Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel, 2016-05-20

| Channels | #fuel index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:10 ericlopez joined #fuel
00:22 eric_lopez joined #fuel
01:26 fatdragon joined #fuel
01:34 eric_lopez joined #fuel
01:48 ilbot3 joined #fuel
01:48 Topic for #fuel is now Fuel 8.0 (Liberty) https://www.fuel-infra.org/ | Paste here http://paste.openstack.org/ | IRC logs http://irclog.perlgeek.de/fuel/
01:52 gardlt joined #fuel
01:56 fatdragon joined #fuel
01:57 gardlt joined #fuel
02:03 gomarivera joined #fuel
02:11 eric_lopez joined #fuel
02:23 eric_lopez joined #fuel
02:25 raunak joined #fuel
02:41 eric_lopez joined #fuel
02:59 eric_lopez joined #fuel
03:24 Samos123 what's the recommended way to enable CentOS repo in Fuel? I want to install libvirt and KVM on the fuel node itself
03:34 gardlt joined #fuel
03:47 gardlt_ joined #fuel
03:52 ericlopez joined #fuel
04:05 eric_lopez joined #fuel
04:29 gomarivera joined #fuel
04:38 javeriak joined #fuel
04:58 javeriak joined #fuel
05:13 javeriak_ joined #fuel
05:14 eric_lopez joined #fuel
05:34 ericlopez joined #fuel
05:53 ericlopez joined #fuel
06:34 Ir0n joined #fuel
06:38 eric_lopez joined #fuel
06:45 DavidRama joined #fuel
06:53 AlexPost joined #fuel
06:53 fzhadaev Samos123: You can just add CentOS-Base repo by common ways. (e.g. adding .repo file to /etc/yum.repos.d )
06:53 eric_lopez joined #fuel
06:59 gardlt_ joined #fuel
07:03 parlos joined #fuel
07:16 fatdragon joined #fuel
07:25 akaszuba joined #fuel
07:33 Samos123 ok thanks for confirming. Will try
07:43 slefeuvre joined #fuel
07:43 slefeuvre Hi all,
07:46 slefeuvre I'm getting a problem with Fuel 7.0 regarding deployement, When controller 2 is deploying, I get an error when is trying to send "openstack endpoint create". In facts, all openstack commands are not reachable from controller 2 when all vip are started in controller 1.
07:46 slefeuvre It seems to be a keystone "v2" problem
07:46 slefeuvre someone already have this kind of problem ?
07:53 Samos123 slefeuvre: it sounds like a network problem between controller 2 and controller 1, haven't had issues with keystonve v2
07:56 javeriak joined #fuel
08:02 pbelamge joined #fuel
08:22 slefeuvre @Samos123: I've found this bug before https://bugs.launchpad.net/fuel/+bug/1479382
08:23 slefeuvre @Samos123: There is no network problem betweens controller, I can ping each other over all networks
08:35 yassine__ joined #fuel
08:50 noshankus joined #fuel
08:55 slefeuvre @Samos123: I can pass through the problem by migrating the crm "vip__public" resource on the 2nd controller during deployment
08:55 LGee Most IRC clients do highlighting when you start the line with their nick
08:55 LGee slefeuvre: this should be enough
08:56 LGee And most do tab completion as well
08:57 slefeuvre Samos123: But this seems to create another problem when checking vip__public at the end of the deployment
08:57 slefeuvre LGee: Thanks to your advices ;)
08:57 LGee :)
08:59 kszukielojc joined #fuel
08:59 slefeuvre Just to have some experience feedbacks, someone already deployed FUEL 7.0 environement with only 2 controller and 3 compute nodes ?
09:07 LGee I'm deploying 7.0 in a single server env
09:17 fatdragon joined #fuel
09:26 nineseconds joined #fuel
09:41 AlexPost joined #fuel
10:18 fatdragon joined #fuel
10:23 javeriak joined #fuel
10:41 LGee joined #fuel
10:43 Miouge joined #fuel
10:48 Miouge joined #fuel
10:49 BobBall joined #fuel
11:12 LGee joined #fuel
11:19 fatdragon joined #fuel
11:54 javeriak joined #fuel
12:20 fatdragon joined #fuel
12:42 DavidRama joined #fuel
12:55 v1k0d3n joined #fuel
13:01 v1k0d3n joined #fuel
13:02 slefeuvre left #fuel
13:03 slefeuvre joined #fuel
13:03 slefeuvre hi all
13:04 slefeuvre I get an error at the end of deployment FUEL 7.0 when this command is launched : /etc/puppet/modules/osnailyfacter/modular/ceph/ceph_ready_check.rb
13:05 slefeuvre After some search, it's look like I have the line in cinder.conf in [databases] section, that is "connection=..." that should not be "sql_connection=..." ?
13:06 v1k0d3n joined #fuel
13:08 krypto joined #fuel
13:10 v1k0d3n joined #fuel
13:14 parlos joined #fuel
13:15 v1k0d3n joined #fuel
13:23 krypto joined #fuel
13:24 javeriak joined #fuel
13:30 krypto joined #fuel
14:01 pbelamge joined #fuel
14:19 MikeG451 joined #fuel
14:21 fatdragon joined #fuel
14:24 raunak joined #fuel
14:24 Miouge joined #fuel
14:25 parlos joined #fuel
14:43 Miouge joined #fuel
14:47 Miouge joined #fuel
15:04 fatdragon joined #fuel
15:06 javeriak joined #fuel
15:20 xarses joined #fuel
15:57 eric_lopez joined #fuel
15:57 gardlt_ joined #fuel
16:05 raunak joined #fuel
16:05 eric_lopez joined #fuel
16:19 Jabadia joined #fuel
16:37 HeOS joined #fuel
17:09 raunak joined #fuel
17:12 artem_panchenko joined #fuel
17:29 xarses slefeuvre: can controller 1 ping the public gateway, and the other nodes public interfaces from the haproxy namespace?
17:30 xarses slefeuvre: why only 2 controllers? there is little protection from easy to happen split brains
17:37 gardlt_ joined #fuel
17:39 jerryz joined #fuel
17:42 gomarivera joined #fuel
17:50 Miouge joined #fuel
17:57 Jabadia joined #fuel
18:13 gardlt joined #fuel
18:35 Miouge joined #fuel
18:36 noons joined #fuel
18:36 krobzaur joined #fuel
18:37 noons Hey guys working on a new fuel deployment and it seems to be failing on the creation of the ceph stack
18:37 noons oddly enough looking at the logs it seems to be deploying osd's everywhere even though just two hosts were destined for ceph.. is that typical?
18:38 xarses noons: the ceph monitors are installed on all the controllers
18:39 xarses as well as radosgw (if you chose that)
18:39 xarses I'd need to understand what the error messages are
18:42 noons basically after the 1800 timeout it fails waiting for the ceph cluster to become healthy
18:42 noons looking at the ceph servers I just see "stuck unclean since forever, current state creating, last acting"
18:45 noons maybe I am best checking with the ceph guys, just odd that I haven't really seen any bugs related to this in fuel
18:46 xarses which fuel version, number of osd nodes?
18:48 xarses noons: also how many ceph-osd roles
18:48 xarses (hosts)
18:53 noons fuel version 8 and 2 ceph-osd roles
19:00 xarses and the output of `ceph -s`?
19:01 noons http://pastebin.com/95Nhrv9e
19:02 xarses did puppet fail on the ceph osd's? you have none active
19:02 xarses the place in the puppet.log will start with something like "MODULAR: ceph-osd"
19:05 noons oddly enough I see nothing about ceph in the puppet log on fuel.. I see that puppet completed but it doesnt look like it actually deployed ceph.. Let me check the actual server itself
19:13 noons it looks like it puppet ran through the OSD, I did see some of these errors directly after the MODULAR: ceph-osd "Cannot find datafile /etc/hiera/override/configuration/cluster.yaml, skipping"
19:17 noons but yea no osd's on these servers..
19:19 xarses seem that puppet should have failed on those 2 nodes
19:27 noons yea weird, says the puppet run completed successfully
19:35 xarses can you pate the log between that MODULAR: and the next occurance in the log?
19:41 xarses noons: ^
19:42 noons sure
19:47 noons https://pastebin.com/VWLnt6xw
19:48 noons hang on went in as private
19:50 gardlt joined #fuel
19:51 noons xarses: sorry about that should work now
19:54 xarses noons: ok, so you should have a ceph.conf and packages on the node correct?
19:55 noons that is correct
19:56 xarses can you post the output from https://gist.github.com/xarses/de6603e8b0009a250d73
20:14 noons joined #fuel
20:18 noons xarses: getting a syntax error with your script, taking a look at it now
20:26 xarses hmm, I thought it was working
20:26 xarses basically, we want the part GUID from the OSD partitions
20:31 raunak joined #fuel
20:32 noons hmm wth... doesn't look like any of the drives were even paritioned..
20:33 noons This should be handled by fuel no?
20:33 xarses yes
20:33 xarses so lets look at the bootstrap/fuel-agent logs then
20:34 xarses you can also look at the partations in the GUI and see what it had
20:35 noons In the gui it shows system on sdb then sdc,sdd,sde,sdf,sdg are all tagged as ceph
20:35 noons let me check the bootstrap logs
20:35 noons sdb seems to be the only one with a parition
20:49 eric_lopez joined #fuel
20:57 sbfox joined #fuel
21:11 eric_lopez joined #fuel
21:15 ralfba joined #fuel
21:15 ralfba hi ... i need to figure out how i can raise the kernel level within the deployment cobbler is doing on the target hosts
21:16 ralfba currently the deployment of ubuntu 14.04.xx still utilizes a 3.13.xx Kernel and not vivid or wily
21:18 ralfba i already tried to change  /var/lib/cobbler/kickstarts/ubuntu-amd64.preseed
21:18 ralfba but this had no effect unfornately
21:20 xarses ralfba: what version of fuel?
21:21 ralfba 8.0 ...
21:22 ralfba i successfully put in a newer kernel within the bootstrap-image, but when FUEL/cobbler deploys ubuntu, it get's down to 3.13.xx again
21:23 xarses Its based on the current default kernel in ubuntu's thrusty repos
21:24 ralfba that's what i thought too ... but default would implement already 4.xx
21:24 ralfba but within the deployment fuel does, it is still 3.13.xx
21:25 xarses again, its based on ubuntu's packages
21:25 xarses let me figure out where that is so you can change it
21:25 ralfba ok - first things first ... am i right to change the the cobbler kickstart file?
21:25 xarses no
21:26 xarses we don't provision with cobbler since 6.1
21:26 xarses its just used to deliver pxe and choose tftp or disk
21:26 ralfba ok - another point i need to learn
21:27 xarses we use fuel-agent in side the bootstrap image to write the image that is build on the fuel node when you first deploy the cluster
21:31 xarses on the fuel node a tool 'fa_build_image' runs to build it
21:31 ralfba ok
21:31 xarses you can see it's call in the astute.log
21:32 xarses heres the package linux-image-generic-lts-trusty
21:32 xarses which is why its 3.13 still
21:35 xarses sill looking for where this was defined in 8.0
21:35 xarses we moved it into the ui in 9.0
21:35 xarses @kozhukalov: still around? ^
21:36 ralfba ok ... nothing in production here, so i could also move ahead to use 9.0
21:37 xarses if you look at the astute log for the call, it might reference where the package list is
21:38 xarses I'm waiting for an 8.0 env to come up so I can look at it
21:39 kozhukalov in 8.0 fuel-agent used default package list since nailgun did not provide list of packages in input data https://github.com/openstack/fuel-agent/blob/stable/8.0/fuel_agent/drivers/nailgun.py#L693
21:40 Jabadia joined #fuel
21:41 xarses kozhukalov: thanks much
21:41 ralfba ok - so i change it there and it should go straight into the next deployment?
21:41 ralfba of course - thx for this quick reply !!!
21:42 kozhukalov yes
21:42 ralfba merci
21:43 kozhukalov it should be mcollective container /usr/lib/python2.7/site-packages/fuel-agent/drivers/nailgun.py
21:43 ralfba got it ...
21:43 ralfba thx again
21:45 Miouge joined #fuel
21:52 ralfba do i need to restart nailgun for getting it into play?
21:57 gardlt joined #fuel
22:13 AlexPost joined #fuel
22:15 eric_lopez joined #fuel
22:40 sbfox joined #fuel
23:11 sbfox joined #fuel
23:17 xarses joined #fuel
23:32 charz joined #fuel
23:40 jerryz joined #fuel
23:41 eric_lopez joined #fuel

| Channels | #fuel index | Today | | Search | Google Search | Plain-Text | summary