| vidd | zykes-, a token is being created in my keystone database, but it still wont let euca do anything | 00:05 |
|---|---|---|
| vidd | =\ | 00:05 |
| vishy | NOVA_API_KEY has nothing to do with euca commands | 00:06 |
| vishy | you only need EC2_ACCESS_KEY and EC2_SECRET_KEY | 00:06 |
| vidd | vishy, what is nova sending to keystone? | 00:07 |
| vidd | i would assume it is sending nova-api stuff | 00:07 |
| vishy | i don't know what you mean | 00:07 |
| vishy | is keystone on the same host as your api server? | 00:08 |
| vidd | zykes-, i see you have added internal services stuff | 00:08 |
| vidd | vishy, it should not matter | 00:08 |
| vidd | if keystone is off, euca works fine | 00:08 |
| vishy | vidd: well you have to set a flag if it isn't | 00:08 |
| zykes- | vidd: it uses internal for dashboard fyi | 00:08 |
| vishy | keystone_ec2_url | 00:08 |
| zykes- | and for others | 00:09 |
| *** rnorwood has joined #openstack | 00:09 | |
| vidd | yes, keystone is on the loacal machine, but im still calling it by IP | 00:09 |
| vidd | i know euca is talking to keystone because a token is getting created in keystone | 00:11 |
| vidd | but keystone is not returning permissions correctly | 00:11 |
| vidd | zykes-, i think your setting up interal services is where i need to look for a solution | 00:12 |
| zykes- | ;p | 00:15 |
| *** Otter768 has joined #openstack | 00:15 | |
| *** nerens has quit IRC | 00:16 | |
| vidd | zykes-, in your script, "testing is only used for the password...right? | 00:17 |
| *** hezekiah_ has quit IRC | 00:18 | |
| *** rnorwood has quit IRC | 00:21 | |
| *** rnorwood has joined #openstack | 00:22 | |
| *** tortuguito is now known as tortu[banhu | 00:24 | |
| zykes- | vidd: yeah | 00:25 |
| vidd | zykes-, how do i stop restart keystone =\ | 00:26 |
| * vidd has GOT to get an init.d for it! | 00:26 | |
| zykes- | :p | 00:29 |
| *** mgw1 has quit IRC | 00:29 | |
| *** Razique has joined #openstack | 00:33 | |
| vidd | zykes-, you spin up keystone manually...what is the command? | 00:34 |
| * vidd has lost it somewhere in his notes | 00:34 | |
| Razique | hi all :D | 00:35 |
| vidd | hello Razique ! | 00:35 |
| Razique | does someone knows how to manage multiple networks per project ? | 00:35 |
| Razique | hey vidd 'sup ? | 00:35 |
| Razique | I'd like to have mutliple networks and vlans per network ; but it doesn't seem to work :/ | 00:35 |
| vishy | Razique: I don't know if anyone has trying that | 00:36 |
| Razique | hey vishy :) | 00:37 |
| vishy | * has tried that | 00:37 |
| Razique | mmm is it possible to have more than one vlan then ? | 00:37 |
| vishy | per project? | 00:37 |
| Razique | no, totally | 00:37 |
| vishy | sure if you use valn manager | 00:37 |
| vishy | * vlan manager | 00:37 |
| vishy | are you trying to have multiple networks per vm? | 00:37 |
| Razique | I'm stuck here | 00:38 |
| Razique | no, I'd like to be able to manage several instances with different vlans | 00:38 |
| vishy | that is vlan manager then | 00:38 |
| Razique | I need to validate a project for a customer (a real one :D ) that involves around 50 instances | 00:38 |
| Razique | there are "zones" that needs to be isolated | 00:39 |
| Razique | vishy: you mean the Linux vlan-manager ? | 00:39 |
| vishy | no | 00:39 |
| vishy | i mean network_manager=VlanManager | 00:39 |
| vishy | instead of FlatDHCPManager | 00:39 |
| Razique | yah, I'm already using the VLAN mode | 00:40 |
| Razique | works great for one VLAN | 00:40 |
| vishy | good, you're set then | 00:40 |
| vishy | :) | 00:40 |
| Razique | but with two, doesn't seem so | 00:40 |
| vishy | each project gets its own vlan | 00:40 |
| vishy | do you have a second project? | 00:40 |
| Razique | yup | 00:40 |
| vishy | and you have two networks defined in the db | 00:41 |
| Razique | http://paste.openstack.org/show/3096/ | 00:41 |
| vishy | with separate vlans? | 00:41 |
| Razique | yup | 00:41 |
| Razique | yup | 00:41 |
| Razique | :D | 00:41 |
| Razique | the instances into the first VLAN work great (can be reached, etc…) | 00:41 |
| Razique | while instances created with the second VLAN cannot be reached | 00:41 |
| vishy | are they getting dhcp? | 00:42 |
| Razique | while the vlan has been created, the bridge, etc.. | 00:42 |
| Razique | yup | 00:42 |
| vishy | where are you trying to reach them from? | 00:42 |
| Razique | http://paste.openstack.org/show/3097/ | 00:42 |
| vishy | are you sure that you have trunked vlan 50 on the ports? | 00:42 |
| Razique | vishy: both are pinged from the nova-network server | 00:43 |
| vishy | there is just one nova-network right? | 00:43 |
| Razique | the switch is a simple one, with no mamangement at all | 00:43 |
| Razique | yup | 00:43 |
| vishy | can you pastie me the output of ifconfig on nework server | 00:43 |
| Razique | the L2 switchs let pass all packets, whatever the VID might be | 00:44 |
| vishy | as well as ip addr show? | 00:44 |
| Razique | http://paste.openstack.org/show/3098/ | 00:44 |
| Razique | http://paste.openstack.org/show/3099/ | 00:44 |
| Razique | thanks :) | 00:44 |
| vishy | and you are sure that the vm is up | 00:44 |
| vishy | you can vnc console etc.? | 00:44 |
| zykes- | wazzup vidd ? | 00:45 |
| Razique | vishy: it's up but the last line is "cloud-init-nonet waiting 120 seconds for a network device." | 00:46 |
| vidd | keystone & returns an error =\ how do you launch it? | 00:46 |
| vishy | Razique: hmm | 00:46 |
| vishy | Razique: so it isn't dhcping properly | 00:46 |
| Razique | vishy: here is the boot log http://paste.openstack.org/show/3100/ | 00:46 |
| vidd | keystone & returns an error =\ how do you launch it? @ zykes- | 00:47 |
| vishy | Razique: is your compute node a separate host? | 00:47 |
| Razique | vishy: yup =) | 00:47 |
| vishy | ok so the dhcp is failing | 00:47 |
| vishy | are both of your vms on the other host? | 00:47 |
| Razique | yes | 00:48 |
| vishy | Razique: can you do a ps aux | grep dnsmasq and paste it | 00:48 |
| *** stewart has quit IRC | 00:48 | |
| Razique | http://paste.openstack.org/show/3101/ | 00:49 |
| Razique | first one : node | 00:49 |
| Razique | second one : controller | 00:49 |
| *** stewart has joined #openstack | 00:49 | |
| vishy | Razique: that all looks correct | 00:51 |
| vishy | can you pastie me iptables -L -n -v on the compute node? | 00:51 |
| Razique | sure :) | 00:51 |
| Razique | http://paste.openstack.org/show/3102/ | 00:51 |
| tjoy | ASDF | 00:51 |
| tjoy | oops | 00:51 |
| tjoy | guess it works | 00:52 |
| vishy | Razique: weird | 00:52 |
| vishy | Razique: i don't see the dhcp rules getting hit | 00:52 |
| Razique | yah, neither do I | 00:53 |
| Razique | u mean that line ? 0 0 nova-compute-inst-328 all -- * * 0.0.0.0/0 192.168.3.3 | 00:53 |
| vishy | Razique: no | 00:54 |
| vishy | 0 0 ACCEPT udp -- * * 192.168.3.1 0.0.0.0/0 udp spt:67 dpt:68 | 00:54 |
| vishy | Razique: if you boot another instance in the first (vlan100) project does it work? | 00:54 |
| Razique | ok, let's see | 00:54 |
| Razique | yup it's working :) | 00:55 |
| Razique | I've now 2 instances into VLAN100 | 00:55 |
| Razique | and both can be pinged | 00:55 |
| vishy | check the output from the iptables command again | 00:56 |
| vishy | 0 0 ACCEPT udp -- * * 192.168.2.1 0.0.0.0/0 udp spt:67 dpt:68 | 00:56 |
| vishy | are there packets on that line now? | 00:56 |
| Razique | 2 674 ACCEPT udp -- * * 192.168.2.1 0.0.0.0/0 udp spt:67 dpt:68 | 00:57 |
| Razique | looks so yah | 00:57 |
| vishy | ok | 00:57 |
| vishy | so that one is working | 00:57 |
| vishy | check syslog on controller | 00:57 |
| Razique | ok wait sorry | 00:57 |
| Razique | the rules exists into two chains | 00:58 |
| Razique | 0 0 ACCEPT udp -- * * 192.168.2.1 0.0.0.0/0 udp spt:67 dpt:68 | 00:58 |
| Razique | nova-compute-inst-322 | 00:58 |
| Razique | nova-compute-inst-329 | 00:58 |
| vishy | sure that is fine | 00:58 |
| vishy | the new one is working | 00:58 |
| Razique | http://paste.openstack.org/show/3103/ | 00:58 |
| vishy | look for messages in syslog about dnsmasq | 00:58 |
| vishy | (on controller) | 00:59 |
| Razique | here vishy http://paste.openstack.org/show/3103/ : | 00:59 |
| vishy | oh thx | 00:59 |
| Razique | mmmm DHCPDISCOVER(br50) 02:16:3e:2f:f9:b1 no address available | 00:59 |
| vishy | yeah | 00:59 |
| vishy | there is your problem | 00:59 |
| vishy | so can you check --dhcp-hostsfile=/var/lib/nova/networks/nova-br50.conf | 01:00 |
| vishy | and see if that eth is listed | 01:00 |
| Razique | the file is empty | 01:01 |
| vishy | aha! | 01:01 |
| vishy | ok so are there errors in nova-network | 01:02 |
| Razique | i don't have the flag -dhcp-hostsfile into my conf | 01:02 |
| Razique | is it ok ? | 01:02 |
| vishy | maybe about being unable to write to the file? | 01:02 |
| Razique | you mean nova wouldn't be able to write the file ? | 01:03 |
| vishy | that is what i'm wondering | 01:03 |
| vishy | maybe if you ran it as root once and then switched back to a regular user | 01:03 |
| vishy | you can just delete the file | 01:04 |
| vishy | killall dnsmasq | 01:04 |
| Razique | I run all commands as root actually | 01:04 |
| vishy | restart nova-network | 01:04 |
| vishy | that should force it to be recreated | 01:04 |
| Razique | yup the file hs been recreated | 01:04 |
| Razique | but remains empty | 01:04 |
| Razique | and nova the fixed ip from the instance has disaperead | 01:05 |
| Razique | http://paste.openstack.org/show/3104/ | 01:06 |
| Razique | weird :/ | 01:06 |
| vidd | zykes-, you still there? | 01:07 |
| *** freeflying has quit IRC | 01:11 | |
| zykes- | vidd: yeah | 01:11 |
| zykes- | atm ;) | 01:11 |
| Razique | vishy: I found "num_networks : 1 from (pid=6058) wait /usr/lib/python2.6/dist-packages/nova/service.py:355" | 01:11 |
| Razique | is there a link with the issue ? | 01:11 |
| zykes- | trying to come up with an idea on how to solve my dns api & instance stuff | 01:11 |
| zykes- | got clues vishy ? | 01:11 |
| vidd | zykes-, in that script you pasted me http://paste.pocoo.org/show/503590/ is the novaadmin in your nova data dase ? | 01:11 |
| Razique | vishy: also 2011-11-06 02:10:05,137 DEBUG nova [-] vlan_start : 100 from (pid=6058) wait /usr/lib/python2.6/dist-packages/nova/service.py:355 | 01:13 |
| zykes- | if i remember correctly vidd it's not needed, it was something i was testing in connection to dashboard | 01:13 |
| vidd | ic | 01:13 |
| vidd | its there and works for you...im going to leave it for now (since its attached to the internal services tenant =] | 01:14 |
| zykes- | ;p | 01:15 |
| zykes- | int_svcs is just a place to put stuff like glance towards swift | 01:15 |
| zykes- | etc | 01:15 |
| zykes- | vishy: ? | 01:15 |
| vishy | Razique: interesting i don't now if that is important or not | 01:17 |
| vishy | * know | 01:17 |
| vishy | Razique: try terminating the instance and launching another one? | 01:18 |
| vishy | when it allocates the ip it should be writing it | 01:18 |
| vishy | Razique: you don't have multi-host set on the network right? | 01:18 |
| Razique | noope | 01:18 |
| Razique | --multi_host=False into both nova.conf | 01:18 |
| vishy | Razique: i'm trying to figure out what could stop it from writing | 01:19 |
| vishy | Razique: and set to false in the db to? | 01:19 |
| vishy | *too | 01:19 |
| vidd | well...no errors...but no availability zones either =\ | 01:19 |
| Razique | vishy: yup http://paste.openstack.org/show/3105/ | 01:20 |
| vishy | Razique: the only way i can think of it being skipped is if multi_host=True | 01:21 |
| Razique | and after the new spawnning, again http://paste.openstack.org/show/3106/ | 01:21 |
| Razique | vishy: mm so should I change something somewhere ? | 01:22 |
| vishy | network_get_associated_fixed_ips(context, network_id): | 01:23 |
| vishy | might be interesting to do a nova-manage shell python | 01:23 |
| vishy | from nova import context | 01:23 |
| vishy | from nova import db | 01:23 |
| Razique | I'm no programmer at all (yet :p ) | 01:23 |
| vishy | db.network_get_associated_fixed_ips(context.get_admin_context(), <network_id>) | 01:24 |
| vishy | and see if you get anything | 01:24 |
| *** julian_c has quit IRC | 01:24 | |
| vishy | that is the only other thing that could be going wrong | 01:24 |
| vishy | that i can tell | 01:24 |
| vishy | Razique: are you running diablo? | 01:25 |
| Razique | vishy: yup | 01:25 |
| Razique | 2011.3 | 01:25 |
| vishy | where did you get it? | 01:25 |
| vishy | stable/diablo branch? | 01:25 |
| Razique | ppa | 01:25 |
| Razique | yah | 01:25 |
| vishy | hmm there were a couple bugs in the ppa | 01:25 |
| Razique | erf | 01:26 |
| vishy | Razique: can you check your fixed ips table and make sure they are associated with the proper network | 01:27 |
| *** mmetheny_ has joined #openstack | 01:28 | |
| *** mmetheny has quit IRC | 01:28 | |
| *** mmetheny_ is now known as mmetheny | 01:28 | |
| Razique | yes they are http://paste.openstack.org/show/3107/ | 01:28 |
| vishy | also check br100.conf and make sure the the ips aren't showing up there | 01:28 |
| Razique | http://paste.openstack.org/show/3108/ | 01:29 |
| Razique | the br100.conf only contains the two leases | 01:29 |
| Razique | and we see on nova-network start the creation of the right network http://paste.openstack.org/show/3109/ | 01:31 |
| *** esquilinho has joined #openstack | 01:31 | |
| * Razique hopes the SQL query doesn't have a LIMIT=1 :D | 01:31 | |
| vishy | Razique: can you check the virtual_interfaces table for weirdness? | 01:32 |
| zykes- | vishy: will essex features a replaced netstack ? | 01:32 |
| zykes- | by quantum and others | 01:32 |
| vishy | zykes-: unknown | 01:32 |
| vishy | depends on how stable netstack is | 01:32 |
| Razique | vishy: http://paste.openstack.org/show/3110/ | 01:33 |
| vishy | reword | 01:33 |
| zykes- | hope it will be :ø | 01:33 |
| vishy | quantum will be shipped, nova may or may not default to it | 01:33 |
| zykes- | and not like keystone is now :p | 01:33 |
| *** livemoon has joined #openstack | 01:33 | |
| Razique | zykes-: hehe | 01:34 |
| *** tortu[banhu has quit IRC | 01:34 | |
| vidd | vishy, is there a launch target date? | 01:34 |
| vishy | vidd: for essex? | 01:34 |
| vidd | yes | 01:34 |
| vishy | early april | 01:34 |
| vishy | the date should be on the wiki | 01:34 |
| livemoon | hi,all | 01:35 |
| vishy | Razique: I'm super confused at why your br50.conf is empty | 01:35 |
| vishy | i would start by throwing some logging statements in linux_net | 01:35 |
| vidd | i dont understand why these launch dtae seem to be "right before ubuntu" | 01:35 |
| Razique | vishy: well yah, it's pretty weird :/ | 01:35 |
| vishy | to try and figure out why it isn't creating the file | 01:35 |
| vidd | hello livemoon | 01:35 |
| vishy | vidd: so we can get into 12.04 :) | 01:36 |
| livemoon | I miss you | 01:36 |
| Razique | hey lionel :) | 01:36 |
| vishy | if you want to give me access I could login and try to find the issue | 01:36 |
| Razique | vishy: what "i would start by throwing some logging statements in linux_net" means ? | 01:36 |
| vidd | vishy, if you want to get into an Ubuntu launch AND work properly with repo-installs, then there should be at least a month | 01:36 |
| vishy | vidd: :) | 01:37 |
| vishy | vidd: we work closely with ubuntu | 01:37 |
| vidd | especially since this next ubuntu is going to be a LTS | 01:37 |
| zykes- | vidd: get RH on board as well ;p | 01:37 |
| vidd | zykes-, redhat scares me =] | 01:37 |
| zykes- | vidd: why ? :p | 01:38 |
| vidd | "better dead than red" is in my blood =] | 01:38 |
| zykes- | vidd: heh | 01:38 |
| * vidd grew up during the cold war | 01:38 | |
| vidd | =] | 01:38 |
| zykes- | well it's still the easiest thing to sell to big companies here where i am so | 01:39 |
| vidd | zykes-, well...im not getting "400" errors anymore | 01:40 |
| vidd | now im getting 401's =\ | 01:40 |
| Razique | vidd: 401 : auth errors | 01:40 |
| vidd | yeah | 01:40 |
| zykes- | vidd: for what ? | 01:41 |
| vidd | so euca is now talking with keystone, they just dont like what each other is saying =\ | 01:41 |
| vidd | well...euca-describe-availability-zones returns nothing | 01:41 |
| vidd | euca-allocate-address returns 401 | 01:42 |
| vidd | i havent even TRIED to add images yet=\ | 01:42 |
| vidd | or keypairs...or anything else =] | 01:43 |
| vidd | This server could not verify that you are authorized to access the document you requested. Either you supplied the wrong credentials (e.g., bad password), or your browser does not understand how to supply the credentials required. | 01:44 |
| *** cowmix has quit IRC | 01:44 | |
| vidd | what is the command to sync keystone's database? | 01:49 |
| vidd | db_sync? | 01:50 |
| vidd | hey! they added "keystone control" | 01:54 |
| vidd | this will help considerably! | 01:54 |
| vidd | we REALLY need this added to the documentation | 01:54 |
| *** stewart has quit IRC | 01:56 | |
| *** Hakon|mbp has quit IRC | 01:57 | |
| vidd | zykes-, do you have "--keystone_ec2_url=http://$keystone_IP:5000/v2.0/ec2tokens" in your nova config? | 01:59 |
| zykes- | yeah | 02:00 |
| zykes- | but not $keystone_ip | 02:00 |
| zykes- | ah, like that correct | 02:00 |
| vidd | replace $keystone_IP with the actual IP of course | 02:00 |
| vidd | zykes-, do you have both a "token" and a "tokens" table in your keystone database? | 02:04 |
| *** freeflying has joined #openstack | 02:06 | |
| vidd | zykes-, im so close to getting this to work i can smell it! | 02:07 |
| vidd | i just need to figure out why the request is coming back unauthorized | 02:07 |
| zykes- | no to the token question vidd | 02:08 |
| *** esquilinho is now known as o_ultimo_do_moic | 02:09 | |
| *** freeflyi1g has joined #openstack | 02:12 | |
| *** rods has quit IRC | 02:12 | |
| Razique | gotta sleep | 02:19 |
| Razique | cya guys | 02:19 |
| Razique | :) | 02:19 |
| zykes- | ;) | 02:19 |
| Razique | :) | 02:20 |
| *** Razique has left #openstack | 02:20 | |
| vidd | zykes-, any ideas? | 02:22 |
| zykes- | vidd: what is it you do when it fails ? | 02:26 |
| *** osier has joined #openstack | 02:27 | |
| vidd | right now euca-allocate-address --debug | 02:27 |
| zykes- | vidd: check nova-api.log | 02:27 |
| zykes- | and see if it fails there | 02:28 |
| vidd | zykes-, http://dpaste.com/648062/ | 02:29 |
| zykes- | vidd: try | 02:30 |
| zykes- | keystone-manager role grant Admin vidd | 02:30 |
| vidd | zykes you ROCK =] | 02:31 |
| vidd | we have a WINNER!!!!! =\ | 02:31 |
| vidd | so now all my users will have to be "admin" | 02:32 |
| zykes- | i had forgotten that in my sample ;p | 02:32 |
| zykes- | vidd: only tenant admins should have it ;p | 02:32 |
| vidd | "userof nova" =] | 02:32 |
| zykes- | now dashboard ? ;) | 02:33 |
| vidd | yes...on to dashboard =] | 02:34 |
| vidd | wait...no...need to get VNCProxy to quit failing | 02:34 |
| vidd | it wants noVNC | 02:34 |
| vidd | but ubuntu does not seem to offer it | 02:34 |
| *** Otter768 has quit IRC | 02:35 | |
| vidd | which is starnge since ubutu makes nova-vncproxy get installed on my sytse | 02:35 |
| *** Moltar has quit IRC | 02:36 | |
| *** vernhart has quit IRC | 02:41 | |
| *** stewart has joined #openstack | 02:45 | |
| vidd | vishy, you still here? | 02:47 |
| *** misheska has quit IRC | 02:48 | |
| *** osier has quit IRC | 02:50 | |
| vidd | zykes-, ive already given up on noVNC | 03:02 |
| vidd | =] | 03:02 |
| vidd | moving on to dashboard | 03:02 |
| *** o_ultimo_do_moic is now known as tortux | 03:06 | |
| vidd | zykes-, in theory....i can copy the entire .dashboard venv from one computer and just drop it carte blanche onto another machine...right? | 03:12 |
| *** tortux has quit IRC | 03:17 | |
| rmk | vidd: I haven't had luck with vncproxy either | 03:22 |
| rmk | I had it working once and never since then | 03:23 |
| vidd | rmk, ive spent 2 weeks (and three format/reinstalls) just getting keystone and euca to work =\ | 03:23 |
| vidd | now dashboard dont want to co-operate | 03:24 |
| vidd | [and here i left my shocked face @ work] | 03:25 |
| vidd | i cant wait to see what happens when i try to get an image loaded into glance | 03:25 |
| vidd | im half-expecting a small nuclear device to go off | 03:26 |
| vidd | zykes-, you still here? | 03:26 |
| *** crescendo has quit IRC | 03:29 | |
| *** littleidea has quit IRC | 03:31 | |
| *** rsampaio has joined #openstack | 04:15 | |
| *** emid has quit IRC | 04:22 | |
| vishy | wow | 04:40 |
| vishy | devstack uses an alternate vncproxy right now | 04:40 |
| vishy | becaue nova-vncproxy doesn't work with chrome | 04:40 |
| vishy | we're going to clean it up and propose it back into trunk | 04:40 |
| vidd | novvncproxy doesnt work with nova | 04:48 |
| vidd | =] | 04:48 |
| vidd | sorry noVNC | 04:49 |
| *** hezekiah_ has joined #openstack | 04:49 | |
| vidd | nova-vncproxy doesnt work with noVNC | 04:49 |
| vidd | noVNC seams to be broken | 04:50 |
| *** misheska has joined #openstack | 04:57 | |
| *** obino has quit IRC | 05:02 | |
| vidd | vishy, you still here? | 05:12 |
| *** crescendo has joined #openstack | 05:21 | |
| *** crescendo has joined #openstack | 05:21 | |
| *** rnorwood has quit IRC | 05:25 | |
| *** dnjaramba has quit IRC | 05:31 | |
| vishy | vidd: maybe | 05:36 |
| vishy | vidd:: novnc works fine if you run it as root and you aren't using chrome | 05:37 |
| vishy | chrome 14 decided to break it | 05:37 |
| vidd | vishy, i cant seem to get it to install and nova-vncproxy complains its not there | 05:37 |
| vishy | interesting | 05:38 |
| vidd | vishy, also...you work on the devstack team right? | 05:38 |
| vishy | yup | 05:38 |
| vidd | the person you get your tty from, smoser recommeds using his new immage https://launchpad.net/cirros | 05:39 |
| vidd | ive used it and it works great | 05:39 |
| vidd | i have the default username and password for it....he doesnt remember what he set up for the old one anymore | 05:40 |
| vidd | =] | 05:40 |
| *** clopez has quit IRC | 05:47 | |
| vishy | cool | 05:52 |
| vishy | the old one is root/password | 05:52 |
| vidd | heh | 05:52 |
| vidd | the new one is "cirros" / "cubswin:)" [the smiley is part of the password] | 05:53 |
| vishy | cubs! | 05:53 |
| vidd | so i take it there is an update coming ? =] | 05:54 |
| vidd | or are you NOT a cubs fan =] | 05:54 |
| *** wilmoore has joined #openstack | 06:08 | |
| vidd | horizon definantly does NOT like mysql =\ | 06:31 |
| vidd | i assume that when you run the "dashboard/manage.py syncdb" your supposed to get propmted to set up a username and password | 06:33 |
| *** lionel has quit IRC | 06:35 | |
| *** lionel has joined #openstack | 06:36 | |
| vidd | getting "ImportError: Could not import settings 'dashboard.settings' (Is it on sys.path?): No module named dashboard.settings" when trying to run dashboard thru apache | 06:36 |
| vidd | i am assuming that this means that the manage.py syncdb is supposed to create the dashboad.settings file or database record that apache file is looking for =\ | 06:38 |
| vidd | and with dashboard set to use mysql, this does not happen | 06:39 |
| *** crescendo has quit IRC | 07:14 | |
| *** sebastianstadil has quit IRC | 07:27 | |
| *** MarkAt2od has joined #openstack | 07:40 | |
| *** livemoon has left #openstack | 07:41 | |
| vidd | i have figured out the issue with dashboard =] | 07:42 |
| vidd | well...most of it anyway =] | 07:42 |
| *** MarkAtwood has quit IRC | 07:42 | |
| *** MarkAt2od has quit IRC | 07:45 | |
| *** MarkAtwood has joined #openstack | 07:45 | |
| *** katkee has joined #openstack | 07:49 | |
| *** Razique has joined #openstack | 07:58 | |
| *** nerens has joined #openstack | 07:58 | |
| *** Razique has quit IRC | 08:02 | |
| *** hezekiah_ has quit IRC | 08:25 | |
| *** alekibango_ has quit IRC | 08:26 | |
| *** alekibango_ has joined #openstack | 08:26 | |
| *** alekibango_ has quit IRC | 08:27 | |
| *** alekibango_ has joined #openstack | 08:27 | |
| *** alekibango has joined #openstack | 08:28 | |
| *** MarkAtwood has quit IRC | 08:44 | |
| *** jedi4ever has joined #openstack | 08:53 | |
| *** sebastianstadil has joined #openstack | 08:57 | |
| *** TheOsprey has joined #openstack | 09:01 | |
| *** sebastianstadil has quit IRC | 09:07 | |
| *** sebastianstadil has joined #openstack | 09:08 | |
| *** sebastianstadil has quit IRC | 09:09 | |
| *** wilmoore has quit IRC | 09:14 | |
| *** TheOsprey_ has joined #openstack | 09:22 | |
| *** TheOsprey has quit IRC | 09:23 | |
| *** imagelife has joined #openstack | 09:45 | |
| *** ChrisAM has quit IRC | 09:48 | |
| *** xinkeT has joined #openstack | 09:49 | |
| *** Teknix has quit IRC | 09:50 | |
| *** ChrisAM1 has joined #openstack | 09:52 | |
| *** nerens has quit IRC | 10:11 | |
| *** po has joined #openstack | 10:14 | |
| *** Ryan_Lane has quit IRC | 10:30 | |
| *** dnjaramba has joined #openstack | 10:42 | |
| zykes- | around vidd ? | 11:01 |
| vidd | zykes-, yep | 11:01 |
| vidd | i have beaten keystone and dashboard into submission | 11:01 |
| zykes- | :D | 11:02 |
| zykes- | horizon or ? | 11:02 |
| vidd | of course | 11:02 |
| zykes- | hint though | 11:02 |
| zykes- | keep track of what commit you have installed | 11:02 |
| zykes- | so if it breaks you can revert it ;) | 11:03 |
| vidd | the tenat and user links throw errors....i think there is an issue with keystone vs. keystoneclient | 11:03 |
| vidd | vnc doesnt work | 11:03 |
| vidd | and instace statuses dont appear to refresh corretly | 11:04 |
| zykes- | hmm | 11:04 |
| zykes- | endpoints are correct ? | 11:04 |
| vidd | the endpoints are correct | 11:04 |
| vidd | and the reboot does actually work (when viewed via euca | 11:05 |
| vidd | oh...and i cant seem to ping or ssh into images | 11:05 |
| zykes- | you can't ping a image ;p | 11:05 |
| zykes- | and instance on the other hand | 11:05 |
| zykes- | you mean to it's floating ipo ? | 11:06 |
| vidd | im tired...its almost my bedtinme | 11:06 |
| vidd | yeah | 11:06 |
| zykes- | go to bed dude | 11:06 |
| zykes- | come back with a freshened set of eyes | 11:06 |
| vidd | and a 2x4, a chisel, and a BFH! | 11:07 |
| vidd | euca cant seem to see the image in glance =\ | 11:08 |
| *** rsampaio has quit IRC | 11:08 | |
| *** rsampaio has joined #openstack | 11:09 | |
| vidd | ok...bed time | 11:13 |
| *** vidd is now known as vidd-away | 11:13 | |
| *** ejat has joined #openstack | 11:51 | |
| *** sebastianstadil has joined #openstack | 12:01 | |
| *** sebastianstadil has quit IRC | 12:12 | |
| *** po has quit IRC | 12:13 | |
| *** freeflyi1g has quit IRC | 12:28 | |
| *** rsampaio_ has joined #openstack | 12:44 | |
| *** rsampaio has quit IRC | 12:44 | |
| *** Hakon|mbp has joined #openstack | 12:49 | |
| *** rsampaio_ has quit IRC | 12:51 | |
| *** rsampaio has joined #openstack | 12:54 | |
| *** mmetheny has quit IRC | 13:28 | |
| *** mmetheny_ has joined #openstack | 13:28 | |
| *** livemoon has joined #openstack | 13:37 | |
| *** familyguy has joined #openstack | 13:44 | |
| familyguy | Thinking of setting up a "lab" to fool around with cloud computing. Got a rack of HP DL360s to play with. Can anyone recommend a good resource for a total n00b to cloud computing to go for a good primer? The docs on the openstack website are rather sparse. | 13:46 |
| guaqua | there's a book on setting up openstack | 13:46 |
| guaqua | i ordered it myself just yesterday | 13:46 |
| guaqua | "Deploying OpenStack" | 13:47 |
| guaqua | that might be of interest | 13:47 |
| *** livemoon has quit IRC | 13:48 | |
| familyguy | guaqua: Cool. That's exactly what I'm looking for. | 13:48 |
| guaqua | i've been dabbling with swift | 13:48 |
| guaqua | just through virtualbox | 13:48 |
| familyguy | Wow, it's only 79 pages long. :) An O'Reilly magazine. heh | 13:49 |
| familyguy | I've been "given" a rack of 24 brand spanking new 12-core boxes so I figured that I'd do something useful with them rather than have interns mining bitcoins on 'em while they burn in for a few months. :) | 13:50 |
| guaqua | :D | 13:51 |
| guaqua | sounds reasonable! | 13:51 |
| guaqua | dell has software for automatically installing openstack on their servers and it might work for you, too | 13:52 |
| guaqua | but i'm a bit skeptical about it | 13:52 |
| guaqua | https://github.com/dellcloudedge | 13:52 |
| guaqua | it's basically a layer above chef to do automatic installation and provisioning | 13:52 |
| familyguy | I'll take a look. These are HP DL360's. | 13:53 |
| guaqua | the skepticism comes from seeing a bit too much ruby there | 13:53 |
| guaqua | the last time i stumbled into ruby was with cloudfoundry. and boy that code is crap | 13:53 |
| *** livemoon has joined #openstack | 13:53 | |
| familyguy | Do userspace apps need to be "aware" of openstack or does all that stuff happen under the covers with the apps being blissfully ignorant of where they're actually running from? | 13:54 |
| guaqua | what do you mean? on virtual machines? | 13:55 |
| guaqua | i'm thinking not. i don't know if there's a way to see instance specific data from within the instance like with ec2 | 13:55 |
| familyguy | Yep, that's what I mean. | 13:56 |
| familyguy | As you can see, I need to crack the books. I've always kinda rolled my eyes at the cloud thing, but now I've got some time and some resources...so I figured...time to see what this is all about. | 13:56 |
| guaqua | "The instance acquires the instance specific configuration from Nova-compute by connecting to a meta data interface running on 169.254.169.254." (http://cssoss.wordpress.com/2011/04/27/openstack-beginners-guide-for-ubuntu-11-04-image-management/) | 13:57 |
| guaqua | dunno how current this information is, but it sounds reasonable | 13:57 |
| guaqua | i only have experience with ec2, none with nova | 13:57 |
| *** rbergeron has quit IRC | 13:58 | |
| *** rbergeron has joined #openstack | 13:58 | |
| guaqua | in general you should stick with images someone else has created, it saves a lot of work | 13:58 |
| guaqua | if that's possible, that is | 13:58 |
| guaqua | at least this is the case with ec2, i have to add again :) | 13:59 |
| *** livemoon has quit IRC | 14:01 | |
| familyguy | There seems to be a lot of buzz with using openstack+ubuntu, but I'm more of a RHEL/CentOS/ScientificLinux kinda person. | 14:05 |
| guaqua | there should be stuff for CentOS too | 14:06 |
| guaqua | but i dunno if you actually end up installing fro | 14:08 |
| guaqua | m source then | 14:08 |
| guaqua | maybe someone else here knows? | 14:08 |
| familyguy | stackops looks promising, but it's based on ubuntu. :( | 14:09 |
| guaqua | that's really interesting | 14:12 |
| *** jedi4ever has quit IRC | 14:19 | |
| *** nRy has joined #openstack | 14:20 | |
| familyguy | not sure if you've seen this: http://ubuntuone.com/p/19vd/ freebie Howto doc...using ubuntu again. Seems pretty nice. Gonna curl up with that for a while to get my feet wet. | 14:27 |
| *** TheOsprey_ has quit IRC | 14:30 | |
| *** TheOsprey_ has joined #openstack | 14:35 | |
| nRy | Anyone know how to use OpenStack to control EC2 instances??? | 14:37 |
| *** familyguy has quit IRC | 14:37 | |
| *** cmagina has quit IRC | 14:41 | |
| *** f-yang has quit IRC | 14:41 | |
| *** TheOsprey_ has quit IRC | 14:44 | |
| *** TheOsprey_ has joined #openstack | 14:47 | |
| *** ChrisAM1 is now known as ChrisAM | 14:53 | |
| *** vidd-away is now known as vidd | 14:58 | |
| *** jeblair has quit IRC | 15:17 | |
| vidd | nRy, what do you need help with? | 15:27 |
| *** imagelife has left #openstack | 15:28 | |
| *** ejat has quit IRC | 15:28 | |
| nRy | vidd: well I am trying to figure out the best way to build a custom control system, a kind of dashboard to control my cloud resources | 15:30 |
| vidd | nRy, is there a reason you dont use dashboard? | 15:31 |
| nRy | trying to figure out the best technology to use, ie: Chef, Puppet, or OpenStack | 15:31 |
| nRy | the EC2 dashboard? | 15:31 |
| nRy | or the OpenStack one? | 15:31 |
| vidd | the openstack one | 15:32 |
| nRy | I have been watching the development of it for a while and I am seriously considering it | 15:32 |
| vidd | or...if you are using someone else's cloud and not your own, then perhaps you should look at hybridfox | 15:33 |
| nRy | I am using a mix of my own cloud and others | 15:33 |
| nRy | so it is both | 15:33 |
| vidd | nRy, i just got the dashboar up and running this morning and i ilike it | 15:33 |
| nRy | can it be customized much? | 15:34 |
| nRy | basically what I need to do is give individual users of our cloud resources the ability to control specific instances | 15:34 |
| vidd | if you know how to, ANYTHING can be customized.... | 15:35 |
| nRy | we have a dashboard of our own that we have been building using Joomla user accounts | 15:35 |
| vidd | its a django project | 15:35 |
| nRy | but I would like to integrate something like the Openstack dashboard into it | 15:35 |
| *** Razique has joined #openstack | 15:35 | |
| nRy | django? | 15:35 |
| nRy | which is a django project? | 15:36 |
| vidd | like a joomla project or a drupal project | 15:36 |
| vidd | but instead of php it uses python | 15:37 |
| nRy | what does? | 15:37 |
| vidd | django | 15:37 |
| nRy | I am familar with django but what does django have to do with Openstack or what we were talking about? | 15:38 |
| vidd | the dashboard is django | 15:38 |
| nRy | the Openstack dashboard is made with Django? | 15:39 |
| nRy | really? | 15:39 |
| vidd | yes | 15:39 |
| nRy | oh ok, cool | 15:39 |
| nRy | so it would be much easier to integrate if our website was using Django instead of Joomla? | 15:39 |
| vidd | how so? | 15:39 |
| vidd | here is what i recomend: | 15:40 |
| vidd | look into keystone...]the authentication central hub] | 15:40 |
| Razique | hi guys | 15:41 |
| nRy | vidd: can you send me a link to keystone? | 15:41 |
| vidd | nRy, get the devstack script to see the total package: | 15:42 |
| nRy | for the dashboard? | 15:42 |
| vidd | https://github.com/cloudbuilders | 15:43 |
| vidd | no...not JUST dashboard...the total package... | 15:43 |
| vidd | nova, glance, swift, quantum, keystone, and the dashboard | 15:43 |
| vidd | all working together in harmony | 15:44 |
| vidd | =] | 15:44 |
| *** Razique has quit IRC | 15:44 | |
| nRy | ah yes ok, thanks | 15:45 |
| nRy | will check out the whole picture | 15:45 |
| vidd | now nRy since you want to take an existing database of users (your joomla people) you will want to look at the [ldap plus others] authentication | 15:47 |
| vidd | in keystone | 15:47 |
| vidd | personally, i think it would be easier to modify your joomla to accept keystone authentcation than keystone/nova/etc to accept joomla auth .... but thats just me =] | 15:49 |
| *** rsampaio has quit IRC | 15:58 | |
| *** rsampaio has joined #openstack | 16:00 | |
| *** MarkAtwood has joined #openstack | 16:08 | |
| *** crescendo has joined #openstack | 16:20 | |
| *** crescendo has joined #openstack | 16:20 | |
| *** qazwsx has joined #openstack | 16:36 | |
| *** bsza has joined #openstack | 16:39 | |
| vidd | ok...i attempted to publish an image into glane an failed miserable...but when i did the cloud publish tarball it worked fine.... | 16:39 |
| vidd | can someone point me to a howto for glance that doesn't read like sereo instructions? =\ | 16:40 |
| *** nijaba has quit IRC | 16:43 | |
| vidd | zykes-, my issue with pinging and ssh-ing was because the image was not uploaded properly into glance | 16:43 |
| vidd | that 4 hours of sleep really helped =] | 16:44 |
| *** nijaba has joined #openstack | 16:45 | |
| *** nijaba has joined #openstack | 16:45 | |
| *** bsza has quit IRC | 16:50 | |
| *** TheOsprey__ has joined #openstack | 16:55 | |
| *** TheOsprey_ has quit IRC | 16:58 | |
| *** nerens has joined #openstack | 16:58 | |
| *** jedi4ever has joined #openstack | 17:07 | |
| *** lvaughn has joined #openstack | 17:12 | |
| *** obino has joined #openstack | 17:14 | |
| *** lvaughn_ has quit IRC | 17:15 | |
| *** hadrian has joined #openstack | 17:33 | |
| *** neunon has quit IRC | 17:44 | |
| *** WormMan has joined #openstack | 17:48 | |
| *** MarkAtwood has quit IRC | 17:55 | |
| *** neunon has joined #openstack | 17:56 | |
| *** neunon has joined #openstack | 17:56 | |
| *** TheOsprey__ has quit IRC | 18:12 | |
| *** stevegjacobs has joined #openstack | 18:15 | |
| *** jeblair has joined #openstack | 18:18 | |
| zykes- | vidd: aha | 18:19 |
| zykes- | 4 hours ? :p | 18:20 |
| *** po has joined #openstack | 18:20 | |
| vidd | zykes-, yeah...i got to sleep in =] | 18:20 |
| zykes- | ;p | 18:22 |
| vidd | my dashboard is acting weird though... | 18:24 |
| vidd | pages have to be reloaded.... | 18:25 |
| vidd | restart and other commands dont complete properly | 18:25 |
| vidd | i think its just because the hardware im using is not designed for this | 18:26 |
| vidd | i do need to figure out why the keystone client dashboard is using isnt talking to the keystone database though | 18:26 |
| vidd | the solution to my dashboard.settings proplem was rather interesting.... | 18:28 |
| *** rnorwood has joined #openstack | 18:28 | |
| vidd | i added dummy info to the django_session table and away i went =] | 18:28 |
| *** ldlework has joined #openstack | 18:29 | |
| *** Razique has joined #openstack | 18:33 | |
| vidd | hello Razique | 18:34 |
| Razique | hi dude :) | 18:34 |
| Razique | 'sup | 18:34 |
| Razique | still battling against Keystone ? | 18:34 |
| vidd | i have a near-fully functional stack | 18:34 |
| vidd | keystone has been beaten into submission.... | 18:35 |
| *** rods has joined #openstack | 18:35 | |
| vidd | dashboard has been denied it venv | 18:35 |
| Razique | ok ok | 18:36 |
| Razique | what issues are u habing now ? | 18:36 |
| Razique | I'm myself trying boot from volume | 18:36 |
| Razique | and instance shutdown | 18:36 |
| vidd | just need to learn how to add an image to glance prperly ..... | 18:37 |
| vidd | and the keystoneclient in dashboard is having an issue with keystone's database | 18:37 |
| Razique | you can use nova-manage image* | 18:38 |
| Razique | or glance upload | 18:38 |
| vidd | Razique, its not a "you should use this" type thing....its a "you types this, click that...."type thing | 18:39 |
| vidd | oh...and vnc is hopeless | 18:40 |
| Razique | nova-manage image image_register XXX.img --is_public=T | 18:40 |
| Razique | :) | 18:40 |
| Razique | nova-mange image kernel_register or ramdisk_register | 18:40 |
| vidd | Razique, i tried that....and the image got mangled =] | 18:40 |
| Razique | what mangled means vidd ? | 18:41 |
| vidd | think "dropped in a wood-chipper" | 18:41 |
| Razique | ah :D | 18:42 |
| vidd | the image clamed to have loaded...i could assign IPs, but no ping and no ssh | 18:42 |
| Razique | silly but have u added the Sec-rules ? | 18:45 |
| Razique | icmp and tcp ? | 18:45 |
| vidd | yes | 18:45 |
| *** hadrian has quit IRC | 18:45 | |
| vidd | the initrd | 18:46 |
| vidd | was 0 bites | 18:46 |
| vidd | it was a three-part-image that i must not have properly carved up and put in to glance | 18:47 |
| vidd | kernel, inird, and img | 18:48 |
| Razique | ok | 18:48 |
| Razique | did the image been sent to the node ? | 18:49 |
| vidd | anyway, i re-created the same tarball with cloud-publish-tarball and it worked just fine | 18:49 |
| Razique | check /var/lib/nova/instance/_base | 18:49 |
| Razique | oh ok :) | 18:49 |
| vidd | so i know it was me, not the system =] | 18:49 |
| vidd | i guess the back-to-back 36 hour days finally beat me across the bac of the head =] | 18:51 |
| vidd | anyway...off to fix keystoneclient =] | 18:51 |
| Razique | cool | 18:53 |
| Razique | here it's like, no matter how many times I deploy openstack | 18:53 |
| Razique | I always find new non-working stuff :D | 18:53 |
| Razique | right now, I can't run anymoreinstances | 18:54 |
| Razique | while the config seems good :D | 18:54 |
| vidd | Razique, its because you keep grabbing from "master" brach | 18:54 |
| Razique | well I try to stick to the same version | 18:54 |
| Razique | but in cases it's working | 18:55 |
| Razique | in others, it doesn't | 18:55 |
| Razique | while there are not any specific errors | 18:55 |
| vidd | what im doing is getting all my gits in one place and then cp them to where they will be deployed from | 18:55 |
| stevegjacobs | My stack is broken since an ubuntu update | 18:57 |
| vidd | dashboard or openstackx brings in a borked python-keystoneclient, so im going to find one that works...and then note that it needs to be overwrite the one that is getting auto-imported =] | 18:57 |
| stevegjacobs | seems that no matter what I try, something doesn't work | 18:57 |
| stevegjacobs | right now, I can't assign ip addresses | 18:57 |
| vidd | stevegjacobs, what did you apt-get? | 18:57 |
| vidd | my working stack is apt-get nova-* glance; git keystone, dashboard, openstackx | 18:59 |
| stevegjacobs | It was a couple of days ago, and I don't remember exactly what I did but it launched an update of almost all nova packages | 18:59 |
| stevegjacobs | I think I was trying to add keystone | 18:59 |
| vidd | did you try to add keystone via apt or git? | 19:00 |
| stevegjacobs | apt | 19:00 |
| stevegjacobs | I have a couple of instances that I don't want to lose | 19:01 |
| vidd | keystone from apt does not play well with nova from apt | 19:01 |
| stevegjacobs | things werer running fine until that update but no control since | 19:01 |
| stevegjacobs | Can I install over the top as you have done and not lose what I have? | 19:02 |
| *** obino has quit IRC | 19:02 | |
| vidd | im not doing git-over-apt | 19:02 |
| vidd | im doing git-over-git | 19:03 |
| vidd | basically, im going to rip out the .eg and replace it with an updated version | 19:03 |
| vidd | and i have not tested it yet to see if it will actually change anything =] | 19:04 |
| *** nyeates has joined #openstack | 19:05 | |
| stevegjacobs | vidd - What do you mean? It looks like you installed all nova packages from apt, then git, dashboard, etc from git | 19:05 |
| vidd | if i understand correctly, git-over apt will work...but i wont promise anything....and i would do it on a test machine, not your working model | 19:05 |
| vidd | right... | 19:05 |
| vidd | dashboard is not a nova-* app | 19:06 |
| vidd | nova-api, nova-network, etc | 19:06 |
| *** guaqua2 has joined #openstack | 19:06 | |
| stevegjacobs | Yeah I know keystone and dashboard are not core yet | 19:07 |
| vidd | when i tried to apt-git my keystone, it nuked my system and i just reformatted and started over | 19:07 |
| vidd | its not that they are not core....they just dont work properly with the binaries in the repo's | 19:07 |
| stevegjacobs | I think that's what happened to me. I can't reformat right now tho | 19:08 |
| stevegjacobs | Unless I can figure out some way to safely preserve a couple of instances - one at least | 19:08 |
| vidd | have you changed your paste.ini or your glance files for the keystone integration yet? | 19:08 |
| stevegjacobs | no | 19:08 |
| vidd | then just apt-get autoremove --purge keystone | 19:09 |
| stevegjacobs | I backed up an d tried to go back to the old auth | 19:09 |
| stevegjacobs | but that is n't working right any more either | 19:09 |
| vidd | so you did change the files | 19:09 |
| stevegjacobs | Only nova.conf | 19:10 |
| stevegjacobs | I think I'll try what you are suggesting | 19:10 |
| vidd | stevegjacobs, before you do anything, make a snapshot of your running instances | 19:11 |
| vidd | and prepare for the live migration | 19:11 |
| stevegjacobs | snapshots aren't working for me either | 19:11 |
| vidd | do you have more than one machine in your deployment? | 19:12 |
| stevegjacobs | three - | 19:13 |
| vidd | the instances you do not wish to lose are they running on your controller or on one of the compute nodes? | 19:14 |
| stevegjacobs | I was also trying to use one strictly for storage, but that isn't happening yet. I have one controller and one good working compute node | 19:14 |
| *** katkee has quit IRC | 19:15 | |
| stevegjacobs | Happily - it is on the compute node! | 19:15 |
| vidd | what database are you using? | 19:16 |
| *** tightwork has joined #openstack | 19:16 | |
| stevegjacobs | mysql | 19:16 |
| stevegjacobs | on the controller | 19:16 |
| tightwork | ew wtf am i doing in here | 19:16 |
| *** tightwork has left #openstack | 19:16 | |
| vidd | have you thought to back up your database to the other machine? | 19:17 |
| stevegjacobs | I don't know if it is a good idea, but I installed phpmyadmin on the controller so I could easily see what is going on with the db | 19:17 |
| vidd | hehe stevegjacobs i have phpmyadmin as well =] | 19:17 |
| vidd | now i only have an all-in-one here...but does anyone know if a compute node is running an instance and the controller is rebooted, does the instance die? | 19:18 |
| *** MarkAtwood has joined #openstack | 19:19 | |
| stevegjacobs | no - would that be a good idea? Would it work to point nova-conf to a db on the other machine? | 19:19 |
| vidd | or is it just not accessable until the controller is back up | 19:19 |
| stevegjacobs | good question. | 19:19 |
| stevegjacobs | I need to go finish cooking dinner - be back in a few minutes - thanks for help! | 19:20 |
| vidd | stevegjacobs, the whole reason i chose mysql (and broke my back getting dashboard to work with it) was because i can move my databases anywhere i want if i want =] | 19:21 |
| *** MarkAtwood has quit IRC | 19:25 | |
| *** MarkAtwood has joined #openstack | 19:32 | |
| *** hggdh has quit IRC | 19:34 | |
| *** Micric has joined #openstack | 19:34 | |
| Micric | Good evening everyone | 19:34 |
| Micric | Is anyone able to anser me a small problem. | 19:35 |
| vidd | Micric, ask away...and then we can see if we have an answer =] | 19:37 |
| Micric | Thank you | 19:38 |
| cloudfly | at the very least we'll try to be entertaining. | 19:38 |
| Micric | I am trying to pass up an image, using cloud-publish-image. One the package is prepared it reports back that the login is rejected. | 19:39 |
| vidd | entertainment is easy...answers are hard =] | 19:39 |
| Micric | I am running swift and glance along with Nova but I cannot find in any of the logs which of the processes is/are rejecting the attempt | 19:40 |
| vidd | Micric, did you build the image or did someone else build it? | 19:40 |
| Micric | I did | 19:40 |
| cloudfly | and you have your api credentials loaded in your environment? | 19:41 |
| vidd | how did you build it and how did you import it to your system | 19:41 |
| *** hggdh has joined #openstack | 19:41 | |
| Micric | Cloudfly: As far as I can tell the credentials are loaded and correct | 19:41 |
| cloudfly | euca-describe-images work? | 19:42 |
| cloudfly | or some equivalent | 19:42 |
| vidd | Micric, to verify your creds are loaded, run "env |grep EC2 | 19:42 |
| vidd | if you get nothing returned, then your creds are not loaded | 19:42 |
| Micric | Vidd, I built the image using kvm-img and used used VNC to build | 19:43 |
| Micric | Removed the partition from the image and prepared the kernel and ramdisk | 19:43 |
| Micric | Cloudfly, yes the credenials are correct and describe-images does return with no errors | 19:43 |
| *** obino has joined #openstack | 19:44 | |
| *** neunon has quit IRC | 19:44 | |
| cloudfly | interesting | 19:45 |
| vidd | Micric, did you prepare the image on the same machine you are trying to publish from or on a different machine? | 19:45 |
| Micric | same machine | 19:45 |
| cloudfly | what is the publish command you are using? | 19:45 |
| cloudfly | and does the nova user / group have access to write to the image store? | 19:45 |
| Micric | cloud-publish-image -t image --K <Kernal file> -R <RAMdisk file> i386 Ubuntu11.10.img mybucket | 19:46 |
| *** neunon has joined #openstack | 19:46 | |
| *** neunon has joined #openstack | 19:46 | |
| Micric | Cloudfly, there s the question. I do not know if the nova user has the rights | 19:47 |
| vidd | Micric, do you have keystone auth? | 19:47 |
| Micric | It is installed using apt-get but I have done no configuration | 19:48 |
| cloudfly | hgwell check out bucket path and images path in your nova flag file | 19:48 |
| cloudfly | then do a process list on your nova processes | 19:48 |
| cloudfly | verify the ACLs on those paths match up to a happy writeable state | 19:49 |
| vidd | Micric, before you go about configuring keystone be advised the apt-get version does not work well with dashboard | 19:49 |
| Micric | was looking at dashboard for later, good to know | 19:49 |
| vidd | Micric, i spent the better part of 2 weeks with keystone and dashboard | 19:50 |
| cloudfly | heh i have not played with keystone yet | 19:50 |
| cloudfly | i did look at the code a bit | 19:50 |
| cloudfly | looks like it may be pretty great | 19:50 |
| guaqua | there doesn't seem to be much material about keystone out there yet | 19:50 |
| vidd | cloudfly, keystone is awesome =] | 19:51 |
| vidd | [now that i got it to work] | 19:51 |
| cloudfly | tempted to try my hand at writing a kerb5 plugin for it =P | 19:51 |
| vidd | cloudfly, when you do set up keystone, your userMUST be set as "Admin" role or you cannot do EC2 anything | 19:53 |
| *** ejat has joined #openstack | 19:54 | |
| *** ejat has joined #openstack | 19:54 | |
| cloudfly | oof | 19:55 |
| cloudfly | makes sense though | 19:55 |
| *** Ryan_Lane has joined #openstack | 19:55 | |
| stevegjacobs | vidd - I think I'll install phpmyadmin and mysql on the other machine and copy that | 19:56 |
| vidd | cloudfly, we have an interesting question.... | 19:56 |
| cloudfly | ? | 19:57 |
| vidd | if you ahve a remote compute node running an instance and you restart the controller, does the instance go down or does it stay running untill the controller is back? | 19:57 |
| cloudfly | stays running | 19:57 |
| stevegjacobs | I'd then like to see if I can get keystone/dashborard running on the controller | 19:57 |
| cloudfly | def | 19:57 |
| stevegjacobs | Hi cloudfly | 19:57 |
| cloudfly | howdy | 19:58 |
| stevegjacobs | vidd is asking for me :-) | 19:58 |
| vidd | stevegjacobs, im going to be running an experimennt.... | 19:58 |
| cloudfly | well i've seen bexar clusters fall over and the instances do keep running on the compute nodes | 19:58 |
| vidd | im going to set up keystone and dash on another machine, along with my databases and see if they will work | 19:59 |
| cloudfly | of course even in bexar with the isolated network controller, the trick was to not host the network controller on the fabric controller. | 19:59 |
| cloudfly | nowadays its much better | 19:59 |
| stevegjacobs | The one instance I'm concerned about hasn't stopped though almost everything else broke on me after I (not intentionally) kicked off an ubuntu upgrade | 20:00 |
| cloudfly | oof | 20:00 |
| cloudfly | well i don't know how your network layer is configured. | 20:00 |
| cloudfly | but the compute controllers will likely desync from rabbitmq | 20:01 |
| Micric | cloudfly, I hate to sound stupid but.......where are the flags set for bucket and image? | 20:01 |
| cloudfly | and they won't become responsive to nova again until services are restarted on them | 20:01 |
| cloudfly | Micric /etc/nova.conf usually | 20:01 |
| stevegjacobs | I had originally installed the system with ubuntu 11:10 packages, got it working etc, and then was trying to install keystone when the upgrade happend | 20:02 |
| Micric | How odd, they are not listed in there. I assume that without a flag it sends to /var/lib/nova | 20:02 |
| cloudfly | maybe | 20:02 |
| cloudfly | should be easy enough to look | 20:03 |
| vidd | now....with everything i have learned, i should [key word SHOULD] be able to get a complete stack up and running in less than 4 hours | 20:03 |
| *** misheska has quit IRC | 20:04 | |
| cloudfly | vidd i mean if you have done it before... that seems achievable | 20:04 |
| vidd | does swift need to be on a server with virt capabilities? | 20:04 |
| guaqua | it sounds reasonable, but it should be quicker :/ | 20:04 |
| cloudfly | but i mean i wouldn't operate on that assumption | 20:05 |
| guaqua | vidd: nope | 20:05 |
| cloudfly | vidd it might help if you are using 10 gig interfaces | 20:05 |
| cloudfly | doubt it though | 20:05 |
| vidd | cloudfly, guaqua i'll be billing the boys upstairs 36 hours for the initial setup =] | 20:06 |
| stevegjacobs | I wanted to put swift on a server without virt capabilities too. | 20:06 |
| guaqua | vidd: sounds reasonable ;) | 20:06 |
| *** misheska has joined #openstack | 20:06 | |
| stevegjacobs | haven't got it running yet that way though | 20:06 |
| guaqua | vidd: although, you've already done the work...so billing for it makes sense | 20:06 |
| vidd | they dont want to pay me for the Proof-of Consept development time...none of it | 20:06 |
| guaqua | stevegjacobs: what's the problem? | 20:06 |
| Micric | I can see it pushed to glance and can follow the location for the glance image repository, but ACL shows full access for everyone | 20:06 |
| guaqua | vidd: that's basically bullshit, but very much how business works... | 20:07 |
| *** misheska has quit IRC | 20:07 | |
| stevegjacobs | I have a controller and a compute node, and I want to add swift on a big storage box. | 20:07 |
| vidd | guaqua they are not going to be standing over my shoulders as i deploy so they will basically have to take my word on how long the setup will take | 20:08 |
| vidd | red lobster anyone =] | 20:08 |
| cloudfly | vidd if you are counting hardware provisioning in that count | 20:08 |
| cloudfly | you may want some breathing room | 20:08 |
| guaqua | swift gets redundant when you have more physical machines, one big box is basically just an interface, none of the redundancy | 20:08 |
| cloudfly | especially if you are reliant on other folks for stuff like network reconfig | 20:09 |
| guaqua | although you can run multiple storage nodes on one computer using different drives etc | 20:09 |
| vidd | cloudfly, i will go from bare bones to fully-functional...and it takes as long as it takjes....plas 10 hours minimum | 20:09 |
| Micric | cloudfly, I can see the ACL's are all set to allow full access for everyone. However I can see the images are being bundled off to glance | 20:10 |
| stevegjacobs | guaqua - I think that is what I need to do for now. 4x 2tb drives available. | 20:11 |
| cloudfly | vidd that's a good contractor =P | 20:11 |
| stevegjacobs | Just need to figure out how to do it. | 20:11 |
| vidd | cloudfly, i know union ppl =] | 20:11 |
| cloudfly | micric hrmm not sure how glance perms work | 20:11 |
| stevegjacobs | But I have another problem first - got to get the two nodes working nicely first | 20:11 |
| guaqua | stevegjacobs: i'd go with a raid setup, but dunno | 20:11 |
| guaqua | so no swift magic for failover :) | 20:12 |
| *** misheska has joined #openstack | 20:12 | |
| Micric | vidd, I have walked away before when they refused to cover any proof work | 20:12 |
| *** misheska has quit IRC | 20:12 | |
| Micric | They always ring back a week later and agree ;) | 20:13 |
| vidd | Micric, my problem is im currently laid off and this is my ticket back to full-time....and they know it | 20:13 |
| Micric | ouch, that is quite the catch 22 than | 20:14 |
| Micric | then* | 20:14 |
| cloudfly | do what you gotta do man | 20:14 |
| cloudfly | thems the breaks | 20:14 |
| vidd | they want to have this up so they can market it but they dont want to send me to texas for the training....they dont want to buy new machines with the specs i ask for .....next they will want me to part the Red Sea! | 20:16 |
| cloudfly | vidd heh, that's how this game is played. you are preaching to the choir. | 20:17 |
| *** nyeates has quit IRC | 20:19 | |
| vidd | zykes-, you here? | 20:20 |
| zykes- | vidd: busy atm | 20:23 |
| *** neunon has quit IRC | 20:23 | |
| vidd | no problem | 20:25 |
| *** JordanRinke has quit IRC | 20:25 | |
| *** huslage has joined #openstack | 20:27 | |
| *** huslage has quit IRC | 20:27 | |
| *** jedi4ever has quit IRC | 20:27 | |
| *** ejat has quit IRC | 20:32 | |
| Micric | And it all goes quiet again | 20:35 |
| *** rnorwood has quit IRC | 20:40 | |
| stevegjacobs | vidd - you sound like you have the same kinda company to work for as I do - | 20:42 |
| vidd | stevegjacobs, yes...i worked for a company.....currently furrlowed...laid off...what ever you want to call it | 20:42 |
| stevegjacobs | Sales promises something that can theoretically be done but that I've never done before, due at midday on Monday | 20:43 |
| stevegjacobs | And I'm up till 3 am tring to get it working :-) | 20:43 |
| cloudfly | =/ | 20:44 |
| cloudfly | i love when sales does that. | 20:44 |
| vidd | i have a data center managing partner claiming he can get "20K a month if we had this" but he cant seem to get it to work | 20:45 |
| cloudfly | personal opinion is that until quantum is out major deployments are a bit pre-mature | 20:46 |
| cloudfly | also i dunno how anyone could be trying to do a major deployment at this point | 20:46 |
| cloudfly | the codes only been out for around a year | 20:46 |
| vidd | so the other partners say "make this work" and "by the way, we are not going to buy any new equiptment until you prove it can work" | 20:46 |
| cloudfly | that's fine | 20:46 |
| cloudfly | but you need eval time | 20:47 |
| cloudfly | plus time to formalize processes | 20:47 |
| zykes- | quantum melange and hopefully a reamped dns :) | 20:47 |
| vidd | cloudfly, my plan is to take my 3 POS home computers in there tomorrow, wow them with dashboard, and get a half-way decent machine.... | 20:48 |
| cloudfly | sounds sketch | 20:48 |
| Micric | Hmmm, dashboard works well | 20:48 |
| vidd | set up a controller, and start converting the existing servers in the data center onto it.... | 20:49 |
| vidd | and take the freed up servers and make more compute nodes | 20:49 |
| cloudfly | you make it sound so simple. | 20:49 |
| cloudfly | =P | 20:49 |
| vidd | i hope to take 24 servers down to 6, and build a decent swift array | 20:50 |
| cloudfly | interesting | 20:50 |
| cloudfly | sounds almost like a proto design for a drop in esxi replacement | 20:51 |
| vidd | cloudfly, the hard part is going to be cloning the existing servers into cloud images | 20:51 |
| cloudfly | time intensive at least | 20:51 |
| cloudfly | the other hard parts will be network layout | 20:51 |
| guaqua | cloudfly: are you referring to nova and the computing parts when you say it isn't mature enough? to the dashboard part? is there something missing from swift? | 20:51 |
| cloudfly | likely there will be some issues there | 20:51 |
| *** Razique has quit IRC | 20:52 | |
| cloudfly | guaqua quantum is my biggest issue | 20:52 |
| cloudfly | until quantum drops into prod use | 20:52 |
| cloudfly | the network layout is going to paradigm shift | 20:52 |
| cloudfly | it's hard to do a major deployment with that on the horizon | 20:52 |
| guaqua | you don't need openstack for virtualization | 20:52 |
| cloudfly | certainly. | 20:53 |
| guaqua | if you want to make an existing infrastructure virtual, it's easier not to try to go to cloud at the same time | 20:53 |
| vidd | guaqua i know i dont need openstack for virtualization | 20:53 |
| cloudfly | guaqua i'll agree there to a degree | 20:53 |
| cloudfly | i do think that openstack vendors need to capitalize on the sudden increase in vsphere pricing | 20:54 |
| cloudfly | but the fundamental difference in cloud scalability and disaster response is difficult for vmware users to adapt to | 20:54 |
| vidd | but at the same time, since im going to be building a market model, i might just as well have the market model acually DO something besides sit on a rack and look pretty =] | 20:54 |
| cloudfly | brb | 20:55 |
| cloudfly | about to update the at home test bed to 11.10 | 20:55 |
| guaqua | openstack and cloud is nice for marketing, but at the moment you need either very tech savvy and enthusiastic clients or just plain not-very-business-savvy clients | 20:55 |
| stevegjacobs | I'm going to try to rescue that system now - | 20:56 |
| stevegjacobs | for a tech savvy enthusiastic client - but we need to get it to do something... | 20:57 |
| guaqua | :) | 20:57 |
| guaqua | that's always needed | 20:57 |
| vidd | ok...im getting ready to format a machine dediacated for swift..... | 20:58 |
| vidd | i have a 320gb hard drive and a 640Gb hard drive in it.... | 20:58 |
| vidd | should i split the 640 in half and work with 3 320 partitions? | 20:59 |
| vidd | and would y'all recommend ext partition or lvm's? | 20:59 |
| notmyname | vidd: (can't stay in channel long, about to walk out the door) I'd recommend that you use whole drives for a swift cluster | 21:00 |
| guaqua | always use lvm :) | 21:00 |
| notmyname | the point of swift is to distribute your data across many possible hardware failure scenarios | 21:01 |
| vidd | both drives on one machine...can i do a "virtual" cluster like that? | 21:01 |
| notmyname | you can use ext, but we've extensively tested xfs and found it to be better for a general use case that other file systems | 21:02 |
| cloudfly | man | 21:02 |
| zykes- | vidd: CORRECT | 21:02 |
| zykes- | pardon the caps | 21:02 |
| notmyname | vidd: you can set it up all in one VM with loopback devices (see the SAIO instructions), but for prod usage, you'll need at least 4 or 5 storage nodes | 21:02 |
| notmyname | sorry to say stuff and leave, but I gotta run | 21:03 |
| *** Micric has quit IRC | 21:03 | |
| vidd | yeah...this is a demo =] | 21:03 |
| * notmyname out | 21:03 | |
| vidd | l8r | 21:03 |
| cloudfly | i always forget is it ahci for booting off usb or ide.. | 21:03 |
| *** rnorwood has joined #openstack | 21:07 | |
| guaqua | are there any user stories for swift? i haven't seen any disaster recovery stories, for example | 21:09 |
| cloudfly | hrmm we're planning a swift roll out shortly | 21:10 |
| cloudfly | we've been using nova forever | 21:10 |
| cloudfly | but swift not so much | 21:10 |
| guaqua | what kind of setup do you have? how many storage nodes, proxies? | 21:10 |
| *** carlp has quit IRC | 21:10 | |
| cloudfly | no clue yet | 21:10 |
| guaqua | okay | 21:10 |
| cloudfly | i am waiting on allocation requirements | 21:10 |
| guaqua | i'm looking into using swift for a public project | 21:11 |
| cloudfly | cool. | 21:11 |
| guaqua | i mean, there's lots of "you should use min 5 nodes on min 3 zones" | 21:11 |
| guaqua | but not that much "this is how a node failure is detected" and "this is how you extend" | 21:12 |
| cloudfly | well my understanding is that every chunk of data is stored 3 times | 21:12 |
| cloudfly | and the swift background processes compare all three | 21:12 |
| cloudfly | if one is off... the other 2 win and it gets written over | 21:12 |
| cloudfly | or rewritten elsewhere | 21:12 |
| cloudfly | something to that effect | 21:13 |
| cloudfly | and it seems to store the 3 to different hosts | 21:13 |
| cloudfly | i don't believe it's geographically aware at this point though | 21:13 |
| cloudfly | which is a pity | 21:13 |
| guaqua | according to documentation, it is. if you set up zones geographically, that is | 21:13 |
| cloudfly | yeah i suppose that's one way to do it | 21:14 |
| cloudfly | not that everyone does that | 21:14 |
| guaqua | theoretically it seems okay, but i'd really like to know how to operate it since that needs to be done anyway | 21:14 |
| cloudfly | yeah best approach there is setup an eval kit and go to town on ot | 21:14 |
| cloudfly | err t | 21:14 |
| cloudfly | ack it | 21:14 |
| cloudfly | and then build some procedural docs | 21:15 |
| guaqua | yeah. i've just done a few days of testing until now | 21:15 |
| cloudfly | i don't really have the gear in my apt to test swift. | 21:15 |
| cloudfly | nova sure | 21:15 |
| cloudfly | but swift requires multiple machines to really test it | 21:16 |
| cloudfly | =/ | 21:16 |
| guaqua | yep. i don't have a test lab either. and using ec2 for that is quite tedious | 21:16 |
| guaqua | (i.e. requires better tools...) | 21:16 |
| cloudfly | so supposedly there is a new shell script for installing nova | 21:19 |
| cloudfly | that works fairly well | 21:19 |
| vidd | hrm....this SAIO says it wants to be done in a VM...... | 21:19 |
| cloudfly | i think i might give that a run on the box i am rebuilding atm | 21:19 |
| vidd | i wonder if i can skip the "make a vm" part and just put it on real hardware | 21:19 |
| vidd | my test machine does not have VM ability | 21:20 |
| cloudfly | i assume you mean hvm | 21:22 |
| cloudfly | i mean you can emulate. | 21:23 |
| cloudfly | anyone can | 21:23 |
| vidd | right | 21:23 |
| vidd | well....im not even too sure about that with THIS equiptment =\ | 21:23 |
| vidd | someone needs to explain to my bosses that "commodity hardware" != "legacy hardware" ! | 21:24 |
| cloudfly | i wish there was a service i could subscribe to that would bring me food when i am working | 21:24 |
| Kiall | I'm sure there is ;) | 21:25 |
| vidd | cloudfly, where do you ive? | 21:25 |
| cloudfly | mountain view | 21:25 |
| vidd | *live | 21:25 |
| vidd | what country | 21:25 |
| cloudfly | usa | 21:25 |
| Kiall | CA, US ;) | 21:25 |
| vidd | have you never heard of "Domino"? | 21:25 |
| cloudfly | yes... the only place that delivers in mountain view | 21:25 |
| guaqua | vidd: google started with commodity, legacy hardware. anything they could use | 21:26 |
| cloudfly | but man... that is so not healthy | 21:26 |
| cloudfly | or very good | 21:26 |
| cloudfly | guaqua things were different back then | 21:26 |
| vidd | cloudfly, then get a veggie with br uck ly | 21:26 |
| cloudfly | yeah i am still pretty sure that qualifies as unhealthy. I'm no doctor, but i kinda feel like a slice from dominos puts one dangerously close to sodium poisoning. | 21:27 |
| vidd | i have a tomato allergy and domino hasn't killed me yet | 21:28 |
| rmk | How does one represent "any port" via the dashboard? | 21:29 |
| rmk | When configuring security groups, of course. | 21:29 |
| stevegjacobs | I just took a break and finished off a nice roast chicken dinner | 21:29 |
| cloudfly | vidd that's terrifying | 21:29 |
| *** code_franco has joined #openstack | 21:30 | |
| Kiall | is it just me, or is https://review.openstack.org/ really unstable? | 21:30 |
| vidd | start = 0 ; end = [whatever the last available port is] | 21:30 |
| cloudfly | seems to be running okay for me kiall | 21:30 |
| cloudfly | but maybe it's spurious | 21:30 |
| vidd | cloudfly, i LOVE tomatoes...they just don't love me back | 21:31 |
| Kiall | cloudfly: yea it is now, but it seems to go down for like 30-60 seconds at a time every so often | 21:31 |
| vidd | an eppy a day keeps the corenor away =] | 21:31 |
| cloudfly | vidd one would think that a tomato allergy would preclude one from eating pizza. | 21:31 |
| cloudfly | Kiall not sure how it works maybe it's doing a cron update or something | 21:31 |
| vidd | cloudfly, one would think =] | 21:32 |
| Kiall | I doubt it TBH! Google's instance for android never had issues ;) | 21:32 |
| vidd | actually...my allergy is not "thoat swells shut" kind....its the "breaks out in hives" kind | 21:33 |
| cloudfly | okay | 21:33 |
| cloudfly | so what's the current preferred method for deploying diablo on a stand alone 11.10 system? | 21:34 |
| Kiall | Using my PPA ;) | 21:34 |
| stevegjacobs | still can't figure out why associating a public ip address is no longer working on my system | 21:34 |
| cloudfly | Kiall i don't doubt that =P | 21:34 |
| cloudfly | any docs on that? | 21:34 |
| vidd | i have to keep my tomato consumption way down (no more than 3 slices every 5 days) or i break out something feirce | 21:34 |
| Kiall | The ubuntu packages are for keystone+dash are just broken | 21:34 |
| cloudfly | i've heard | 21:34 |
| Kiall | s/are// | 21:34 |
| Kiall | https://launchpad.net/~managedit/+archive/openstack | 21:35 |
| Kiall | Thats the stable/diablo branches of everything... | 21:35 |
| cloudfly | how often do you update your ppa? | 21:35 |
| Kiall | "12 minutes ago" ;) | 21:35 |
| cloudfly | hah | 21:35 |
| cloudfly | that's what i like to hear =P | 21:35 |
| vidd | when i first learned about it, i had to take hydrocortizone pills for a month to clear it up | 21:35 |
| Kiall | I keep it updated maybe once a week with anything thats changed in the stable/diablo branches | 21:36 |
| cloudfly | vidd i think you eat too much pizza... and I say that as a loyal son of brooklyn | 21:36 |
| stevegjacobs | Kiall would installing this over the top of my broken ubuntu packages stand a chance of fixing it? | 21:36 |
| Kiall | if it was only the packages, you installed, most likely. | 21:36 |
| Kiall | if it was only the packages you installed, most likely.* | 21:37 |
| *** code_franco has quit IRC | 21:37 | |
| Kiall | if you ran devstack or any "python setup.py" etc, then your /usr/local is filled with stuff that will override the packages | 21:37 |
| Kiall | any packages* | 21:37 |
| cloudfly | might go with dev stack tbh | 21:38 |
| cloudfly | kinda want to try some stuff | 21:38 |
| Kiall | Yea, if you want to dev .. use dev stack .. | 21:38 |
| Kiall | if you want a production install.. then dev stack, as it says in the name, is not suitable | 21:39 |
| cloudfly | heh | 21:39 |
| cloudfly | absolutely | 21:39 |
| stevegjacobs | I just want something that will work and run a few instances and do some real work | 21:39 |
| Kiall | stevegjacobs: well .. my PPA packages work .. installed, edit the *.conf files to suit .. done. | 21:40 |
| Kiall | and devstack the same... | 21:40 |
| Kiall | but devstack is harder to maintain... | 21:40 |
| vidd | devstack was very useful in assisting me with setting up dashboard in apache | 21:41 |
| Kiall | yea - I've picked apart the script more than a few times to understand everything | 21:42 |
| stevegjacobs | I set up everything so far from ubuntu packages - I just looked in /usr/local/bin - nothing there - | 21:42 |
| cloudfly | is wiki'd approach still preferred method for dev stack deployment? | 21:42 |
| stevegjacobs | so is that a good indication that I can use Kiall's ppa? | 21:42 |
| Kiall | stevegjacobs: then, no harm in trying the PPA packages, they can easily be removed as there packages if you want devstack in the end (which can't be said for devstack) | 21:43 |
| Kiall | stevegjacobs: if you dpkg -P all the packages then rm -rf /var/lib/(nova|glance|keystone) .. rm -rf /etc/(nova|glance|keystone) to be sure they are gone... | 21:44 |
| vidd | and a quick note to anyone who cares...to get any site up without having to put anything in the /var/www/ directory, just start the site-enabled filename with 00*-dir-name | 21:44 |
| cloudfly | stevegjacobs Kiall's packages are probably preferable | 21:44 |
| vidd | i did not overwrite the 000-default file, a created a 001-dash file | 21:45 |
| Kiall | vidd: yea, i just lob a dashboard.conf into /etc/apache2/conf.d/ with my packages | 21:48 |
| *** Ruetobas has quit IRC | 21:50 | |
| vidd | Kiall, i suppose if you want to go the phpmyadmin way....=] | 21:51 |
| Kiall | vidd: yea.. I figure its fine.. your other web apps should be in instances ;) | 21:52 |
| cloudfly | nova.sh is fun | 21:52 |
| *** rnorwood has quit IRC | 21:54 | |
| vidd | Kiall, i was about to format my fully functional box so i can test having keystone, glance, swift, mysql, and dashboard running on seperate machines.... | 21:55 |
| vidd | but then i was like....wait...just change the nova.conf file! | 21:55 |
| *** TREllis has quit IRC | 21:56 | |
| vidd | This way, if it dont work as expected i can just change it back instead of going through the whole process again =] | 21:57 |
| *** Ruetobas has joined #openstack | 21:57 | |
| Kiall | vidd: huh? | 21:58 |
| vidd | i have 3 machines to work with....2 cant do hardware virtualization and have less than a gig of ram each | 21:59 |
| *** MarkAtwood has quit IRC | 21:59 | |
| Kiall | then use LXC on them rather than KVM ;) | 21:59 |
| *** MarkAtwood has joined #openstack | 22:00 | |
| Kiall | (or .. get suitable hardware!) | 22:00 |
| vidd | so i am setting the one with alot of hards drive space up as a faux swift array, and the relic to run the services (keystone, mysql,etc) | 22:00 |
| *** miclorb_ has joined #openstack | 22:01 | |
| vidd | Kiall, i tried that b4 and the instances would not load at all....even with devstack | 22:04 |
| Kiall | weird, I've never tried it.. maybe it doesnt work ;) | 22:04 |
| zykes- | vidd: what problems areyou having now ? | 22:04 |
| stevegjacobs | Does it work to have some nodes with kvm and some with lxc? | 22:04 |
| cloudfly | you know... peoples houses are always so clean in the movies | 22:04 |
| zykes- | Kiall: go experience with kvm / openvz ? | 22:04 |
| zykes- | stevegjacobs: why lxc ? | 22:04 |
| vidd | zykes-, just the dashboard not able to see the users or the tenants | 22:05 |
| zykes- | ah o | 22:05 |
| Kiall | zykes-: nope, only toyed with LXC a few months back... | 22:05 |
| stevegjacobs | I have some hardware that doesn't support kvm that I would like to reuse | 22:05 |
| *** sebastianstadil has joined #openstack | 22:07 | |
| *** vidd-away has joined #openstack | 22:08 | |
| *** vidd-away has joined #openstack | 22:08 | |
| stevegjacobs | Kiall - just tried your packages - all good till then end - error on installing openstack-dashboard http://paste.openstack.org/show/AcIp2RhZMI47OKDfzscD/ | 22:08 |
| *** vidd has quit IRC | 22:08 | |
| zykes- | stevegjacobs: you can also use xen with older hardware and do pvirt | 22:08 |
| *** vidd-away is now known as vidd | 22:08 | |
| vidd | zykes-, i got disconnected...did you get my issue discription? | 22:09 |
| *** david__c has quit IRC | 22:13 | |
| *** localhost has quit IRC | 22:13 | |
| *** localhost has joined #openstack | 22:15 | |
| *** TREllis has joined #openstack | 22:17 | |
| *** TREllis has quit IRC | 22:18 | |
| stevegjacobs | Kiall - not sure I understand what I need to do next after installing from your packages - | 22:18 |
| Kiall | stevegjacobs: : sorry - I AFK | 22:18 |
| vidd | stevegjacobs, you configure the apps | 22:19 |
| Kiall | stevegjacobs: apt-get -f install .. it looks like i need to put a harder dependency on the python package | 22:19 |
| *** TREllis has joined #openstack | 22:20 | |
| Kiall | stevegjacobs: after installing the packages, start by setting up keystone.. | 22:20 |
| Kiall | normally that means changing the /etc/keystone/keystone.conf to point at a MySQL database | 22:21 |
| Kiall | and loading some sample data | 22:21 |
| Kiall | Some sample keystone data: https://github.com/cloudbuilders/devstack/blob/master/files/keystone_data.sh | 22:21 |
| Kiall | replacing all the %BLA% variables in that file | 22:22 |
| Kiall | other than %tenant_id% that is.. | 22:22 |
| vidd | yeah that tenant ID is important =] | 22:22 |
| Kiall | then .. once keystone is setup (ie you've ran those commands without an error)... move onto glance | 22:23 |
| Kiall | again, edit /etc/glance/*.conf to use mysql, set it the service token you setup in keystone.. then onto nova... | 22:24 |
| Kiall | again .. mysql + set service token + tell it to use ketstone in nova.conf... | 22:25 |
| Kiall | I actually have most of this scripted .. i might just clear out the private parts and put them somewhere ;) | 22:26 |
| cloudfly | man | 22:26 |
| cloudfly | dev stack has come a long ways from novascript. | 22:26 |
| vidd | cloudfly, its come a long way in the last 2 weeks | 22:27 |
| vidd | its got swift in it now...and quantum actually works | 22:27 |
| cloudfly | stellar | 22:27 |
| cloudfly | need to cut teeth on quantum | 22:27 |
| vidd | well...i hear that it works =] | 22:27 |
| cloudfly | heh | 22:27 |
| cloudfly | for lesser values of "works" | 22:28 |
| cloudfly | i am sure | 22:28 |
| Kiall | lol | 22:28 |
| vidd | 2 weeks ago, there was just a psydo-quantum in there...now it acually downloads the package and everything! | 22:29 |
| cloudfly | one thing that's been true of this project since forever... it moves damned fast =P | 22:29 |
| zykes- | vidd: what ? | 22:30 |
| vidd | yeah...tell me about it! the keystone database changes at least twice since i started working on this early october | 22:30 |
| zykes- | Kiall: i was wondering on howto do with a DNSaaS service | 22:30 |
| Kiall | AFAIK, DNSaaS is something rackspace is still only considering open sourcing... | 22:31 |
| cloudfly | seems kind of superfluous | 22:31 |
| vidd | zykes-, i was saying...my dashboard wont show the user or the tenant pages | 22:32 |
| zykes- | Kiall: Kiall i was thinking of starting on one if not... | 22:32 |
| zykes- | vidd: k | 22:32 |
| zykes- | tried running the dev server it comes with vidd and see if it gives an err ? | 22:32 |
| zykes- | or enabling logging on it | 22:32 |
| vidd | zykes-, i have the devs enabled....dash wont load outside venv without the eggs to back it up =\ | 22:33 |
| zykes- | :| | 22:34 |
| *** MarkAtwood has quit IRC | 22:34 | |
| vidd | its something in the python-keystoneclient | 22:34 |
| vidd | i think the databse and the client are not lining up | 22:35 |
| zykes- | does it have the python-keystone client installed? | 22:35 |
| vidd | yes | 22:35 |
| stevegjacobs | Kiall I've set up sql-connection to mysql - but there is a setting default_store = sqlite - does that need to be changed? | 22:35 |
| zykes- | funkhy | 22:35 |
| zykes- | stevegjacobs: no need to change that | 22:35 |
| zykes- | you just need to change the connection uri | 22:36 |
| Kiall | stevegjacobs: 2 seconds .. about to push up the scripts I used to set everything up.. | 22:38 |
| Kiall | https://github.com/managedit/openstack-setup | 22:40 |
| *** arBmind has joined #openstack | 22:40 | |
| Kiall | Its probably not 100% complete, since I was making it as I went and missed a few steps.. | 22:41 |
| Kiall | but usually just stuff like restarting services before running the *-test.sh files | 22:41 |
| zykes- | Kiall: you think it's a vad idea with the dnsaas ? | 22:41 |
| Kiall | no, DNS is a logical addition to OS... | 22:41 |
| zykes- | i mean rs hasn't come out with anything yet so :p | 22:42 |
| Kiall | vidd: maybe those scripts will help you too ;) | 22:42 |
| Kiall | zykes-: they probably will at some point :) | 22:42 |
| vidd | Kiall, i have the template that works with my stored version of keystone.... | 22:42 |
| Kiall | they said they would, and frankly, their kept their word so far :) | 22:43 |
| zykes- | Kiall: heh, i was wanting to start on something but not sure if i will then :p | 22:43 |
| Kiall | they've kept their word so far* | 22:43 |
| zykes- | Kiall: i had an idea on dnsaas using powerdns as the dns server | 22:43 |
| zykes- | but ok | 22:43 |
| zykes- | ;p | 22:43 |
| vidd | zykes-, ya never know....your stuff might be better than rackspace's =] | 22:44 |
| Kiall | powerdns = win when it comes to easy of management ;) | 22:44 |
| vidd | what is powerDNS writen on? | 22:44 |
| *** mgw has joined #openstack | 22:45 | |
| *** mgw has left #openstack | 22:45 | |
| Kiall | no idea TBH | 22:45 |
| zykes- | lemme see vidd | 22:46 |
| Kiall | probably C or C++ | 22:46 |
| *** jj0hns0n has joined #openstack | 22:47 | |
| vidd | its GPL'd....rackspace wont like that =] | 22:47 |
| zykes- | what's wrong with the GPL ? ;p | 22:47 |
| vidd | zykes-, dunno....rackspace seems to like apache better =\ | 22:48 |
| zykes- | Kiall: i was wondering on howto architecture itr | 22:49 |
| cloudfly | hrmm | 22:49 |
| cloudfly | where does dev stack put the api keys by default? | 22:50 |
| vidd | ahhhh there are already fully functional .debs AND RPM's another negative it would appear =] | 22:50 |
| vidd | cloudfly, in the stackrc file i believe | 22:50 |
| zykes- | vidd: what's bad with that ? | 22:51 |
| vidd | i didnt say it was BAD....it just seams these are things openstack dont like =] | 22:51 |
| zykes- | eh, what's bad with that? they have pkgs in distros.. | 22:52 |
| vidd | and some of them actually work | 22:52 |
| zykes- | uhm, all pdns packages work in distros ? | 22:52 |
| zykes- | at least el* and ubuntu stuff | 22:52 |
| vidd | zykes-, im refering to OS packages | 22:53 |
| zykes- | ah | 22:53 |
| zykes- | wonder who i could contact at RS regarding dns ;p | 22:54 |
| cloudfly | hrmm | 22:54 |
| arBmind | hi, does it make sense to use openstack for a single server to manage virtual machines? | 22:54 |
| arBmind | documentation states, that single server setup is only for testing | 22:55 |
| vidd | arBmind, do you intend to expand to multiple machines? is this a proof of concept? | 22:55 |
| vidd | arBmind, single server is INTEDED for testing.... | 22:56 |
| arBmind | we want to use this for our small site, and development server - project management and stuff | 22:56 |
| arBmind | currently we run proxmox to control virtual machines | 22:56 |
| vidd | but, like in my case, you can start with a single all-in-one and then expand to multiple machines as your needs increase | 22:57 |
| arBmind | but this seems to lack the network management | 22:57 |
| Kiall | stevegjacobs: did you get the link to those scripts? | 22:57 |
| *** rnorwood has joined #openstack | 22:57 | |
| arBmind | and now i am looking for a replacement | 22:57 |
| vidd | arBmind, check out the devstack script | 22:58 |
| vidd | im sure that your networking needs [and then some] will be met | 22:58 |
| zykes- | Kiall: what you thin k? | 22:58 |
| Kiall | about? sorry, was AFK ;) | 22:59 |
| vidd | run it in a VM ... for testing purposes..... | 22:59 |
| arBmind | vidd, I tried that on my home linux server, but always ran into issues | 22:59 |
| vidd | arBmind, what type of issues? | 22:59 |
| zykes- | Kiall: dnsaas, writing one | 22:59 |
| vidd | what kind of equiptment is your home server? | 22:59 |
| zykes- | since rs is most likely doing it | 22:59 |
| arBmind | vidd, first it complained about a directory it was not allowed to create... | 22:59 |
| Kiall | zykes-: oh .. i'd ask on the OS mailing list to see what the RS people are doing.. | 23:00 |
| vidd | does your hove server have kvm ability...and is it enabled in bios | 23:00 |
| vidd | arBmind, i recommend you "sudo su" then run the script | 23:00 |
| arBmind | vidd, it has and it's enabled (intel i5 processor) | 23:00 |
| arBmind | vidd, that's what i did | 23:01 |
| arBmind | vidd, I used ubuntu mini 11.10 image... | 23:01 |
| vidd | how long ago did you grab the script? | 23:01 |
| arBmind | today, 12 hours ago or so | 23:01 |
| vidd | and you have a proper install of ubuntu | 23:02 |
| arBmind | i don't know how I can install ubuntu wrong | 23:03 |
| arBmind | i reserved some lvm space on the hdd for images, but the script never went through | 23:03 |
| *** jj0hns0n has quit IRC | 23:03 | |
| vidd | there are ppl out there that can burn water ! | 23:03 |
| vidd | does your server have internet access? | 23:04 |
| arBmind | server can access the internet though my home router | 23:04 |
| arBmind | the script installed all kinds of packages | 23:05 |
| vidd | what does the message say? | 23:05 |
| arBmind | I have not written it down, that's why I asked the more general setup question first :) | 23:06 |
| vidd | so i take it then you are not near that machine =] | 23:06 |
| arBmind | I type to the machine right now... but with a different hard disk... | 23:07 |
| vidd | ive gotten the devstack script to install on a pIII with 192 mb ram 80 gb hard drive so im sure your surever can handle it =] | 23:08 |
| vidd | so you have a dual-boot? | 23:09 |
| arBmind | My issue so far is that all the documentation seems to be a bit outdated | 23:09 |
| arBmind | no, i switched the hdd with cables ;) | 23:09 |
| vidd | this project does move fast....before the ink dries, there are 15 updates =] | 23:09 |
| *** PiotrSikora has quit IRC | 23:10 | |
| arBmind | that's what keeps me looking around :) | 23:10 |
| *** PiotrSikora has joined #openstack | 23:10 | |
| arBmind | really cool project, but hard to get into | 23:10 |
| cloudfly | depends on what you are doing i think | 23:11 |
| Kiall | vidd: which is why you should use the stable repos ;) | 23:11 |
| vidd | well...to see where the issue is, we need to see the issue =] | 23:11 |
| vidd | Kiall, im refering to the devstack script | 23:12 |
| arBmind | ok, I will give it a fresh run tomorrow, and report it here | 23:12 |
| vidd | arBmind, what is this machine running now? | 23:12 |
| arBmind | its running ubuntu server with desktop, used as a storage and media server | 23:13 |
| arBmind | or the production server? | 23:13 |
| *** MarkAtwood has joined #openstack | 23:14 | |
| vidd | you dont want to put the devstack script on a production machine | 23:14 |
| cloudfly | well i mean unless you do. in which case i'd love to go to your parties. | 23:14 |
| vidd | can you spin up a VM on your production machine and run the script in that? | 23:15 |
| vidd | cloudfly, i dont like parties that include driving cars off cliffs as the prelude to the main attraction =] | 23:15 |
| cloudfly | I do. | 23:16 |
| arBmind | it theoretically possible, but proxmox requires to install the virtual machine from a cd image, that would take a while | 23:16 |
| vidd | ouch...yeah...this is your replacement source! | 23:17 |
| arBmind | yes | 23:17 |
| vidd | new fully functioal VM's in 15 seconds or less | 23:17 |
| vidd | just take a saved image and tell it to run | 23:18 |
| *** lorin1 has joined #openstack | 23:18 | |
| cloudfly | the mcdonalds of virtual machines | 23:18 |
| arBmind | that's what i am looking for :) | 23:18 |
| *** lorin1 has quit IRC | 23:18 | |
| cloudfly | that'd make a hell of an app store | 23:18 |
| *** lorin1 has joined #openstack | 23:19 | |
| vidd | and if you need to you can always super-size your order and spin up images 10 at a time =] | 23:19 |
| cloudfly | i'll take a number 3 with postgres in place of the mysql and an extra side of logging. | 23:19 |
| arBmind | but i fear that openstack is only a starting point | 23:21 |
| arBmind | we are already investigating chef | 23:21 |
| arBmind | but are still unsure... | 23:21 |
| vidd | make mine a -t m1.large with a side order of 25 GB storage and hold the logging ...to go | 23:21 |
| cloudfly | chef for what? | 23:21 |
| cloudfly | well i mean nova does the iaas. | 23:21 |
| cloudfly | the paas is a whole nother deal | 23:21 |
| arBmind | storing all the configuration for each vm | 23:21 |
| vidd | i never understood that....isnt chef what you use to BUILD os deployments? | 23:21 |
| *** miclorb_ has quit IRC | 23:22 | |
| cloudfly | it's configuration management | 23:22 |
| cloudfly | os deployments are more like kickstart | 23:22 |
| cloudfly | or deb bootstrap | 23:22 |
| *** shentonfreude has joined #openstack | 23:22 | |
| cloudfly | chef does stuff like config files and package sets | 23:22 |
| vidd | we have a plug-in for that =] | 23:22 |
| Kiall | cloudfly: yea, like you need for installing OS :) | 23:22 |
| cloudfly | not at all | 23:22 |
| zykes- | vidd: for what ;p | 23:23 |
| cloudfly | you only need that AFTER you installed the os | 23:23 |
| cloudfly | immediatelty after sure | 23:23 |
| vidd | zykes-, chef with OS | 23:23 |
| cloudfly | but after for certain | 23:23 |
| Kiall | sure, you bootstrap the HW with kickstart, and chef to manage the config.. | 23:23 |
| cloudfly | yep | 23:23 |
| Kiall | since kickstart is a once off thing, and confog changes over time (obliviously!) | 23:23 |
| cloudfly | what's interesting is that openstack has no agent for vms | 23:23 |
| zykes- | vidd: ok ;p | 23:23 |
| cloudfly | as in an internal configuration management | 23:24 |
| cloudfly | quantum i think will help here | 23:24 |
| zykes- | cloudfly: what you mean ? | 23:24 |
| zykes- | quantum is not for configuring vm's cloudfly | 23:24 |
| cloudfly | spinning up an operations openstack deployment and then joining instances from that openstack unit to become the chef servers for user clouds | 23:24 |
| cloudfly | the idea is to cut the operations control off from the user accessed api serversd | 23:25 |
| vidd | quantum is for getting the network to work =] | 23:25 |
| cloudfly | yes | 23:25 |
| cloudfly | and it helps with overlaying one openstack clusters vms into anothers. | 23:25 |
| cloudfly | which i think is pretty essential | 23:25 |
| zykes- | cloudfly: say what ? | 23:25 |
| cloudfly | i'll need some specifics | 23:26 |
| cloudfly | direct my response | 23:26 |
| zykes- | cloudfly: quantum is for controlling your os network components | 23:26 |
| cloudfly | yes | 23:26 |
| cloudfly | and some of the folks building plugins for quantum | 23:26 |
| zykes- | nothing to do with config management | 23:26 |
| cloudfly | zykes... depends on your design | 23:27 |
| zykes- | in the terms of a "cms" | 23:27 |
| arBmind | * installing virtual machine for another run | 23:27 |
| cloudfly | grrr... | 23:27 |
| cloudfly | stack.sh hung on mysql restart | 23:27 |
| zykes- | cloudfly: no not at all, it's a networking service, as in to control where ports are located etc | 23:28 |
| vidd | i wonder if quantum will fix my "i only have one nic" issue =\ | 23:29 |
| cloudfly | zykes and what i was pointing out | 23:29 |
| cloudfly | is that it can be used to allow for federation between openstack clusters. | 23:29 |
| zykes- | vidd: only have one nic issue ? | 23:29 |
| vidd | yeah....sometimes my system puts the floating ip as the eth0 ip | 23:30 |
| vidd | so when i try to ssh into the vm, im actually getting into my server | 23:31 |
| vidd | im wondering if hard-coding the br100 in the /etc/networking/interfaces will help | 23:32 |
| *** clopez has joined #openstack | 23:33 | |
| *** lorin1 has quit IRC | 23:33 | |
| vidd | be back in a bit | 23:34 |
| *** vidd is now known as vidd-away | 23:35 | |
| arBmind | is there a way to forward ports from the host machine to the virtual machines? | 23:35 |
| arBmind | a setup without public ips for the vms? | 23:35 |
| *** aaron has joined #openstack | 23:36 | |
| cloudfly | you could probably do it with iptables | 23:36 |
| arBmind | cloudfly, but that would defeat the advantage of openstack security groups, would it not? | 23:37 |
| cloudfly | yep | 23:38 |
| arBmind | but i fear that would not count as an argument to request more ip addresses for a root server here :( | 23:39 |
| *** lionel has quit IRC | 23:39 | |
| *** aaron has quit IRC | 23:39 | |
| *** lionel has joined #openstack | 23:39 | |
| stevegjacobs | I am still here - just finished keystone setup - keystone-test.sh success :-) | 23:40 |
| *** nerens has quit IRC | 23:46 | |
| Kiall | stevegjacobs: if you find anything missing from those scripts, let me know.. | 23:47 |
| *** qazwsx has quit IRC | 23:48 | |
| arBmind | Kiall, what's the purpose of your scripts? | 23:48 |
| Kiall | a simpler version of devstack that uses ubuntu packages, rather than git clones... | 23:49 |
| Kiall | a much much simplier version of devstack ;) | 23:49 |
| arBmind | that would help a lot... | 23:50 |
| Kiall | ;) | 23:50 |
| arBmind | is the dashboard finished already? | 23:51 |
| Kiall | well, it works... but its not an official OS project until essex is released | 23:53 |
| cloudfly | this is horizon? | 23:53 |
| cloudfly | or some other thing? | 23:53 |
| Kiall | yea - horizon | 23:54 |
| stevegjacobs | Kiall, so far so good! This is all installing on the controller node. What about compute node(s)? | 23:54 |
| *** nerens has joined #openstack | 23:54 | |
| *** nerens has quit IRC | 23:54 | |
| Kiall | You probably want to edit nova.sh then :) | 23:55 |
| Kiall | Those scripts setup a all in one OS install, which, is probably worth doing. to start with. | 23:57 |
| Kiall | then once its up and running, install a second, remove the first with dpkg -S etc | 23:57 |
| *** jj0hns0n has joined #openstack | 23:58 | |
| cloudfly | it would be nice is devstack grabbed a few images by default. | 23:58 |
| *** bryguy has quit IRC | 23:58 | |
| arBmind | vidd, i ran the stack.sh script successfully on my virtual machine here, first time - that gives me hope | 23:58 |
| *** bryguy has joined #openstack | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!