image downloading to node cache(check size in /var/lib/nova/instance/_baze or remove all in it); 3)kvm-pxe module needed; 4) network appears much later | 12:15 |
| *** openpercept has joined #openstack | 12:15 | |
| alexn65 | livemoon: I mean not proper arch | 12:15 |
| livemoon | alexn65: thanks. I will reinstall node and test it again | 12:16 |
| *** PiotrSikora has quit IRC | 12:20 | |
| *** hggdh has quit IRC | 12:20 | |
| alexn65 | livemoon: for nothing :) But if problem is in cache - it will probably remain cause these dirs will not be removed | 12:21 |
| *** PiotrSikora has joined #openstack | 12:21 | |
| *** javiF has joined #openstack | 12:24 | |
| *** ahasenack has quit IRC | 12:24 | |
| *** ahasenack has joined #openstack | 12:25 | |
| *** kernelfreak has joined #openstack | 12:26 | |
| livemoon | alexn65:tomorrow I will try again and hope to get result | 12:29 |
| *** dirkx_ has quit IRC | 12:29 | |
| *** hggdh has joined #openstack | 12:31 | |
| uvirtbot | New bug: #860516 in keystone "keystone nova integration steps in README need updating" [Undecided,New] https://launchpad.net/bugs/860516 | 12:31 |
| *** brendan-maguire has joined #openstack | 12:31 | |
| *** brendan-maguire has left #openstack | 12:32 | |
| livemoon | uvirtbot: Hi, have you tested the dashboard? | 12:33 |
| uvirtbot | livemoon: Error: "Hi," is not a valid command. | 12:33 |
| *** jtanner has quit IRC | 12:34 | |
| *** jtanner has joined #openstack | 12:34 | |
| *** lorin1 has joined #openstack | 12:35 | |
| livemoon | uvirtbot: :) | 12:37 |
| uvirtbot | livemoon: Error: ":)" is not a valid command. | 12:37 |
| *** AD\denman has joined #openstack | 12:41 | |
| uvirtbot | New bug: #860520 in nova "image snapshots broken (disk config change causes glance plugin AttributeError)" [High,In progress] https://launchpad.net/bugs/860520 | 12:42 |
| AD\denman | Hi, all. Has anyone had any luck getting RBD volumes working with nova-volumes? I can get the volume create, but when I attempt to attach the volume to the instance, I get an error that says: "libvirtError: cannot resolve symlink rbd/volume-00000001: No such file or directory" in the nova-compute logs | 12:42 |
| *** javiF has quit IRC | 12:42 | |
| *** atkisc has joined #openstack | 12:44 | |
| *** AD\denman is now known as enmand | 12:44 | |
| *** dirkx_ has joined #openstack | 12:44 | |
| *** marrusl has quit IRC | 12:45 | |
| *** livemoon has quit IRC | 12:45 | |
| *** bsza has joined #openstack | 12:45 | |
| Jaesang | Hi, | 12:46 |
| *** livemoon has joined #openstack | 12:46 | |
| Jaesang | I launched an instance- (in flatdhcp mode) | 12:46 |
| *** PeteDaGuru has joined #openstack | 12:46 | |
| *** lborda has joined #openstack | 12:47 | |
| Jaesang | But dhcp server can't send IP address to instance | 12:47 |
| Jaesang | What do I check ? | 12:49 |
| viraptor | has anyone tried to recompile the github version into a package? my build is failing the authors check (dan@nicira.com missing) | 12:49 |
| *** rustam has quit IRC | 12:50 | |
| *** eallrich has quit IRC | 12:51 | |
| *** Razique_ has quit IRC | 12:52 | |
| *** pmyers has joined #openstack | 12:52 | |
| *** pmyers has joined #openstack | 12:53 | |
| alexn65 | Jaesang: check dnsmask processes on compute node and connectivity for network, mb --network_host on others compute nodes | 12:56 |
| *** ivoks has joined #openstack | 12:59 | |
| *** Tsel has joined #openstack | 13:01 | |
| *** nelson____ has quit IRC | 13:03 | |
| *** nelson____ has joined #openstack | 13:03 | |
| *** lborda has quit IRC | 13:05 | |
| *** osier has joined #openstack | 13:05 | |
| *** fayce has quit IRC | 13:06 | |
| *** dgags has quit IRC | 13:07 | |
| atkisc | hi, are you running dashboard ,everybody ? | 13:07 |
| *** msivanes has joined #openstack | 13:07 | |
| *** ziyadb_ has joined #openstack | 13:08 | |
| *** Vek has joined #openstack | 13:09 | |
| *** brodrigue has quit IRC | 13:09 | |
| *** Razique has joined #openstack | 13:10 | |
| *** hggdh has quit IRC | 13:10 | |
| livemoon | atkisc:yz,have you used dashboard? | 13:16 |
| *** hggdh has joined #openstack | 13:17 | |
| livemoon | Jaesang: maybe you meet the similar situation with me. | 13:17 |
| *** huslage has joined #openstack | 13:18 | |
| atkisc | no | 13:18 |
| livemoon | atkisc: go on. | 13:19 |
| atkisc | ? | 13:20 |
| livemoon | atkisc: ji xu ,jia you | 13:20 |
| atkisc | 不搞了 | 13:21 |
| *** aa has quit IRC | 13:21 | |
| Jaesang | livemoon: How can you fix that? | 13:21 |
| atkisc | 问也没人解答 | 13:21 |
| jaypipes | winston-d: I'm around now... | 13:22 |
| atkisc | are you running dashboard ? | 13:22 |
| alexn65 | livemoon: no, Jaesang has different, cause you even don`t get to stage when network interfaces upping | 13:22 |
| *** tangyaguang has quit IRC | 13:23 | |
| enmand | Arg | 13:23 |
| * enmand kicks RBDDriver | 13:23 | |
| *** aa has joined #openstack | 13:24 | |
| *** webx has quit IRC | 13:25 | |
| *** imsplitbit has joined #openstack | 13:26 | |
| *** hadrian has joined #openstack | 13:27 | |
| *** ccorrigan has quit IRC | 13:32 | |
| livemoon | Jaesang: you can get ip? | 13:32 |
| livemoon | Jaesang: can you get ip? | 13:33 |
| Jaesang | livemoon: not yet | 13:33 |
| livemoon | atkisc: I can not get ip too.I only can ping it. | 13:34 |
| livemoon | alexn65: I can not get ip too.I only can ping it. | 13:34 |
| *** bcwaldon has joined #openstack | 13:35 | |
| *** doude has joined #openstack | 13:35 | |
| *** rsampaio has joined #openstack | 13:37 | |
| livemoon | quit | 13:40 |
| *** livemoon has quit IRC | 13:40 | |
| *** Jaesang has quit IRC | 13:40 | |
| *** koolhead17 has quit IRC | 13:49 | |
| *** valeriy2 has quit IRC | 13:50 | |
| *** notaj has quit IRC | 13:50 | |
| *** truijllo has joined #openstack | 13:51 | |
| truijllo | hi guys | 13:51 |
| *** kbringard has joined #openstack | 13:51 | |
| *** PiotrSikora has quit IRC | 13:51 | |
| truijllo | I'm trying to have a working nova environment to study | 13:51 |
| truijllo | I've installed nova+glance , official diablo releases | 13:51 |
| truijllo | and all seem to work properly but.... | 13:52 |
| truijllo | if I try to use keystone ( release https://github.com/openstack/keystone/commit/0425fba560e7d68e52f922667972765e64ad17dc like suggested in last meeting ) | 13:52 |
| *** alexn65 has quit IRC | 13:53 | |
| truijllo | running nova-api I obtain ImportError: <module 'nova.api.ec2' from '/usr/lib/python2.6/site-packages/nova/api/ec2/__init__.pyc'> has no 'ToToken' attribute | 13:53 |
| truijllo | any suggestions ? :-/ | 13:53 |
| enmand | truijllo: I believe there is a nova-api.paste file with Keystone, did you copy that to the correct location (possibly /etc/nova)? | 13:53 |
| *** PiotrSikora has joined #openstack | 13:53 | |
| truijllo | yes, I did | 13:54 |
| enmand | And completely restarted OpenStack? | 13:54 |
| truijllo | yes | 13:54 |
| truijllo | I use for n in api compute network objectstore scheduler vncproxy direct-api ; do sudo service openstack-nova-$n start; done | 13:55 |
| truijllo | and only API Server fails :-/ | 13:55 |
| *** Fredric has joined #openstack | 13:56 | |
| *** joesavak has joined #openstack | 13:56 | |
| *** zul has joined #openstack | 13:57 | |
| *** katkee has quit IRC | 13:57 | |
| kbringard | I personally like for i in $( ls /etc/init.d/nova-* | awk -F/ '{ print $NF }' ); do service $i restart; done | 13:59 |
| atkisc | . | 13:59 |
| *** nyeates has joined #openstack | 14:00 | |
| *** robbiew has joined #openstack | 14:00 | |
| FabriceB | f or in (cd /etc/init.d && echo nova-*) | 14:01 |
| *** benrodrigue has joined #openstack | 14:02 | |
| *** gnu111 has joined #openstack | 14:03 | |
| *** bourke has quit IRC | 14:03 | |
| *** robbiew has quit IRC | 14:04 | |
| *** PerlStalker has joined #openstack | 14:04 | |
| *** bourke has joined #openstack | 14:05 | |
| *** dirkx_ has quit IRC | 14:05 | |
| truijllo | of course, this is only a bash matter | 14:05 |
| *** martine has joined #openstack | 14:07 | |
| *** livemoon has joined #openstack | 14:08 | |
| *** ldlework has joined #openstack | 14:08 | |
| *** martine has quit IRC | 14:11 | |
| *** jdag has joined #openstack | 14:12 | |
| *** bcwaldon has quit IRC | 14:12 | |
| *** jfluhmann has quit IRC | 14:14 | |
| *** jsavak has joined #openstack | 14:14 | |
| *** rsampaio has quit IRC | 14:15 | |
| *** joesavak has quit IRC | 14:16 | |
| *** tryggvil_ has quit IRC | 14:20 | |
| viraptor | ttx: do you have some pointer at the commit used to build the diablo final package? | 14:22 |
| ttx | viraptor: which project ? | 14:23 |
| viraptor | for nova | 14:23 |
| viraptor | it seems like for example the tagged version doesn't pass the tests, because of .mailmap typo | 14:23 |
| *** javiF has joined #openstack | 14:24 | |
| ttx | It was built from http://bazaar.launchpad.net/~hudson-openstack/nova/milestone-proposed/revision/1201 | 14:24 |
| viraptor | I'm confused... how could it pass the tests correctly with issue in http://bazaar.launchpad.net/~hudson-openstack/nova/milestone-proposed/view/head:/.mailmap#L18 (missing <>) - when the ppa buildlog shows test_authors_up_to_date passing? | 14:25 |
| viraptor | ahhh.... unless the ppa build environment didn't include .git - is that right? | 14:26 |
| ttx | viraptor: yes, it starts from the tarball | 14:26 |
| *** hugokuo has joined #openstack | 14:28 | |
| *** mattray has joined #openstack | 14:29 | |
| *** scalability-junk has joined #openstack | 14:29 | |
| *** cp16net has joined #openstack | 14:29 | |
| viraptor | cool, thanks | 14:30 |
| scalability-junk | hey I just wanted to check if there are any resources on how to setup openstack in a small scale environment? or is this not possible? let's say to check out and run sort of production with 2 dedicated servers? | 14:30 |
| *** mattray has quit IRC | 14:30 | |
| *** mattray has joined #openstack | 14:30 | |
| *** rnirmal has joined #openstack | 14:31 | |
| *** LiamMac has joined #openstack | 14:32 | |
| *** atkisc has left #openstack | 14:34 | |
| *** robbiew has joined #openstack | 14:35 | |
| *** dgags has joined #openstack | 14:35 | |
| *** code_franco has joined #openstack | 14:35 | |
| *** Tsel has quit IRC | 14:36 | |
| livemoon | Hi, Using FLATDHCP, br100 ip is 192.168.200.20, when I start nova-network, nova always sets br100 to 192.168.200.1. Why does it happen? Gateway? | 14:37 |
| livemoon | And I see "gateway" field in table "networks" in mysql seted 192.168.200.1. What's meaning of the gateway here | 14:38 |
| *** osier has quit IRC | 14:43 | |
| *** shiny_ has quit IRC | 14:43 | |
| *** tsel has joined #openstack | 14:43 | |
| *** ldlework has quit IRC | 14:44 | |
| gnu111 | livemoon: did you set the br100 ip to 192.168.200.20 manually? nova-network acts as a gateway for the vms: http://docs.openstack.org/diablo/openstack-compute/admin/content/configuring-flat-dhcp-networking.html | 14:44 |
| livemoon | gnu111:yes, I set br100 in /etc/network/interfaces | 14:46 |
| *** nerens has quit IRC | 14:47 | |
| gnu111 | what is your --flat_interface in nova.conf? | 14:47 |
| livemoon | gnu111: --flat_interface=eth0 | 14:48 |
| kbringard | you'll also want to make sure that gateway is set properly in the networks table in the DB | 14:48 |
| kbringard | like you mentioned above | 14:48 |
| livemoon | then I manually set gateway in mysql. | 14:49 |
| kbringard | that table gets populated when you do nova-manage network create | 14:49 |
| *** Ryan_Lane has joined #openstack | 14:51 | |
| livemoon | nova-network will set my br100 to the gateway which has been set in db | 14:51 |
| kbringard | yep, because nova-network acts as the default gateway for the defined networks | 14:52 |
| livemoon | which ip can I set ? host ip as gateway? | 14:53 |
| kbringard | well, it's an internal IP | 14:53 |
| kbringard | it depends on how your network is setup | 14:53 |
| kbringard | but, think of it like how you have a router on your internet at hme | 14:54 |
| kbringard | you have your WAN IP and your LAN IP | 14:54 |
| kbringard | if you ifconfig on your laptop, the default gateway is going to be 10.0.0.1 or whatever | 14:54 |
| kbringard | which is an IP on the router | 14:54 |
| kbringard | it can be whatever you want it to be | 14:54 |
| *** jfluhmann has joined #openstack | 14:55 | |
| kbringard | so long as it's one hop away from the machine you want to use | 14:55 |
| livemoon | but there are two same ip in my network. | 14:55 |
| livemoon | one is my route. the other is host br100. | 14:56 |
| *** tryggvil has joined #openstack | 14:56 | |
| kbringard | right, so you need to change the br100 to be something else | 14:56 |
| kbringard | because all the VMs are going to use that as their default route | 14:57 |
| *** adjohn has joined #openstack | 14:57 | |
| *** ahasenack has quit IRC | 14:57 | |
| *** ldlework has joined #openstack | 14:58 | |
| *** kaigan_ has quit IRC | 14:58 | |
| livemoon | thanks, I will think it more | 14:58 |
| *** katkee has joined #openstack | 14:58 | |
| kbringard | be default it goes VM -> br100 -> eth0 -> internet | 14:58 |
| kbringard | of course, change out the interface names accordingly | 14:58 |
| kbringard | between br100 and eth0 is where the natting for the VM happens | 14:59 |
| *** mrjazzcat has joined #openstack | 14:59 | |
| kbringard | which actually brings me to my thought… has anyone tried taking out the network controller entirely? and just using it as a DHCP orchestrator? | 14:59 |
| *** dragondm has joined #openstack | 14:59 | |
| livemoon | In my network. I set vm the same network address with my host | 14:59 |
| *** mrjazzcat has left #openstack | 14:59 | |
| *** ahasenack has joined #openstack | 15:00 | |
| livemoon | my br100(eth0) is 192.168.200.20 | 15:00 |
| livemoon | and my vm ip range is 192.168.200.25-29 | 15:01 |
| kbringard | so you probably want to update the gateway entry in the networks table to be .20 | 15:01 |
| livemoon | Now I use 192.168.200.20 as gateway in db | 15:01 |
| kbringard | then tear down the br100 stuff and restart nova-network | 15:01 |
| *** jedi4ever has quit IRC | 15:02 | |
| livemoon | but I have another node | 15:02 |
| livemoon | It has nova-network itself | 15:02 |
| *** atkisc has joined #openstack | 15:02 | |
| livemoon | I see "mutil-host" in nova-manage network create command | 15:03 |
| livemoon | how to use mutil-host | 15:04 |
| *** joesavak has joined #openstack | 15:04 | |
| kbringard | I don't know much about that, I've mostly used VLAN mode and only recently done basic FLatDHCP stuff | 15:04 |
| *** dolphm has joined #openstack | 15:04 | |
| livemoon | kbringard: thank you. | 15:04 |
| kbringard | vishy wrote this awhile back about HA network, perhaps it'll help: http://unchainyourbrain.com/openstack/13-networking-in-nova | 15:04 |
| *** jsavak has quit IRC | 15:04 | |
| livemoon | kbringard: 谢谢 | 15:06 |
| chmouel | vishy: http://pastie.org/private/hepz165pavvmicpvruvqg | 15:06 |
| atkisc | virshy: dashboar + keystone ? success ? | 15:07 |
| kbringard | livemoon: you're welome | 15:07 |
| dolphm | atkisc: vishy: was that a bug in keystone or dashboard? | 15:07 |
| kbringard | livemoon: 不客氣 | 15:07 |
| kbringard | is that right? | 15:08 |
| *** ldlework has quit IRC | 15:08 | |
| livemoon | kbringard: cool | 15:08 |
| livemoon | atkisc: 先放下dashboard吧 | 15:09 |
| atkisc | 放不下了 | 15:09 |
| livemoon | so many people are waiting for dashboard. | 15:11 |
| livemoon | Does anyone know "mutil-host" in nova-manage network | 15:12 |
| *** vladimir3p has joined #openstack | 15:14 | |
| Razique | hey kbringard :) | 15:14 |
| livemoon | oh, I know it via google. Good night all. | 15:15 |
| kbringard | hey Razique | 15:15 |
| kbringard | g'night livemoon | 15:15 |
| Razique | good night | 15:16 |
| livemoon | kbringard is a good guy. | 15:16 |
| kbringard | so if anyone is in the mood to wax intellectual about nova network stuff… | 15:16 |
| Razique | put me on track dude :) | 15:17 |
| Razique | I just arrived! | 15:17 |
| *** reidrac has quit IRC | 15:17 | |
| livemoon | Razique :) | 15:17 |
| *** FabriceB has quit IRC | 15:18 | |
| kbringard | I was thinking about vishy's HA network thing, but in truth I don't know a ton about it (I've been lazy and not kept up on it's development) | 15:18 |
| kbringard | I was considering doing a hybrid HA, where I offload the routing and meta-data ip forwarding to upstream hardware | 15:18 |
| kbringard | that's easy enough to do | 15:18 |
| livemoon | what about db? | 15:19 |
| kbringard | and basically leaves us with NAT and DHCP as the network controller's jobs | 15:20 |
| livemoon | db HA? | 15:20 |
| kbringard | livemoon: I do M-M replication using Linux-HA for hot failover | 15:20 |
| kbringard | for DB | 15:20 |
| livemoon | ok | 15:20 |
| kbringard | so, as far as the network thing goes… I was thinking about ways to abstract that another layer and have the natting happen in a separate cluster of machines, instead of directly on the compute nodes | 15:21 |
| *** rsampaio has joined #openstack | 15:22 | |
| medberry | livemoon, vishy knows multi_host but then he created it. | 15:22 |
| kbringard | vishy knows most everything because he created most of it ;-) | 15:22 |
| *** rsampaio has quit IRC | 15:22 | |
| *** ldlework has joined #openstack | 15:23 | |
| livemoon | grab vishy | 15:23 |
| *** rsampaio has joined #openstack | 15:23 | |
| bpaluch | hmm I think I'm running into the same problem you are having livemoon | 15:24 |
| livemoon | bpaluch: which problem? | 15:24 |
| chmouel | vishy: http://pastie.org/private/5exr69jdogy39qktr9eg | 15:25 |
| bpaluch | I want VM traffic to traverse my network hardware in vlan mode | 15:25 |
| *** neogenix has joined #openstack | 15:25 | |
| bpaluch | I have a lot of servers that can't be VM's in this envo and I don't want to nat my vm's I just want them on a certain vlan | 15:25 |
| *** ldlework has quit IRC | 15:25 | |
| livemoon | bpaluch: I use FLATDHCP not vlan | 15:25 |
| *** llang629 has joined #openstack | 15:26 | |
| bpaluch | oh maybe I just read half of what you were dealing with | 15:26 |
| livemoon | bpaluch: kbringard use vlan mode | 15:26 |
| *** nyeates has quit IRC | 15:26 | |
| *** llang629 has left #openstack | 15:27 | |
| livemoon | my test envirment is poor, no vlan switch. so I just use FLAT | 15:27 |
| chmouel | ARGS=($@) | 15:27 |
| chmouel | echo ${ARGS[@]:1:2} | 15:27 |
| bpaluch | I think I may have configured my cisco switch wrong, I need to consult my cisco guru in my office, I've never trunked a switch to a server like this. I have a vlan which gateway is on my main router, then a switch that is trunked off that router, then ports on that switch connected to the second nic on my servers. | 15:28 |
| livemoon | chmouel is writing scripts | 15:29 |
| bpaluch | anyway, my vlan gw is 10.253.52.129, so i'm guessing my br100 has to be 10.253.52.130 on my host? | 15:29 |
| *** ldlework has joined #openstack | 15:29 | |
| chmouel | livemoon: yeah playing with bash array | 15:30 |
| livemoon | chmouel: good child | 15:30 |
| *** ameade has joined #openstack | 15:30 | |
| livemoon | sleeping, see you tomorrow. | 15:31 |
| *** dobber has quit IRC | 15:31 | |
| *** mdomsch has joined #openstack | 15:31 | |
| vishy | jaypipes: magic https://gist.github.com/1245389 | 15:31 |
| Razique | you can use clan, juste the with have to let pass tagged paquets | 15:31 |
| livemoon | bpaluch:if I test vlan, I will chat with you about the problem. | 15:32 |
| *** jsavak has joined #openstack | 15:32 | |
| kbringard | vishy: did you see my thought above… about using upstream routing hardware but then having an entire NAT service instead of a network controller on each compute node? | 15:32 |
| vishy | kbringard no i didn't look | 15:33 |
| livemoon | vishy: Does mutil-host only have two value: T or F? | 15:33 |
| jaypipes | vishy: ewww. :) | 15:33 |
| vishy | livemoon: yes | 15:33 |
| kbringard | ah, well, that's pretty much it… any thoughts on such a beast? | 15:33 |
| *** joesavak has quit IRC | 15:33 | |
| vishy | jaypipes: parsing json without python is a little tough :) | 15:34 |
| jaypipes | vishy: that is an admirable hack indeed :) | 15:34 |
| vishy | jaypipes: i get errors about missing min_ram values when i try to use a current client with an old install btw | 15:34 |
| vishy | i assume that is expected | 15:35 |
| jaypipes | vishy: that is a bug. | 15:35 |
| jaypipes | vishy: :( | 15:35 |
| livemoon | kbringard: It seems simple and effective arch | 15:35 |
| *** shaibn has quit IRC | 15:37 | |
| *** livemoon has quit IRC | 15:37 | |
| *** livemoon has joined #openstack | 15:38 | |
| *** dendro-afk is now known as dendrobates | 15:39 | |
| *** arreyder has quit IRC | 15:39 | |
| *** arreyder has joined #openstack | 15:39 | |
| *** imsplitbit has quit IRC | 15:40 | |
| kbringard | livemoon: yea, I'm just not entirely sure how to tell nova to offload the NAT functionality to this other set of machines | 15:40 |
| livemoon | but I think it is not suitible for large envirement | 15:40 |
| *** mattray has quit IRC | 15:41 | |
| *** nyeates has joined #openstack | 15:41 | |
| livemoon | Goodbye, learn from you all tomorrow. | 15:41 |
| *** obino has quit IRC | 15:41 | |
| kbringard | 'night | 15:41 |
| *** livemoon has quit IRC | 15:41 | |
| uvirtbot | New bug: #860666 in nova "Exception messages don't get displayed correctly in test suite" [Undecided,New] https://launchpad.net/bugs/860666 | 15:42 |
| Razique | Why openstack uses Erlang ? | 15:42 |
| *** Gordonz has joined #openstack | 15:43 | |
| *** v0id has joined #openstack | 15:46 | |
| jaypipes | Razique: it doesn't. RabbitMQ is written mostly in Erlang. | 15:47 |
| Razique | ah ok, thanks :) | 15:47 |
| *** bcwaldon has joined #openstack | 15:48 | |
| *** Fredric has quit IRC | 15:48 | |
| *** zz_bonzay is now known as zz_bonzay|away | 15:49 | |
| *** zz_bonzay|away is now known as bonzay | 15:49 | |
| bpaluch | vishy: can you explain the HW gateway section of exisint HA a little? what does pass the dhcpoption=3, ip of gateway mean? Do I pass that upon creating the instance or is that in a conf file somewhere? | 15:50 |
| kbringard | bpaluch: you'd put that in your dnsmasq.conf | 15:50 |
| *** bonzay is now known as bonzay|away | 15:50 | |
| kbringard | so that when it hands out the DHCP lease to the VM, it tells the VM to use the ip specified in the option as it's default gateway | 15:51 |
| *** bonzay|away is now known as zz_bonzay|away | 15:51 | |
| kbringard | which would allow you to route out of a hardware router directly, rather than using the nova-network host as a software router | 15:51 |
| bpaluch | ah ic | 15:51 |
| kbringard | but, you have to setup some stuff manually on the router, that nova-network usually sets up | 15:52 |
| kbringard | like, the ip forward to make 169.254.169.254 forward to the meta-data service | 15:53 |
| bpaluch | will nova-network set up dhcp still? | 15:53 |
| kbringard | and you have to make sure the router can route for the network(s) the VMs are on | 15:53 |
| *** openpercept has quit IRC | 15:53 | |
| kbringard | yea, in that scenario nova-network still does DHCP and NAT | 15:53 |
| bpaluch | yeah I already have my vlan's set up there anyway | 15:53 |
| bpaluch | I think this is the best solution, the projects I work on are network heavy, and I can't see a software switch being very scalable for them. | 15:54 |
| kbringard | nat being floating IP, since of course the router would have to do NAT for the VM network if it wasn't publicly routable IP space | 15:54 |
| bpaluch | but I could set up nat on my router too then | 15:55 |
| kbringard | yea | 15:55 |
| kbringard | but, for floating IPs | 15:55 |
| bpaluch | hmm thanks now I just have to read up on dnsmasq.conf | 15:55 |
| *** imsplitbit has joined #openstack | 15:55 | |
| kbringard | unless you modify nova-network to talk to your hardware router | 15:55 |
| bpaluch | is that something that quantum is supposed to do? | 15:56 |
| bpaluch | hmm I need to get my cisco reps in here. | 15:56 |
| kbringard | yea, assuming there is a cisco plugin for quantum :-D | 15:56 |
| kbringard | it would go like | 15:56 |
| bpaluch | I think they are one of the contributers ;) | 15:56 |
| kbringard | euca-associate-address -> API server -> quantum -> cisco plugin -> NAT setup on the hardware | 15:56 |
| *** irahgel has quit IRC | 15:57 | |
| Razique | time to go home | 15:59 |
| Razique | cya stackers :) | 15:59 |
| *** Razique has left #openstack | 15:59 | |
| *** swift82 has quit IRC | 15:59 | |
| *** ldlework has quit IRC | 16:01 | |
| *** irahgel has joined #openstack | 16:02 | |
| *** HowardRoark has joined #openstack | 16:03 | |
| *** HowardRoark has quit IRC | 16:05 | |
| *** Ryan_Lane has quit IRC | 16:05 | |
| *** elasticdog has joined #openstack | 16:07 | |
| *** FabriceB has joined #openstack | 16:09 | |
| gnu111 | what's the easiest way to upgrade from cactus to diablo in RHEL. I finally got a diablo setup working properly. But I also have a cactus setup with several users and few images loaded. If possible would like not to re-install that setup. | 16:09 |
| *** katkee has quit IRC | 16:10 | |
| *** stewart_ has joined #openstack | 16:10 | |
| *** stewart has quit IRC | 16:11 | |
| *** imsplitbit has quit IRC | 16:11 | |
| *** nicolas2b has quit IRC | 16:11 | |
| *** aliguori has quit IRC | 16:13 | |
| scalability-junk | Would you advice against openstack for a 2 server configuration to start with? | 16:14 |
| *** cp16net_ has joined #openstack | 16:15 | |
| kbringard | scalability-junk: I don't think anyone in here would advise against openstack for any reason :-D | 16:15 |
| *** llang629 has joined #openstack | 16:16 | |
| scalability-junk | kbringard: yeah, but I tried to setup an openstack setup with 2 servers each server running all components, so you can have failover with the second server. and this setup was just too small back then. So I wanted to check, if openstack ist still only for the big players or if there is any way to make it work in small scales too? | 16:17 |
| *** cp16net_ has quit IRC | 16:17 | |
| *** cp16net_ has joined #openstack | 16:17 | |
| *** cp16net has quit IRC | 16:18 | |
| *** cp16net_ is now known as cp16net | 16:18 | |
| kbringard | it should work fine | 16:18 |
| scalability-junk | kbringard: not really against openstack, more advising against small scale | 16:18 |
| *** code_franco has quit IRC | 16:18 | |
| scalability-junk | kbringard: any docs, about setting up openstack in a 2 server configuration, still enabling failover of all components? | 16:18 |
| *** llang629 has left #openstack | 16:18 | |
| *** rnirmal has quit IRC | 16:19 | |
| *** jsavak has quit IRC | 16:21 | |
| kbringard | scalability-junk: not really… although I think the network controller is really the only thing that needs coordination | 16:21 |
| kbringard | everything else should just pull messages out of the queue and work autonomously | 16:21 |
| kbringard | http://unchainyourbrain.com/openstack/13-networking-in-nova | 16:21 |
| *** jsavak has joined #openstack | 16:21 | |
| kbringard | that article gives a good outline of vishy's ideas for network HA | 16:21 |
| kbringard | where you put a network controller on each compute node and it acts as the controller for all the VMs on that node | 16:22 |
| kbringard | that's probably a good place to start | 16:22 |
| kbringard | as far as API, etc, I've setup multiple API servers behind a load balancer, done DB replication and failover, etc | 16:22 |
| scalability-junk | ok I'll take a look at that. thanks | 16:23 |
| scalability-junk | kbringard: what about running openstack compute and storage on one server good idea or really bad one? | 16:24 |
| kbringard | it depends on what kind of load you're going to put on the system | 16:24 |
| kbringard | if this is a production system that you plan to sell to people | 16:24 |
| *** obino has joined #openstack | 16:26 | |
| *** ldlework has joined #openstack | 16:26 | |
| *** heckj has joined #openstack | 16:26 | |
| scalability-junk | kbringard: more a private production setup | 16:28 |
| *** heckj has quit IRC | 16:29 | |
| *** heckj has joined #openstack | 16:29 | |
| kbringard | :shrug: it should be OK, but I mean, the more you spread it out, the less disruption a single hardware failure causes :-) | 16:29 |
| *** blakeyeager has joined #openstack | 16:30 | |
| scalability-junk | kbringard: Yeah I know. I would love to spread it out onto a few servers, but I'm a student, I love to play with such things and don't want to use as is IaaS. But the money isn't available to support more than 1 perhaps 2 servers ;) | 16:31 |
| *** heckj has quit IRC | 16:32 | |
| *** heckj has joined #openstack | 16:32 | |
| *** foexle has quit IRC | 16:34 | |
| *** rustam has joined #openstack | 16:37 | |
| *** nacx has quit IRC | 16:37 | |
| *** llang629 has joined #openstack | 16:39 | |
| *** llang629 has left #openstack | 16:39 | |
| *** chemikadze has quit IRC | 16:39 | |
| *** jdurgin has joined #openstack | 16:40 | |
| *** huslage has quit IRC | 16:40 | |
| *** chemikadze has joined #openstack | 16:40 | |
| *** huslage has joined #openstack | 16:40 | |
| *** huslage has joined #openstack | 16:41 | |
| *** irahgel has quit IRC | 16:41 | |
| *** jedi4ever has joined #openstack | 16:41 | |
| *** aliguori has joined #openstack | 16:42 | |
| tsel | You can configure all nova components on one server with one Ethernet? would you specify it on the nova.conf as both flat_interface and public_interface? | 16:45 |
| heckj | tsel: yep, that works | 16:46 |
| tsel | Thanks. | 16:49 |
| *** po has joined #openstack | 16:49 | |
| *** ejat has joined #openstack | 16:55 | |
| *** lhrc_ has joined #openstack | 16:55 | |
| *** ejat has quit IRC | 16:56 | |
| *** ejat has joined #openstack | 16:56 | |
| *** nerens has joined #openstack | 16:56 | |
| *** Ryan_Lane has joined #openstack | 16:57 | |
| *** ohnoimdead has joined #openstack | 16:59 | |
| *** zul has quit IRC | 17:01 | |
| *** zul has joined #openstack | 17:01 | |
| *** bcwaldon has quit IRC | 17:03 | |
| *** blueblade has quit IRC | 17:04 | |
| *** hggdh has quit IRC | 17:05 | |
| *** bcwaldon has joined #openstack | 17:05 | |
| *** hggdh has joined #openstack | 17:06 | |
| *** hggdh has joined #openstack | 17:06 | |
| *** localhost has quit IRC | 17:08 | |
| *** code_franco has joined #openstack | 17:09 | |
| *** localhost has joined #openstack | 17:09 | |
| *** zul has quit IRC | 17:11 | |
| *** zul has joined #openstack | 17:13 | |
| bpaluch | kbringard: if I put -dhcpoption=3,ipOfGateway in my dnsmasq.conf what would my bridge connection IP be? | 17:14 |
| *** lhrc_ is now known as mikeyp-3 | 17:17 | |
| kbringard | hmm, thats a good question… I guess it wouldn't really matter so long as it was something in the VM subnet's broadcast domain so that it can send/receive DHCP requests | 17:17 |
| kbringard | and was connected to an interface that can route to whatever upstream hardware you're using | 17:18 |
| rmk | Has anyone created images which are self contained? As in, do not separate the initrd and kernel? | 17:18 |
| kbringard | rmk: yep | 17:18 |
| bpaluch | is the dnsmasq service restarted by nova-network? | 17:18 |
| rmk | Cool had a few questions if you don't mind | 17:18 |
| kbringard | I've used virt-install to install Cent images | 17:19 |
| kbringard | bpaluch: kind of… it'll start it, but it won't re-start it | 17:19 |
| rmk | kbringard: Nova is passing root=/dev/vda -- but that obviously implies there's no partitioning whatsoever on the image | 17:19 |
| rmk | So how do you get a bootloader onto that? | 17:19 |
| kbringard | so you'll probably want to kilall dnsmasq (assuming the only dnsmaq processes were started by nova-network) then restart it | 17:19 |
| bpaluch | oh so if I change my dnsmasq conf file, then I have to restart it myself? | 17:19 |
| bpaluch | ok, hmm hope that is why it didn't work. | 17:20 |
| kbringard | yea, I think HUPing it will work | 17:20 |
| kbringard | but, I generally just killall dnsmasq, then restart nova-network to let it bring it back upo | 17:21 |
| rmk | kbringard: I've basically been taking a xenserver image and ultimately getting it to .raw, then extracting the root partition. All well and good if I want to separate the kernel/initrd, but I really do not. | 17:21 |
| kbringard | rmk: well, it needs /boot | 17:21 |
| rmk | Why? | 17:21 |
| rmk | If grub is there it really does not. The question is how to get grub on there. | 17:21 |
| kbringard | that's how you get the kernel on the /dev/vda | 17:21 |
| kbringard | I've not converted an existing image | 17:21 |
| kbringard | I've only used virt-install to build a new image from a kickstart profile | 17:22 |
| rmk | ok how do you typically layout the images? | 17:22 |
| kbringard | so I have not futzed with it at that level | 17:22 |
| *** dragondm has quit IRC | 17:22 | |
| kbringard | I just create a single / partition | 17:22 |
| kbringard | on /dev/vda | 17:22 |
| kbringard | then /boot /initrd, etc are all directly on / | 17:22 |
| kbringard | instead of being being their own mount in the libvirt.xml | 17:22 |
| rmk | Right that makes sense | 17:23 |
| kbringard | but again, this was all done automagically for me by virt-install | 17:23 |
| rmk | Isn't nova expecting /dev/vda to literally be the root filesystem? | 17:23 |
| *** adjohn has quit IRC | 17:23 | |
| rmk | I guess my problem is there isn't enough room for grub because of the way I'm doing conversions. | 17:24 |
| kbringard | ah, that could be | 17:25 |
| kbringard | yea, I think nova expects /dev/vda to be the root | 17:25 |
| rmk | Which should be fine come to think of it, so long as I can get grub at the beginning of that. | 17:25 |
| kbringard | but like you said, that's where grub should live, too | 17:25 |
| *** javiF has quit IRC | 17:30 | |
| hacim | ok i've been following the guide for setting up swift, and I've gotten to this point: http://docs.openstack.org/diablo/openstack-object-storage/admin/content/verify-swift-installation.html | 17:30 |
| hacim | this seems to presume I have some swauth installed? | 17:31 |
| vishy | bpaluch: you need to set it as an option for dnsmasq | 17:32 |
| vishy | there is a flag to use an additional conf file for dnsmasq, so you can put it in a file then pass | 17:32 |
| vishy | --dnsmasq_config_file | 17:32 |
| vishy | = whatever | 17:33 |
| notmyname | hacim: you are using swift 1.4.3? | 17:33 |
| bpaluch | is that for when I create an instance? | 17:33 |
| vishy | bpaluch: ^^ | 17:33 |
| annegentle | hacim: yes "You can run these commands from the proxy server if you have installed swauth there." it is an assumption | 17:33 |
| *** hugokuo has quit IRC | 17:33 | |
| vishy | bpaluch: no it has to be set in your flagfile, and you'll have to killall dnsmasq and restart nova-network for it to take affect | 17:33 |
| hacim | notmyname: I'm not yet, using the debian packages which are still 1.3 | 17:33 |
| rmk | Yeah so grub wants a partition table but if I add one, then /root immediately becomes /dev/vda1, which nova won't like | 17:33 |
| annegentle | however you can still verify the installation using the steps, just the actual commands will vary | 17:33 |
| bpaluch | oh | 17:33 |
| vishy | bpaluch: and you'll have to use some magic to get metadata to work | 17:34 |
| notmyname | hacim: ah. (then you are using the wrong package source) | 17:34 |
| *** rastem has joined #openstack | 17:34 | |
| bpaluch | magic on my router right? I just need to forward traffic from 169 address to my nova-api host right? | 17:34 |
| bpaluch | hmm this solution kind of doens't work when you have multiple vlans then does it. | 17:34 |
| bpaluch | because different vlan different route gateway | 17:34 |
| hacim | annegentle: ok, so its not an option. i understood from a previous step that it wasn't required | 17:35 |
| rastem | Just reading about openstack and the Swift object store. Is this technology bottlenecked by the proxy servers? | 17:35 |
| notmyname | rastem: no | 17:35 |
| atkisc | vishy: openstack error : not get info | 17:35 |
| notmyname | hacim: swauth has been removed from the swift codebase and put into a separate project. you can keep using swauth, or you can use the included tempauth | 17:36 |
| annegentle | hacim: swauth is one of three options for auth with swift - swauth, tempauth, and keystone | 17:36 |
| notmyname | hacim: (in later code than you are running now) | 17:36 |
| rastem | notmyname: so the proxy servers are peers? | 17:36 |
| annegentle | thanks notmyname | 17:36 |
| notmyname | hacim: swauth is prod-ready. tempauth is meant for POC and dev purposes | 17:36 |
| notmyname | rastem: yes | 17:36 |
| *** dirkx_ has joined #openstack | 17:36 | |
| vishy | notmyname: managed to get keystone working with a swift install. I'm wondering about somehow combining it with swauth to get acls... | 17:37 |
| enmand | Hey, has anyone had any luck getting rbd to work well with nova-volume? I seem to get an error everytime I try to attach a volume to an instance | 17:37 |
| rastem | notmyname: ok ty | 17:37 |
| notmyname | vishy: I thought there was a grand plan by the architects to get ACLs ;-) | 17:37 |
| vishy | enmand: i know it was working at some point in history, but I'm not sure of the current state | 17:38 |
| enmand | Hmm, OK | 17:38 |
| vishy | notmyname: ah yes. That may even exist by 2020 | 17:38 |
| enmand | I can create the volumes, and delete them | 17:38 |
| enmand | But I get "libvirtError: cannot resolve symlink rbd/volume-00000001: No such file or directory" when I try and attach it to an instance | 17:39 |
| *** ldlework has quit IRC | 17:39 | |
| kbringard | bpaluch: yea, I don't know how well it'll work with VLANs… you'd have to setup your dnsmasq.conf to hand out a different gateway for each network | 17:39 |
| hacim | well, I do have swauth-prep, etc. what I dont have is something listening on port 8080 | 17:39 |
| kbringard | which sounds painful | 17:39 |
| *** ohadlevy has quit IRC | 17:40 | |
| bpaluch | kbringard: yeah hmm not the optimal solution, not even sure how I could have dnsmasq do that | 17:40 |
| kbringard | I think you can setup dnsmasq to hand out different stuff based on the subnet making the request, but either way | 17:41 |
| kbringard | hairy, to be sure | 17:41 |
| *** rastem has quit IRC | 17:41 | |
| notmyname | vishy: I'm sure it's possible. I'm looking at keystone (and the 2.0 spec in general) over the next few weeks to figure out how to integrate cloud files. I hope to be more informed about it by next week so I can give you a reasonable answer | 17:41 |
| * hacim starts up the swift-proxy-server | 17:41 | |
| bpaluch | vishy: I have my vm getting the HW route address, do I need to do something special with the bridge connection on the host machine? I can't pass traffic to the gateway on the VM, but I can ssh to the vm through the host, but nowhere else. | 17:42 |
| hacim | hm, ok swift-proxy is running, and listening on 8080, but the swprep isn't returning anything when it starts | 17:43 |
| notmyname | hacim: anything in the proxy log? and swauth is in the proxy pipeline, right? | 17:44 |
| notmyname | swauth prep may not actually give any output. I don't remember | 17:44 |
| hacim | notmyname: i'm looking around for a log, but so far haven't found one | 17:44 |
| hacim | notmyname: it doesn't return, just sits there | 17:45 |
| notmyname | hacim: /var/log/syslog (unless you've changed your syslog configs) | 17:45 |
| hacim | i'm not sure I understand what you mean by 'in the proxy pipeline' | 17:45 |
| *** darraghb has quit IRC | 17:45 | |
| notmyname | hacim: in the proxy-server.conf file there is a pipeline command. swauth should be there (along with a [filter:swauth] section below it) | 17:45 |
| hacim | notmyname: so syslog shows it stopping and starting, but other than that, no output | 17:46 |
| hacim | ah | 17:46 |
| notmyname | hacim: it? | 17:46 |
| *** wariola has joined #openstack | 17:48 | |
| hacim | notmyname: the swift-proxy-server... but you were right, I didn't have it in the pipeline, and now that I add it I get a python error "pkg_resources.DistributionNotFound:" - which is probably because I need to get it from git | 17:49 |
| hacim | but, perhaps it is better if I wait until I get a newer version of software to play with so I run into less snags | 17:49 |
| vishy | bpaluch: hmm, well you hve to make sure that it is bridged in to an interface that has access to that ip | 17:49 |
| notmyname | hacim: good idea :-) | 17:49 |
| notmyname | hacim: mtaylor can set you straight on where to get the updated packages | 17:50 |
| *** dirkx_ is now known as dirkx | 17:50 | |
| mtaylor | hacim: what's up? | 17:50 |
| hacim | notmyname: i'm using debian unstable, and the packager is working on getting the new ones uploaded | 17:50 |
| hacim | mtaylor: no need... i just realized I'm using older code and I should wait until zigo updates the packages in debian before I flounder around more :) | 17:51 |
| bpaluch | vishy, the bridge in interface has access, but the port on my cisco router is a trunked port, so its possible I did that wrong. I have switchport mode trunk , and switchport trunk encapsulation dot1q on the port, | 17:51 |
| *** ohadlevy has joined #openstack | 17:51 | |
| *** rastem_ has joined #openstack | 17:51 | |
| mtaylor | hacim: well, in the mean time we do have packages at https://launchpad.net/~openstack-release/+archive/2011.3/ - although they are currently targetted at ubuntu | 17:51 |
| bpaluch | I actually can't ping the gateway from my host node... | 17:51 |
| mtaylor | hacim: adding debian packages to a repo we provide is on the todo list | 17:52 |
| *** rastem_ has joined #openstack | 17:53 | |
| *** retr0h has joined #openstack | 17:53 | |
| *** ameade has quit IRC | 17:53 | |
| vishy | bpaluch: can you ping the gateway from the host machine? | 17:53 |
| bpaluch | oh wait, I think my bridge interface has the wrong route, even though it has the correct network | 17:53 |
| vishy | bpaluch: interesting | 17:53 |
| hacim | mtaylor: ok, thanks! | 17:53 |
| *** rastem_ has left #openstack | 17:54 | |
| bpaluch | http://paste.openstack.org/show/2568/ | 17:54 |
| *** rastem has joined #openstack | 17:54 | |
| rastem | Hi, another question: what's the reason containers can't be nested? | 17:54 |
| bpaluch | 10.253.52.192 is the hosts network on eth0 | 17:54 |
| bpaluch | and eth1 is supposed to have the vlan network | 17:54 |
| vishy | bpaluch: and what ip does the vm have? | 17:55 |
| bpaluch | 10.253.52.132 | 17:55 |
| vishy | bpaluch: how is that going to work? | 17:55 |
| bpaluch | I have --vlan_interface=eth1 in my nova.conf | 17:55 |
| rastem | I have a need to map a directory structure from data coming out of machines, so need 'directories' | 17:56 |
| vishy | bpaluch: are you using vlan mode? | 17:56 |
| bpaluch | vishy: my router is handling the VM's. | 17:56 |
| bpaluch | yeah vlan mode | 17:56 |
| vishy | ok | 17:56 |
| vishy | i don't see any traffic on the vlan at all | 17:56 |
| vishy | so if you do brctl show | 17:57 |
| hacim | annegentle: if swauth is prod-ready, tempauth is for POC, what is keystone for? | 17:57 |
| vishy | it shows vnet0 bridged into vlan191? | 17:57 |
| vishy | hacim: integration | 17:57 |
| bpaluch | vishy: no | 17:57 |
| bpaluch | br1008000.02163e5037a4novlan161 | 17:57 |
| vishy | bpaluch: no? where is it bridged? | 17:57 |
| bpaluch | oh wait | 17:58 |
| bpaluch | yes it is, I didn't see vnet0 listed there too | 17:58 |
| vishy | * i meant 161 | 17:58 |
| vishy | ok so vnet0 and vlan161 are bridged | 17:58 |
| bpaluch | http://paste.openstack.org/show/2569/ | 17:58 |
| bpaluch | yeah | 17:58 |
| vishy | but no traffic is showing on vlan161 | 17:58 |
| *** jedi4ever has quit IRC | 17:58 | |
| gnu111 | what does this error mean: nova: Class Client could not be found | 17:59 |
| vishy | so your vlan is not set right | 17:59 |
| vishy | gnu111: probably missing glance dependency | 17:59 |
| vishy | bpaluch: try ip addr | 17:59 |
| *** mikeyp-3 has quit IRC | 17:59 | |
| gnu111 | it is in the compute node. Do I need to install glance? | 18:00 |
| vishy | bpaluch: make sure vlan161 is on eth1 like it supposed to be | 18:00 |
| vishy | gnu111: it needs glance client | 18:00 |
| gnu111 | ok | 18:00 |
| vishy | gnu111: python-glance should have it | 18:00 |
| bpaluch | 10: vlan161@eth1: <BROADCAST,MULTICAST,M-DOWN> mtu 1500 qdisc noop state DOWN | 18:00 |
| bpaluch | oh its down | 18:00 |
| bpaluch | hmm | 18:00 |
| vishy | hehe | 18:00 |
| bpaluch | wt heck wonder why and when that happen | 18:01 |
| *** mattray has joined #openstack | 18:01 | |
| *** dirkx has quit IRC | 18:01 | |
| vishy | bpaluch: do you have port security turned on on your router? | 18:01 |
| bpaluch | no | 18:02 |
| vishy | bpaluch: up it? | 18:02 |
| vishy | bpaluch: it looks like eth1 might be down as well? | 18:02 |
| bpaluch | yeah it is testing to see if it work now | 18:02 |
| bpaluch | Its working!! | 18:03 |
| bpaluch | thanks! | 18:03 |
| vishy | bpaluch: awesome | 18:04 |
| vishy | bpaluch: you will still have issues with metadata forwarding | 18:04 |
| vishy | which could be a bit hairy | 18:04 |
| vishy | but at least you have connectivity | 18:04 |
| vishy | :) | 18:04 |
| *** stuntmachine has joined #openstack | 18:04 | |
| bpaluch | I can figure out metadata forwarding on my router I think. I'll put a static route in | 18:05 |
| bpaluch | I just need to forward to my nova-api host right? | 18:05 |
| *** dirkx has joined #openstack | 18:06 | |
| vishy | bpaluch: you can do that | 18:09 |
| vishy | bpaluch: i think it will work if you are running nova-network on every host as well | 18:09 |
| vishy | the prerouting rule should kick in | 18:09 |
| *** martine has joined #openstack | 18:09 | |
| *** willi__ has joined #openstack | 18:10 | |
| vishy | bpaluch: the static route can work, but you have to make sure that traffic can get back into the vlans | 18:12 |
| vishy | bpaluch: so you might have to do some tweaking to get it right | 18:12 |
| *** jedi4ever has joined #openstack | 18:13 | |
| *** bpaluch has quit IRC | 18:14 | |
| uvirtbot | New bug: #860779 in openstack-ci "launchpad bug update hook should handle bugs targeted to more than one project" [Medium,Triaged] https://launchpad.net/bugs/860779 | 18:15 |
| *** willi__ has quit IRC | 18:16 | |
| *** hggdh has quit IRC | 18:18 | |
| *** bpaluch has joined #openstack | 18:20 | |
| *** FabriceB has quit IRC | 18:20 | |
| *** cereal_bars has joined #openstack | 18:21 | |
| *** ldlework has joined #openstack | 18:23 | |
| uvirtbot | New bug: #860778 in keystone "missing dependencies" [Undecided,New] https://launchpad.net/bugs/860778 | 18:23 |
| *** rastem has quit IRC | 18:23 | |
| *** koolhead17 has joined #openstack | 18:23 | |
| *** dirkx has quit IRC | 18:25 | |
| *** hggdh has joined #openstack | 18:25 | |
| *** hggdh has joined #openstack | 18:26 | |
| *** clauden_ has quit IRC | 18:28 | |
| *** clauden_ has joined #openstack | 18:28 | |
| *** dragondm has joined #openstack | 18:28 | |
| *** Razique has joined #openstack | 18:29 | |
| gnu111 | euca-terminate-instances does not disassociate the public ip address. is this a bug? or something wrong with my config here. | 18:29 |
| *** jedi4ever has quit IRC | 18:29 | |
| *** stuntmachine has quit IRC | 18:29 | |
| *** code_franco has quit IRC | 18:30 | |
| *** jedi4ever has joined #openstack | 18:30 | |
| *** ameade has joined #openstack | 18:30 | |
| *** Razique has quit IRC | 18:31 | |
| *** Razique has joined #openstack | 18:32 | |
| *** fabiand__ has joined #openstack | 18:33 | |
| *** fabiand__ has left #openstack | 18:34 | |
| *** katkee has joined #openstack | 18:35 | |
| *** Razique has quit IRC | 18:37 | |
| *** hggdh has quit IRC | 18:38 | |
| *** mgius has quit IRC | 18:39 | |
| *** hggdh has joined #openstack | 18:44 | |
| *** stewart_ is now known as stewart | 18:47 | |
| *** wariola has quit IRC | 18:47 | |
| *** mgius has joined #openstack | 18:48 | |
| *** rustam has quit IRC | 18:56 | |
| *** bpaluch has joined #openstack | 18:57 | |
| *** konetzed has left #openstack | 18:59 | |
| *** Eyk^off is now known as Eyk | 19:00 | |
| *** tsel has quit IRC | 19:01 | |
| *** ohadlevy has quit IRC | 19:03 | |
| *** jedi4ever has quit IRC | 19:05 | |
| *** ejat has quit IRC | 19:05 | |
| *** obino has quit IRC | 19:06 | |
| *** obino has joined #openstack | 19:06 | |
| *** ohadlevy has joined #openstack | 19:07 | |
| *** jakedahn has quit IRC | 19:08 | |
| *** dirkx has joined #openstack | 19:08 | |
| *** mrmartin has joined #openstack | 19:09 | |
| *** johnpur has joined #openstack | 19:11 | |
| *** ChanServ sets mode: +v johnpur | 19:11 | |
| *** rastem has joined #openstack | 19:12 | |
| rastem | quick question: proxy servers are the front-facing servers, and storage nodes are the ones that sit on a private network, right? | 19:13 |
| *** g2 has joined #openstack | 19:13 | |
| andyandy_ | I am trying to use the cloudpipe with this command "sudo nova-manage vpn run --user=admin --project=project" but I have an error in nova-network:(nova.rpc): TRACE: FixedIpNotFoundForNetwork: Fixed IP address (%(address)s) does not exist in network (%(network_uuid)s, do you have a suggestion? | 19:14 |
| heckj | rastem: yep, the proxy is what does the interaction. There's no "You must have the separated", but if you had a front and back network, that would be a logical separation. | 19:15 |
| rastem | heckj: ah, so you could in theory have proxy servers be storage nodes? | 19:15 |
| heckj | rastem: I believe it is possible to run the proxy processes on the storage nodes. Most larger installs have them separated, but I don't believe that's mandatory | 19:16 |
| rastem | heckj: ok cool ty for the help | 19:17 |
| *** nyeates has quit IRC | 19:17 | |
| heckj | rastem: np | 19:17 |
| *** ohadlevy has quit IRC | 19:18 | |
| *** dolphm has quit IRC | 19:20 | |
| g2 | Hi, | 19:20 |
| g2 | I am trying to create a network using "nova-manage network create private 10.10.1.0/24 1 256 --bridge=br100", but it fails with an error "Command failed, please check log for more info" | 19:21 |
| *** MarkAtwood has joined #openstack | 19:21 | |
| heckj | --bridge is the wrong extension. I believe you want "--flat-interface" | 19:21 |
| g2 | There is no information in the logs. Can someone give me pointers about what may be going on/ | 19:21 |
| g2 | heckj: You mean "nova-manage network create private 10.10.1.0/24 1 256 --flat-interface=br100" | 19:22 |
| g2 | I get an error, no such option --flat-interface | 19:23 |
| Vinsh_HP | _ | 19:23 |
| *** ohadlevy has joined #openstack | 19:23 | |
| heckj | g2: just a sec, looking up my install notes | 19:24 |
| Vinsh_HP | its an _ not a - | 19:24 |
| Vinsh_HP | --flat_interface | 19:24 |
| g2 | VInsh_HP: I get the same error for --flat_interface also | 19:24 |
| heckj | g2: --flat_interface is the flag for the config file, and in my notes I have this as an install command line: "nova-manage network create private --fixed_range_v4=10.0.0.0/24 --num_networks=1 --network_size=256 --bridge=br100 --bridge_interface=eth0 --multi_host=T" | 19:25 |
| heckj | Note, I was using --multi-host in this too - you can ignore that. | 19:25 |
| *** sdake_ has joined #openstack | 19:25 | |
| *** sdake has quit IRC | 19:26 | |
| bpaluch | vishy: http://paste.openstack.org/show/2572/ this is how you can have different gateways per subnet in dnsmasq.conf in case you you were curious | 19:26 |
| g2 | heckj: It worked, many thanks! | 19:26 |
| g2 | One more questions: The network that is created doesn't have any uuid. Why is that? | 19:27 |
| g2 | I run "nova-manage network list" and all the networks have "None" as uuid | 19:27 |
| *** sdake_ has quit IRC | 19:29 | |
| heckj | g2: sorry, don't know - deeper than I can go on the code right now | 19:29 |
| *** sdake has joined #openstack | 19:30 | |
| g2 | heckj: Okay, thanks for your help. | 19:31 |
| *** rastem has quit IRC | 19:32 | |
| *** mrmartin has quit IRC | 19:34 | |
| *** dgags has quit IRC | 19:34 | |
| *** MarcMorata has quit IRC | 19:37 | |
| *** dgags has joined #openstack | 19:37 | |
| *** code_franco has joined #openstack | 19:39 | |
| rmk | Getting this running everything under Virtualbox -- Error: internal error no supported architecture for os type 'hvm' -- I've already specified libvirt-type=qemu, what else am I missing? | 19:39 |
| PerlStalker | Strange. I can log into my dashboard as a regular user but not an admin user. | 19:40 |
| *** dolphm has joined #openstack | 19:41 | |
| rmk | nm dumb mistake | 19:42 |
| kbringard | rmk, you should check out boxgrinder, too, if you're interested in RH based images | 19:42 |
| *** jblesage has joined #openstack | 19:43 | |
| PerlStalker | I'm getting the error "Unexpected error: 'NoneType' object has no attribute 'id'" when I try to log in as an admin. | 19:43 |
| rmk | kbringard: cool, will do -- thanks for the suggestion | 19:43 |
| PerlStalker | Based on the log, it looks like the problem is with the tenant somewhere. | 19:43 |
| *** mattray has quit IRC | 19:44 | |
| kbringard | rmk: they even have an appliance image that you can load into OpenStack, launch and then use to build VMs :-D | 19:44 |
| *** MarkAtwood has quit IRC | 19:48 | |
| *** nati has joined #openstack | 19:50 | |
| DuncanT1 | Is anybody using the nova native api to manage volumes, rather than the euca API? | 19:51 |
| *** mdomsch has quit IRC | 19:53 | |
| *** MarkAtwood has joined #openstack | 19:53 | |
| annegentle | DuncanT1: volumes are in an extension to the compute api | 19:54 |
| DuncanT1 | annegentle: Yes, unfortunately I can only get 'details' to work, no other functions | 19:54 |
| annegentle | DuncanT1: yeah I was gonna say, I couldn't get it all to work | 19:55 |
| DuncanT1 | annegentle: Ok, thanks, that's one datapoint at least. I'll keep hacking for a while but I'm coming to the conclusion it isn't fully implemented yet | 19:57 |
| annegentle | DuncanT1: nor documented as a result :) | 19:57 |
| annegentle | no worky, no doccy | 19:57 |
| *** cereal_bars has quit IRC | 19:58 | |
| g2 | DuncanT1: I am using the OS volume api | 19:58 |
| *** nerdstein has quit IRC | 19:58 | |
| DuncanT1 | g2: Can you provide any hints at all? All I can get working is 'detail' | 19:58 |
| g2 | I have implemented the support of nova-volumes in novaclient. The change is pending the pull | 19:58 |
| *** mfer has quit IRC | 19:58 | |
| *** mfer has joined #openstack | 19:59 | |
| DuncanT1 | g2: Is there a tree I could have a look at please? | 19:59 |
| annegentle | go g2 go | 19:59 |
| *** MarkAtwood has quit IRC | 19:59 | |
| g2 | DunconT1: Here is the pull request | 20:00 |
| g2 | https://github.com/rackspace/python-novaclient/pull/121#issuecomment-2181004 | 20:00 |
| g2 | Here is the public repo: https://github.com/gagupta/python-novaclient | 20:00 |
| g2 | I am working on adding the CLI for volumes in novaclient next. | 20:01 |
| DuncanT1 | g2: That is fantastic, thanks, I'll pull down a copy and see what happens | 20:01 |
| heckj | g2: Woot! | 20:01 |
| *** jblesage has quit IRC | 20:01 | |
| *** joesavak has joined #openstack | 20:02 | |
| *** zul has quit IRC | 20:04 | |
| *** MarkAtwood has joined #openstack | 20:04 | |
| *** Razique has joined #openstack | 20:04 | |
| *** dolphm has quit IRC | 20:04 | |
| *** jsavak has quit IRC | 20:05 | |
| *** joesavak has quit IRC | 20:06 | |
| g2 | To use it just do the following: | 20:08 |
| g2 | from novaclient.v1_1 import servers | 20:08 |
| g2 | from novaclient.v1_1 import client | 20:08 |
| g2 | cs = client.Client(NOVA_USERNAME, NOVA_API_KEY, NOVA_PROJECTID, NOVA_URL) | 20:08 |
| g2 | vol_list = cs.volumes.list() | 20:08 |
| *** Binbin has quit IRC | 20:09 | |
| *** Binbin has joined #openstack | 20:09 | |
| *** nati has quit IRC | 20:11 | |
| *** hggdh has quit IRC | 20:12 | |
| Razique | back | 20:13 |
| *** joesavak has joined #openstack | 20:14 | |
| *** MarkAtwood has quit IRC | 20:14 | |
| *** jsavak has joined #openstack | 20:14 | |
| smoser | wonder if anyone can help | 20:15 |
| smoser | up to about 2 weeks ago, https://gist.github.com/1169889 would work on oneiric nova | 20:15 |
| *** hggdh has joined #openstack | 20:16 | |
| smoser | something happened recently (close to milestone) and now, when you run an instance, you're cut off from the host | 20:16 |
| *** ohadlevy has quit IRC | 20:16 | |
| smoser | http://paste.ubuntu.com/698090/ is the last bits of 'tail -f /var/log/nova' before my all-in-one goes down. | 20:17 |
| *** ohadlevy has joined #openstack | 20:17 | |
| *** jakedahn has joined #openstack | 20:22 | |
| *** ohadlevy has quit IRC | 20:22 | |
| *** ziyadb_ has quit IRC | 20:22 | |
| *** martines has quit IRC | 20:23 | |
| *** ohadlevy has joined #openstack | 20:23 | |
| negronjl | having an issue trying to get keystone working on ubuntu/oneiric. After installing the package, it's dependencies and config file ( http://paste.ubuntu.com/698105/ ), I keep getting the following error whenever I try to run "keystone-manage tenant add admin": ERROR: (OperationalError) unable to open database file None None. Any ideas on this? | 20:23 |
| DuncanT1 | g2: After a bit of messing, it works fantastically. Thanks! | 20:24 |
| uvirtbot | New bug: #860872 in glance "s3 backend cannot stream remote images" [High,In progress] https://launchpad.net/bugs/860872 | 20:26 |
| *** ziyadb has joined #openstack | 20:28 | |
| *** tryggvil has quit IRC | 20:32 | |
| *** ziyadb has quit IRC | 20:33 | |
| *** code_franco has quit IRC | 20:33 | |
| *** martines has joined #openstack | 20:34 | |
| *** katkee has quit IRC | 20:34 | |
| *** Razique has quit IRC | 20:36 | |
| uvirtbot | New bug: #860885 in keystone "(OperationalError) unable to open database file None None" [Undecided,New] https://launchpad.net/bugs/860885 | 20:41 |
| *** joesavak has quit IRC | 20:49 | |
| *** trent has joined #openstack | 20:49 | |
| *** jsavak has quit IRC | 20:49 | |
| *** ewindisch has joined #openstack | 20:51 | |
| *** pmyers has quit IRC | 20:53 | |
| *** cp16net has quit IRC | 20:55 | |
| *** pmyers has joined #openstack | 20:58 | |
| *** lorin1 has quit IRC | 21:01 | |
| *** koolhead17 has quit IRC | 21:01 | |
| *** pmyers has quit IRC | 21:03 | |
| *** bdwarr6 has joined #openstack | 21:03 | |
| *** troytoman-away is now known as troytoman | 21:03 | |
| *** mattray has joined #openstack | 21:06 | |
| *** mcohen has joined #openstack | 21:07 | |
| *** mcohen has left #openstack | 21:07 | |
| *** breardo has quit IRC | 21:09 | |
| gnu111 | smoser: are you using Diablo? maybe it is this bug: https://bugs.launchpad.net/ubuntu/+source/nova/+bug/859587 | 21:10 |
| *** tryggvil_ has joined #openstack | 21:11 | |
| *** huslage has quit IRC | 21:11 | |
| *** msivanes has quit IRC | 21:12 | |
| *** ziyadb has joined #openstack | 21:12 | |
| *** nerens has quit IRC | 21:15 | |
| *** pmyers has joined #openstack | 21:15 | |
| *** Ruetobas has quit IRC | 21:16 | |
| *** mfer has quit IRC | 21:17 | |
| *** mfer has joined #openstack | 21:19 | |
| *** nelson____ has quit IRC | 21:21 | |
| *** nelson____ has joined #openstack | 21:21 | |
| *** johnpur has quit IRC | 21:22 | |
| * scalability-junk damn why is openstack so great, hate that :D | 21:23 | |
| *** Ruetobas has joined #openstack | 21:23 | |
| *** MarkAtwood has joined #openstack | 21:26 | |
| *** martine has quit IRC | 21:27 | |
| *** pmyers has quit IRC | 21:28 | |
| *** Eyk is now known as Eyk^off | 21:28 | |
| *** aliguori has quit IRC | 21:30 | |
| *** bcwaldon has quit IRC | 21:30 | |
| *** adjohn has joined #openstack | 21:31 | |
| *** mies has quit IRC | 21:31 | |
| *** pmyers has joined #openstack | 21:33 | |
| *** troytoman is now known as troytoman-away | 21:34 | |
| *** PotHix has quit IRC | 21:36 | |
| *** troytoman-away is now known as troytoman | 21:39 | |
| *** pmyers has quit IRC | 21:40 | |
| *** mfer has quit IRC | 21:42 | |
| *** lts has quit IRC | 21:43 | |
| PerlStalker | Is there a simple howto for getting openstack installed on ubuntu? | 21:44 |
| *** gnu111 has quit IRC | 21:44 | |
| scalability-junk | PerlStalker: yeah | 21:45 |
| scalability-junk | storage or compute? | 21:46 |
| heckj | PerlStalker: If you mean compute, I wrote one up for Diablo the other day: http://www.rhonabwy.com/wp/2011/09/23/installing-openstack-diablo-release-nova-and-glance/ | 21:46 |
| PerlStalker | scalability-junk: both eventually but compute for now. | 21:46 |
| *** zul has joined #openstack | 21:46 | |
| *** dgags has quit IRC | 21:47 | |
| *** dgags has joined #openstack | 21:53 | |
| *** cole has joined #openstack | 21:54 | |
| *** bsza has quit IRC | 21:55 | |
| *** dgags has quit IRC | 21:57 | |
| *** aliguori has joined #openstack | 21:58 | |
| *** magg has joined #openstack | 22:00 | |
| magg | hello | 22:00 |
| magg | how do i know which version i have installed | 22:01 |
| kbringard | did you install from packages or code? | 22:01 |
| magg | packages | 22:01 |
| kbringard | you should be able to dpkg -l | grep nova | 22:01 |
| kbringard | or, rpm -aq | grep nova | 22:01 |
| kbringard | if you're a RH kind of person | 22:01 |
| *** Razique has joined #openstack | 22:02 | |
| magg | says 2011.2 | 22:02 |
| kbringard | that's Cactus, iirc | 22:02 |
| magg | damn | 22:02 |
| kbringard | I need to head out, see y'all tomorrow | 22:02 |
| *** kbringard has quit IRC | 22:02 | |
| magg | how do i install diablo then? in ubuntu from a ppa | 22:03 |
| magg | ppa:nova-core/trunk? | 22:04 |
| magg | help plz | 22:04 |
| *** adjohn has quit IRC | 22:06 | |
| *** mattray has quit IRC | 22:07 | |
| *** primeministerp2 has joined #openstack | 22:07 | |
| *** adjohn has joined #openstack | 22:07 | |
| magg | trunk or universe? | 22:09 |
| *** Jamey_ has joined #openstack | 22:09 | |
| magg | for ubuntu 11.04 | 22:10 |
| *** troytoman is now known as troytoman-away | 22:10 | |
| *** troytoman-away is now known as troytoman | 22:10 | |
| scalability-junk | magg: a sec | 22:10 |
| scalability-junk | try this: | 22:10 |
| magg | oks | 22:10 |
| scalability-junk | sudo apt-get install python-software-properties | 22:10 |
| scalability-junk | sudo add-apt-repository ppa:openstack-release/2011.3 | 22:10 |
| scalability-junk | sudo apt-get update | 22:10 |
| scalability-junk | that should actually update to diablo | 22:11 |
| magg | thanks let me try it | 22:11 |
| scalability-junk | magg here are some additional information: http://docs.openstack.org/diablo/openstack-compute/admin/content/migrating-from-cactus-to-diablo.html | 22:16 |
| magg | thanks | 22:16 |
| scalability-junk | no worries | 22:16 |
| *** adjohn has quit IRC | 22:16 | |
| scalability-junk | just a question on HA and failover. do I need replication for the cloudcontroller services or would it be fine to just backup the mysql database every few minutes and reinstall with a backup on failover? | 22:19 |
| *** mikeyp-3 has joined #openstack | 22:22 | |
| retr0h | scalability-junk: kbringard and I just setup rubyrepl http://www.rubyrep.org/ on our mysql masters and just use heartbeat with shared vip .. standard stuff | 22:23 |
| scalability-junk | retr0h: may i ask how your overall setup looks like and how much overall servers you got? | 22:24 |
| rmk | Does anyone have issues with VMs coming back up after a reboot using qemu (not kvm)? | 22:24 |
| rmk | I have no problems with this on kvm. | 22:24 |
| rmk | I have the proper flag set to restart vms on reboot. | 22:25 |
| *** pmyers has joined #openstack | 22:25 | |
| retr0h | scalability-junk: right now 3 proto type sites with ~20 or so compute servers... | 22:25 |
| retr0h | with another 3 sites coming up with few racks each | 22:26 |
| scalability-junk | compute servers as node/physical servers? | 22:26 |
| scalability-junk | *as in | 22:26 |
| retr0h | scalability-junk: ya | 22:26 |
| cole | can someone fill me in on the openstack / ceph announcement that just happened? was it just announcing the .35 release of ceph/librados? | 22:27 |
| scalability-junk | may I ask a second question: how do you use the 20 nodes for the 3 sites? small bits or just 7 server instances as huge vms for each site? | 22:28 |
| retr0h | scalability-junk: some are for storage (swift) some are for compute, network controller, api, etc.. nothing too fancy in the current env | 22:29 |
| *** juddm has joined #openstack | 22:30 | |
| scalability-junk | retr0h: but further up the latter. you use openstack for pooling the resources and then create small vms for loadbalancing, ssl terminal, webserver and database servers? | 22:31 |
| *** troytoman is now known as troytoman-away | 22:31 | |
| scalability-junk | are you using huge vm instances for those components or a more smaller ones? | 22:31 |
| retr0h | the orchestration above the 3 is not handled by openstack... thats glue that can be tackled any number of ways... | 22:31 |
| retr0h | thats something we are still working out | 22:33 |
| *** PerlStalker has quit IRC | 22:33 | |
| *** cole has quit IRC | 22:34 | |
| *** jfluhmann has quit IRC | 22:35 | |
| scalability-junk | retr0h: yeah I was just curious. so you are using dedicated servers for database api server etc. and use mysql replication to replicated the preferences right? | 22:35 |
| retr0h | yah | 22:35 |
| retr0h | zactly | 22:35 |
| *** pixelbeat has quit IRC | 22:36 | |
| retr0h | everything is in the database.. so long as that is sane.. restarting the nova* processes just restore from what was int he db | 22:36 |
| retr0h | personally i would rather we not use a standard database and use something more distributed like zookeeper but that gunna get everyone mad at me saying that ;P | 22:37 |
| *** pmyers has quit IRC | 22:38 | |
| WormMan | isn't that just a matter of getting sqlalchemy dealing with whatever you want to use? | 22:38 |
| scalability-junk | WormMan: but as far as I know sqlalchemy is sorf of just let you use postgre and mysl... | 22:40 |
| scalability-junk | *-is | 22:40 |
| *** benrodrigue has quit IRC | 22:42 | |
| *** MarkAtwood has quit IRC | 22:44 | |
| *** juddm has quit IRC | 22:47 | |
| *** nerdstein has joined #openstack | 22:48 | |
| *** bdwarr6 has quit IRC | 22:52 | |
| *** medberry is now known as med_out | 22:52 | |
| *** nerdstein has quit IRC | 22:53 | |
| *** dendrobates is now known as dendro-afk | 22:55 | |
| *** Binbin has quit IRC | 22:57 | |
| *** Binbin has joined #openstack | 22:57 | |
| vishy | retr0h: not that insane. We're discussing other opitions at the summit | 22:58 |
| *** clauden__ has joined #openstack | 22:59 | |
| *** clauden_ has quit IRC | 22:59 | |
| *** Jamey_ has quit IRC | 22:59 | |
| *** magg has quit IRC | 23:00 | |
| *** nethoopty has joined #openstack | 23:02 | |
| *** miclorb_ has joined #openstack | 23:05 | |
| *** pmyers has joined #openstack | 23:07 | |
| *** dragondm has quit IRC | 23:09 | |
| *** dysinger has joined #openstack | 23:09 | |
| *** Razique has quit IRC | 23:11 | |
| *** Razique has joined #openstack | 23:13 | |
| *** rustam has joined #openstack | 23:14 | |
| *** Ryan_Lane has quit IRC | 23:18 | |
| *** v0id has quit IRC | 23:22 | |
| *** cereal_bars has joined #openstack | 23:24 | |
| *** ziyadb has quit IRC | 23:25 | |
| *** cereal_bars has quit IRC | 23:25 | |
| *** ziyadb has joined #openstack | 23:27 | |
| *** Ryan_Lane has joined #openstack | 23:36 | |
| *** Ryan_Lane has quit IRC | 23:37 | |
| *** ldlework has quit IRC | 23:41 | |
| *** jmckenty has joined #openstack | 23:41 | |
| *** Gordonz has quit IRC | 23:43 | |
| *** nethoopty has quit IRC | 23:44 | |
| *** Razique has quit IRC | 23:46 | |
| *** stuntmachine has joined #openstack | 23:48 | |
| *** joonwon has joined #openstack | 23:48 | |
| *** neogenix has quit IRC | 23:50 | |
| *** osier has joined #openstack | 23:52 | |
| *** Ryan_Lane has joined #openstack | 23:53 | |
| *** pmyers has quit IRC | 23:53 | |
| *** jmckenty has quit IRC | 23:53 | |
| *** pmyers has joined #openstack | 23:55 | |
| *** tryggvil_ has quit IRC | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!