Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel-dev, 2014-06-16

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

All times shown according to UTC.

Time Nick Message
05:21 e0ne joined #fuel-dev
05:22 teran joined #fuel-dev
05:24 teran_ joined #fuel-dev
06:07 e0ne joined #fuel-dev
06:18 e0ne joined #fuel-dev
06:26 al_ex3 joined #fuel-dev
06:50 asledzinskiy joined #fuel-dev
07:02 salmon_ joined #fuel-dev
07:04 e0ne joined #fuel-dev
07:13 teran joined #fuel-dev
07:16 e0ne joined #fuel-dev
07:16 al_ex3 joined #fuel-dev
07:22 dnikishov joined #fuel-dev
07:36 jaypipes joined #fuel-dev
07:56 saibarspeis joined #fuel-dev
08:05 teran joined #fuel-dev
08:27 artem_panchenko joined #fuel-dev
08:33 sbanka joined #fuel-dev
08:46 teran joined #fuel-dev
08:51 salmon_ joined #fuel-dev
08:52 teran_ joined #fuel-dev
08:52 mattymo bogdando, ping
08:53 bogdando mattymo, pong
08:54 vkozhukalov joined #fuel-dev
08:54 mattymo looking at your logrotate fix
08:55 mattymo I don't like escaping on dashes so much
08:55 mattymo because what if user has a custom hostname
08:55 mattymo it might be better to escape on starting with number
08:55 mattymo because a hostname cannot start with a digit
08:55 mattymo (usually)
08:55 bogdando as far as I know wa always add a node number with dash?
08:55 mattymo we'll allow 100% custom hostnames at some point
08:55 bogdando no autonumeration?
08:56 mattymo let me find a thread on it
08:56 bogdando if so, we should use another dir for symlinks
08:56 bogdando so the nailgun part should be updated
08:57 bogdando sadly, logrotate doesn't support regular expressions, only wildcards
08:57 Tatyanka_Leontov joined #fuel-dev
08:58 mattymo yeah I hate that
08:58 bogdando and it cannot exclude dirs, afaik
09:04 akasatkin joined #fuel-dev
09:12 mattymo yeah exclusions are tough too
09:13 mattymo I guess dash is the best route for now and we should just merge it
09:20 mihgen bogdando: actually I have a question about logging too
09:20 mihgen are we gonna fix multiple files in 5.1 ?
09:21 mattymo mihgen, what does "multiple files" mean?
09:21 mihgen for every service?
09:21 bogdando mihgen, we have it already in master
09:21 mihgen bogdando: excellent! did we add it in 5.1?
09:21 bogdando that is the multiline fix https://review.openstack.org/#/c/90980/9/deployment/puppet/openstack/templates/00-remote.conf.erb
09:21 mihgen we don't have it in 5.0, right
09:22 bogdando *.* @@(o)<%= rserver['server']-%>:<%= rserver['port'] -%>;RemoteLog             here is a (o) mode for multilines...
09:22 bogdando yes, we merged it after 5.0
09:23 mihgen mattymo: we had multiple files per service. For example, nova has several python modules, so we had log file per module
09:23 mihgen now it is finally fixed as I understand, and it is very very cool :)
09:23 bogdando mihgen, oh, multiple *files* issue was fied as well
09:23 bogdando s/fixed
09:24 bogdando now we can see OSt logs in Fuel UI for all deployment scenarios
09:24 bogdando and they are differentiated by the service names instead of python qualnames, e.g. nova-compute
09:25 mattymo is it now 1 file per service instead of broken up?
09:27 bogdando mattymo, yes, there us a new option use_syslog_rfc_format that provides a service names in syslog app names (aka tags)
09:27 mihgen excellent
09:27 bogdando so we would have nova-conmpute, nova-conductor, etc
09:28 bogdando so, I'd say there is 1 file per *sub*service
09:29 mattymo as long as it is more appealing for a user, I'm for it
09:43 sbog joined #fuel-dev
09:43 sbartel joined #fuel-dev
09:43 sbartel hello
09:43 sbartel i have a question about fuel iso generation
09:44 holser bogdando, I think it should be this way <%= @rserver['server']-%>:<%= @rserver['port'] -%>
09:44 holser @
09:45 sbartel when making a make iso (version 4.1 for information) some times the process seems to freeze  during the dependancy resolution
09:46 sbartel with for example this message  : MySQL-server-wsrep-5.5.28_wsrep_23.7-12.x86_64 de proprietary dependancy problems  --> MySQL-server-wsrep conflicts with MySQL-server
09:46 b-zone joined #fuel-dev
09:46 sbartel any idea of what i can do to increase the process speed ?
09:48 mattymo sbartel, run it on a system with more cpu cores
09:48 mattymo dependency resolution could take 1 whole minute
09:48 sbartel ok I'll try thank you
09:48 mattymo but if it's hanging for 10 minutes, then there's something wrong.. maybe with network communication to a repo
09:49 sbartel it is hanging for more than 20 minues for some  dependancy
09:49 sbartel so guess it may be a network issu
09:49 mattymo that's bad
09:50 sbartel i'll try to increace the cpu core
09:50 sbartel and figure ou what can be wrong on my network configuration
09:51 sbartel thank you
10:09 teran joined #fuel-dev
10:10 teran_ joined #fuel-dev
10:11 teran_ joined #fuel-dev
10:51 hyperbaba joined #fuel-dev
10:51 hyperbaba Hi there
10:52 hyperbaba Does anybody knows how to enable quotas on openstack (using fuel 5.0) without complete redeployment ?
11:01 teran joined #fuel-dev
11:23 tnurlygayanov joined #fuel-dev
11:29 MiroslavAnashkin hyperbaba: Quotas is an OpenStack feature. Please check quota section in nova.conf <http://docs.openstack.org/trunk/config-reference/content/list-of-compute-config-options.html>,http://docs.openstack.org/user-guide-admin/content/dashboard_set_quotas.html and http://docs.openstack.org/user-guide-admin/content/cli_set_quotas.html
11:31 hyperbaba MiroslavAnashkin: Yes, i am aware of that. I was wondering can i call puppet recepie to do the changes for me after initial deployment.
11:31 MiroslavAnashkin hyperbaba: Basically you have to set `quota_driver=nova.quota.DbQuotaDriver` in /etc/nova.conf on all deployed nodes and restart OpenStack services
11:32 hyperbaba MiroslavAnashkin: and setup the default values in nova/neutron/cinder ?
11:34 MiroslavAnashkin quota is set in scope of openstack::controller class deployment. It is not too easy to set is with puppet call due to dependencies.
11:36 MiroslavAnashkin Yes, you may set your own default values for quotas in nova.conf
11:37 hyperbaba MiroslavAnashkin: Thank you very much for the information. Do you maybe know when andraug is online . I need to ask him couple of questions regarding ceph and openstack. He did make some changes in openstack regarding this.
11:43 MiroslavAnashkin andraug should be available in 4-5 hours. He is located in California and it it 4:40 AM there.
11:45 hyperbaba MiroslavAnashkin: Thank you once more.
13:07 IvanBerezovskiy joined #fuel-dev
13:09 IvanBerezovskiy hi, guys. I am working on ceilometer. I want to setup swift statistics from ceilometer in simple mode. Could anyone help and tell how object storage works in simple mode?
13:10 vrovachev joined #fuel-dev
13:21 e0ne joined #fuel-dev
13:36 evgeniyl bookwar: could you please retrigger https://fuel-jenkins.mirantis.com/job/stackforge-verify-fuel-web/3860/console ?
13:37 bookwar done
13:56 e0ne_ joined #fuel-dev
13:59 ikalnitsky joined #fuel-dev
14:03 jobewan joined #fuel-dev
14:57 angdraug joined #fuel-dev
15:07 xarses joined #fuel-dev
15:09 enikanorov_ joined #fuel-dev
15:14 mattymo angdraug, xarses https://review.openstack.org/#/c/98743/
15:14 mattymo for fixing the docker0 bug
15:15 mattymo but this patch breaks if you don't merge these 2 first: https://review.openstack.org/#/c/98973/ and https://review.openstack.org/#/c/99411/
15:20 jaypipes joined #fuel-dev
15:29 enikanorov__ joined #fuel-dev
15:34 angdraug mattymo: should 98743 be dependent on the other two?
15:34 angdraug it's not right now...
15:34 mattymo both 98973 first, then 99411, then 98743
15:35 mattymo s/both//
15:38 IvanBerezovskiy left #fuel-dev
15:41 angdraug can you rebase 98743 over 99411 then?
15:42 angdraug that would fix the -1's from Artem and yourself, right?
16:03 rmoe joined #fuel-dev
16:15 nuritv joined #fuel-dev
16:28 sbog joined #fuel-dev
16:28 sbog left #fuel-dev
16:39 angdraug joined #fuel-dev
16:45 xarses angdraug: https://review.openstack.org/#/c/99807/1/specs/5.1/ml2-neutron.rst
16:45 christopheraedo Hope someone can +2 this one soon, I'm updating docs and would like to refer to it. Excellent work to get someone a fuel development environment very quickly.  https://review.openstack.org/#/c/99298/
16:49 angdraug christopheraedo: did you test it?
16:49 angdraug ah, I see you comment
16:50 angdraug needs a rename and a comment banner on top
16:50 angdraug I'd call it prepare-build-env.sh or somesuch
16:50 angdraug and needs a comment on top to explain what this script is for
16:51 christopheraedo yes, I tested it under ubunu 14, 13 and 12 - it does what it is supposed to :)  I agree with your comments - I'm just glad to see we have it now, it's going to be a big part of helping people move faster with Fuel
16:51 angdraug christopheraedo: can you highjack this commit to do the rename/documenting? I'd gladly merge it hten
16:52 christopheraedo yes I would be happy to, I'll do that now
16:56 angdraug thanks!
17:19 angdraug mattymo: are you still around to rebase those 3 commits into a deps chain, or do you mind if I do?
17:27 bogdando remote logs issue https://bugs.launchpad.net/fuel/+bug/1328877 looks resolved with the patch suggested (tested manually)
17:27 bogdando so please review https://review.openstack.org/#/c/100273/
17:27 bogdando we need this fix for CI asap
17:32 angdraug bogdando: ok, I'm looking at it, lets see if CI comes back with a +1
17:39 Topic for #fuel-dev is now #fuel-dev Fuel dev docs http://docs.mirantis.com/fuel-dev/ | Paste here http://paste.openstack.org/ | IRC logs http://irclog.perlgeek.de/fuel-dev/ | gerrit traffic @ #fuel-tracker
17:40 e0ne joined #fuel-dev
17:41 bogdando angdraug, I was testing at the existing lab, and logs have returned back to /var/log/remote in rsyslog container at least. Need to figure out how ti 'fix' empty logs in UI tho - perhaps some other containers have to be restarted...
17:41 bogdando or we could use fresh ISO build from the patch, just for sure...
17:46 rmoe bogdando: I'm trying to understand bug 1329991
17:46 rmoe is the issue that the pattern matches BOTh the node-xxx.domain.tld directory and the symlink to it?
17:47 rmoe which then causes the same log files to be processed twice?
17:47 bogdando yup
17:47 rmoe ok
17:48 rmoe so it has nothing to do with having both 10-fuel.conf and 20-fuel.conf which appear identical?
17:49 bogdando yes, 10* and 20* configs ensure time based and size based criteria respectively
17:50 bogdando the fix is to either do not include symlinks from matching, or put all symlinks to another dir
17:50 bogdando the former is suggested as a patch
17:50 bogdando the 2nd could be possible only by the nailgun side
17:51 rmoe your fix looks good, I just wanted to make sure I fully understood the problem
17:55 angdraug bogdando: https://review.openstack.org/#/c/100273/6/deployment/puppet/openstack/manifests/logging.pp line 231
17:57 angdraug how is this related to the similar resource in rsyslog-only.pp?
18:00 angdraug client side only?
18:01 angdraug should they be based on the same template?
18:07 e0ne joined #fuel-dev
18:09 angdraug nm, figured it out. didn't realize we don't use class[nailgun] anymore
18:36 teran joined #fuel-dev
18:44 teran joined #fuel-dev
18:52 vkozhukalov joined #fuel-dev
19:03 e0ne joined #fuel-dev
19:09 artem_panchenko joined #fuel-dev
19:33 sbanka joined #fuel-dev
19:49 xarses angdraug: https://github.com/stackforge/fuel-library/blob/master/deployment/puppet/neutron/manifests/agents/l3.pp#L28-42
19:49 xarses can you explain the purpose?
19:58 angdraug (for the benefit of lurkers) as discussed offline, ^ is essentially to prevent the service from being started by dpkg
20:50 e0ne joined #fuel-dev
21:16 zynzel who is responsible for building packages for ubuntu? Im almost sure that in 4.X socket path in mysql was /var/lib/mysql/mysql.sock and now is /var/run/mysqld/mysqld.sock
21:16 zynzel the same with pidfile
21:17 angdraug zynzel: fuel-osci team builds packages for centos and ubuntu
21:18 angdraug it seems to me that they did the right thing, socket file doesn't belong in /var/lib
21:18 zynzel angdraug: in centos socket is in /var/lib/mysql ;)
21:19 zynzel and my mysql-wss RA search socket there..
21:20 zynzel but maybe old RA for galera, dont check DB health via socket/queries.
21:20 zynzel hmm, maybe this is good time to introduce different default for centos/ubuntu in RA.
21:20 zynzel i will check this tomorrow
21:21 zynzel angdraug: thanks
22:29 xarses angdraug: https://github.com/xarses/fuel-library/compare/bp;ml2-neutron?expand=1
23:47 xarses joined #fuel-dev

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