Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel-dev, 2014-06-24

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

All times shown according to UTC.

Time Nick Message
00:24 rmoe joined #fuel-dev
00:32 teran joined #fuel-dev
00:34 teran_ joined #fuel-dev
00:45 e0ne joined #fuel-dev
01:07 teran joined #fuel-dev
01:20 xarses joined #fuel-dev
01:38 teran_ joined #fuel-dev
01:45 e0ne joined #fuel-dev
01:47 e0ne joined #fuel-dev
02:09 teran joined #fuel-dev
02:40 teran_ joined #fuel-dev
02:45 e0ne joined #fuel-dev
02:52 novel joined #fuel-dev
02:58 aedocw joined #fuel-dev
03:02 e0ne joined #fuel-dev
03:11 teran joined #fuel-dev
03:39 vkozhukalov joined #fuel-dev
03:42 teran_ joined #fuel-dev
03:45 e0ne joined #fuel-dev
04:15 teran joined #fuel-dev
04:25 a_teem joined #fuel-dev
04:25 e0ne joined #fuel-dev
04:26 al_ex6 joined #fuel-dev
04:46 teran_ joined #fuel-dev
05:18 teran joined #fuel-dev
05:49 teran_ joined #fuel-dev
06:15 dshulyak_ joined #fuel-dev
06:20 teran joined #fuel-dev
06:29 vkozhukalov joined #fuel-dev
06:48 sambork joined #fuel-dev
06:48 vkozhukalov joined #fuel-dev
06:51 teran_ joined #fuel-dev
06:58 vkozhukalov1 joined #fuel-dev
07:01 xarses joined #fuel-dev
07:02 teran joined #fuel-dev
07:04 artem_panchenko joined #fuel-dev
07:05 salmon_ joined #fuel-dev
07:10 akasatkin joined #fuel-dev
07:11 dshulyak_ xenolog: hi, can you take a look at this bug //bugs.launchpad.net/fuel/+bug/1322553? it affects library somehow or not?
07:23 tnurlygayanov joined #fuel-dev
07:28 accela-dev joined #fuel-dev
07:29 aglarendil joined #fuel-dev
07:29 xenolog joined #fuel-dev
07:29 MiroslavAnashkin joined #fuel-dev
08:07 al_ex6 joined #fuel-dev
08:08 vkozhukalov joined #fuel-dev
08:10 brain461 joined #fuel-dev
08:14 fandi joined #fuel-dev
08:15 ddmitriev joined #fuel-dev
08:26 sambork joined #fuel-dev
08:32 bogdando joined #fuel-dev
08:44 teran joined #fuel-dev
09:01 vkozhukalov1 joined #fuel-dev
09:04 sbartel hello
09:04 sbartel concerning this bug
09:04 sbartel https://bugs.launchpad.net/fuel/+bug/1326476
09:05 teran joined #fuel-dev
09:05 sbartel is there any workaround to generate the iso on ubuntu 14.04?
09:08 bookwar sbartel: install multistrap package from Ubuntu 12.04
09:09 sbartel ok
09:09 sbartel I'll try that
09:09 sbartel i tried the opposite (installing multistrap from 14.10)
09:10 e0ne joined #fuel-dev
09:12 bookwar sbartel: and it didn't work?
09:12 sbartel it solve the issu about multistrap ($forceyes variable error)
09:13 bookwar yes, but that's not enough
09:13 sbartel but then I have problem with base-files and  bash  as describe in the fuel bug 1326476
09:13 bookwar we really need to file a bug in upstream for this, just need some time to get a bit deeper in the issue and reproduce it with minimal set of dependencies
09:14 bookwar i doubt Ubuntu guys will be interested in building our iso :)
09:15 sbartel ok
09:15 sbartel this this the ubuntu bug about multistrap in 14.04 https://bugs.launchpad.net/ubuntu/+source/multistrap/+bug/1313787
09:20 bookwar sbartel: that was filed by monester
09:22 sbartel is he someone from mirantis?
09:23 bookwar yes
09:23 sbartel anyway what is strange that installing multistrap and realpath from precise seems to correct the problem
09:24 bookwar he is our second devops engineer
09:24 sbartel so there is still an issue in utopic version of multistrap with iso process
09:24 teran joined #fuel-dev
09:25 tuve_ joined #fuel-dev
09:26 brain461 joined #fuel-dev
09:35 sambork joined #fuel-dev
09:38 enikanorov_ joined #fuel-dev
09:44 rshon joined #fuel-dev
09:59 evgeniyl Guys, I reverted patch which was merged yesterday, because of critical bug https://review.openstack.org/#/c/102164/ review please, and we start new build image for our qa team.
10:01 salmon_ evgeniyl: what kinf od bug?
10:03 evgeniyl salmon_: after provisioning, agent doesn't retrieve nailgun_uid which is used for mcollective config, as result orchestrator thinks that all nodes stuck in provisioning (because it cannot find them by names).
10:04 evgeniyl salmon_: I want to rollback it, reopen the ticket and we will have time to propose required fix and ask our qa team to test it separately, and then merge it.
10:05 evgeniyl sambork: ^
10:05 salmon_ evgeniyl: ok, I see
10:06 sambork Ok, I will take a look on this issue
10:08 evgeniyl sambork: ok, could you please set +1 if revert is ok?
10:11 sambork done
10:16 akasatkin1 joined #fuel-dev
10:18 odyssey4me joined #fuel-dev
10:20 tzn joined #fuel-dev
10:22 tnurlygayanov hi there
10:22 tnurlygayanov I have an idea
10:23 tnurlygayanov Sahara project requires some advanced chages in the default configuration (which allow to increase the performance of Sahara cluster)
10:24 tnurlygayanov can we change some configuration to improve the performance of Sahara clusters, if user will select 'Instal Sahara'?
10:25 dmitryme tnurlygayanov: please elaborate, what do think should be changed?
10:26 odyssey4me Anyone else getting package conflicts when trying to 'make all' from fuel-main/master?
10:26 tnurlygayanov dmitryme hm... I can't describe all possible options (need to ask Sahara team), but I know, for example, about option for HDD disks of VMs - if we will configure no-random access for hdd disks of VMS it will increase the performance of clusters
10:27 tnurlygayanov dmitryme probably you know some parameters which will allow to increase performance too?
10:29 tnurlygayanov I know that Andrey Lazarev knows many parameters, which we should change in default configuration to improva Sahara performance
10:29 brain461 joined #fuel-dev
10:29 mattymo odyssey4me, I just built an iso a few minutes ago from master
10:30 tnurlygayanov but before we will do this I want to confirm that we are ready to improve performance of Sahara in 5.1
10:30 mattymo you should wipe your local mirror, pull latest master, and try again
10:30 mattymo odyssey4me, there are a # of conflicts, but there is a check for an exact number
10:30 odyssey4me mattmo - will do, I'm using Ubuntu 13.04 if that matters
10:31 saibarspeis Can you tell me what are the docker0 and vethXX[a-z] in the new Fuel 5.0 ?
10:31 saibarspeis I can't find any documentation about it
10:32 saibarspeis I'm of course talking about fuelmenu in Fuel 5.0
10:35 saibarspeis Anyone ?
10:36 odyssey4me saibarspeis - from previous conversations I've seen on here, it's a bad idea to muck with anything other than the eth0 settings
10:36 dmitryme tnurlygayanov: I know only of one parameter Andrew tweaked - disk_cachemodes=‘writeback’. It makes writing to disk more ‘aggressive’ in terms that disk write returns once data is written to the cache, before it is written to disk. tI is acceptable to tweak it for Hadoop, but it might not be good for other VMs.
10:36 odyssey4me the others are used for docker and at the moment a lot of assumptions are made in the deployment about those - they're never exposed outside of the fuel master anyway
10:38 saibarspeis odyssey4me: Thank you so much. That's what I needed to know. So the problem of changing anything in configuration via fuelmenu remains.
10:38 saibarspeis I install fuel 4.1 or 5.0 using it's PXE interface (eth0) temporary routed to the internet during the install stage
10:38 saibarspeis Then I configure eth1 as my "outside" interface, where I want my dashboard
10:57 evgeniyl nurla, bookwar merged revert.
10:57 nurla evgeniyl: nice!
11:01 e0ne joined #fuel-dev
11:19 vrovachev joined #fuel-dev
11:25 brain4611 joined #fuel-dev
11:27 brain4611 joined #fuel-dev
11:50 ogelbukh gentlemen, quick question
11:51 mattymo gentleperson, go ahead
11:51 mattymo ogelbukh, ^
11:51 ogelbukh is it possible to remove a node from fuel unallocated nodes?
11:51 ogelbukh i removed it from all environments and basically turned it off
11:51 mattymo I believe it can be purged from the DB but it will be discovered again if you bootstrap it
11:51 ogelbukh fuel detects that node is gone, but keeps it configured
11:52 mattymo evgeniyl, any comment here^?
11:52 ogelbukh so, no way to decomission it without manual db-level intervention?
11:52 mattymo ogelbukh, I believe no
11:53 ogelbukh ok, thanks
11:53 e0ne_ joined #fuel-dev
11:56 evgeniyl ogelbukh: no, there is no way to remove unallocated node from cluster
11:57 ogelbukh evgeniyl: mattymo: thank you
12:11 brain461 joined #fuel-dev
12:22 brain461 joined #fuel-dev
12:33 sambork joined #fuel-dev
12:36 guillaume_ joined #fuel-dev
12:39 dshulyak_ joined #fuel-dev
12:41 ddmitriev joined #fuel-dev
12:43 jaypipes joined #fuel-dev
13:10 odyssey4me ogelbukh - there is, but only through a direct API query
13:10 odyssey4me let me dig that up for you
13:10 odyssey4me https://bugs.launchpad.net/fuel/+bug/1326116
13:11 odyssey4me the description indicates how to do it with a curl query
13:13 mattymo ikalnitsky, ping
13:13 ikalnitsky mattymo: i'm here
13:13 mattymo ikalnitsky, re: 1333631 - I know why this fails
13:13 mattymo after upgrade, /var/www/nailgun on the nginx container resets to owner root:root with permission 0700 (not executable by apache user)
13:14 mattymo therefore the forbidden error
13:14 mattymo in start.sh script we chmod it: https://github.com/stackforge/fuel-main/blob/master/docker/nginx/start.sh#L6
13:14 ikalnitsky mattymo: yeah, i've figured out that and assign it to evgeniyl. so how it should be fixed?
13:18 odyssey4me_ joined #fuel-dev
13:20 evgeniyl mattymo: hmm, who resets this permissions if we set them in starts.sh? puppet?
13:21 mattymo evgeniyl, I believe it's a problem when the new container mounts this volume
13:21 mattymo it resets the permissions
13:22 mattymo it's because the lxc/libcontainer doesn't enable userspace isolation or passthrough or whatever it's called
13:22 mattymo if we ran nginx and forced it to be root, it would work around the issue... but that opens up worse problems
13:23 mattymo I'm not 100% sure what else could reset this
13:27 evgeniyl mattymo: ok, I'll debug this problem, also I saw the similar problem on Nikolay's env (on master node without upgrades), he had problem with static files which had root ownership, as result nginx was not able to serve static files.
13:29 mattymo evgeniyl, that's unusual too. the files should be world readable
13:29 mattymo and we can fix this in the rpm spec for nailgun
13:31 casanch1 joined #fuel-dev
13:42 vkramskikh hi guys
13:43 vkramskikh we try to solve this https://bugs.launchpad.net/fuel/+bug/1325889
13:43 vkramskikh and there is a proposal to make it look somehow like this: http://172.18.204.100:8000/#cluster/1/nodes
13:44 vkramskikh http://i.imgur.com/JCGAwVT.png
13:44 vkramskikh what do you think? do you like it?
13:46 ikalnitsky vkramskikh: why i see cluster info here http://172.18.204.100:8000/#clusters ? i mean, why i see some cluster info on the page where all clusters are shown ?
13:48 vkramskikh ikalnitsky: it's not a final version, just an example
13:48 vkramskikh if we decide to reimplement it like this, we will implement it properly
13:56 vkozhukalov joined #fuel-dev
13:56 mattymo vkramskikh, it's because of permission of /var/www/nailgun
13:57 mattymo chmod -R 755 /var/www/nailgun from nginx container and try again
13:57 mattymo and if you can come up with steps to reproduce this, I would love it. something unexpected is resetting these permissions
14:00 ikalnitsky vkramskikh: btw, why we need this info? we can see this info when clicks onto cluster name?
14:00 vkramskikh yes, but it is very unobvious
14:00 vkramskikh so we may want to move this info to a better place :)
14:04 ogelbukh odyssey4me: thank you, it works!
14:06 odyssey4me ogelbukh: no problem - happy to help :)
14:06 odyssey4me got that little gem from xarxes, so thanks to him really :)
14:10 brain461 joined #fuel-dev
14:17 jobewan joined #fuel-dev
14:20 teran joined #fuel-dev
14:34 teran joined #fuel-dev
14:45 IvanBerezovskiy joined #fuel-dev
14:51 jobewan joined #fuel-dev
14:52 jobewan joined #fuel-dev
14:55 jobewan joined #fuel-dev
14:58 jobewan joined #fuel-dev
15:01 mrasskazov :w
15:02 mrasskazov sorry
15:13 brain461 joined #fuel-dev
15:14 teran_ joined #fuel-dev
15:17 vkozhukalov joined #fuel-dev
15:31 tuve joined #fuel-dev
15:41 izinovik fuelers, can somebody explain me what all these messages on fuel master node console mean: http://pbrd.co/1lN0vJl
15:48 izinovik btw, is it normal for dhcrelay on FUEL master node NOT to listen on eth0, only on docker0?
16:03 IvanBerezovskiy left #fuel-dev
16:06 rmoe joined #fuel-dev
16:20 evg izinovik: it should be running with "-i eth0 -i docker"
16:20 izinovik evg, I already fixed this issue by hands
16:20 izinovik but why eth0 was absent in /etc/sysconfig/dhcrelay ?
16:23 evg izinovik: i've just looked into a fresh master: 'INTERFACES="eth0 docker0"' in my case
16:30 artem_panchenko joined #fuel-dev
16:32 evg izinovik: bwt, in your console your're seeing bitcoin's mining proccess
16:47 nzrf_ joined #fuel-dev
16:53 izinovik evg, how can I access FUEL masters console, just switch with Alt+F2, Alt+F3, Alt+F4, etc ?
16:57 evg izinovik: yes, it's just centos node
16:57 xarses joined #fuel-dev
16:57 evg izinovik: may be you mean ssh?
16:58 angdraug joined #fuel-dev
16:58 izinovik evg, no,no I mean ordinary console or physical console
16:58 dshulyak_ joined #fuel-dev
16:59 evg izinovik: Alt+Fx
16:59 izinovik I just restart my FUEL master and again eth0 is absent in /etc/sysconfig/dhcrelay
17:01 evg izinovik: run   ps axf | grep dhcrelay
17:01 evg izinovik: or, sorry. i've misread
17:02 izinovik evg, dhcrelay is running (as service dhcrelay status reports)
17:02 izinovik evg, but it only runs this way: /usr/sbin/dhcrelay -i docker0 172.17.0.2
17:02 evg izinovik: cat /etc/sysconfig/dhcrelay | grep INTERFACES
17:03 izinovik [root@fuel log]# grep INTERFACES /etc/sysconfig/dhcrelay
17:03 izinovik INTERFACES=" docker0"
17:06 evg izinovik: do you really have "eth0" interface?
17:06 izinovik evg, you mean do I see it in `ip addr` and `ifconfig` output? Yes.
17:08 MiroslavAnashkin http://docs.mirantis.com/fuel/fuel-5.0/user-guide.html#changing-pxe-network-parameters-during-installation
17:09 MiroslavAnashkin This section appeared on our site today
17:09 evg izinovik: how have you installed your master node? have you use fuelmenu? is it hw or vm installation?
17:38 e0ne joined #fuel-dev
17:42 e0ne joined #fuel-dev
17:53 tuve joined #fuel-dev
17:54 odyssey4me is there a way that I can make nailgun (in fake mode) log each function/procedure/class that it's going through? I have a failure in a change I'm making and can't seem to find where it's failing.
17:54 odyssey4me the debug logging is just not enough
18:04 vkozhukalov joined #fuel-dev
18:22 ogelbukh haha
18:22 ogelbukh it's occured to me that deleted node stays in cobbler
18:23 ogelbukh after it's deleted from db
18:24 ogelbukh isn't is a bug?
18:24 ogelbukh or nailgun doesn't manage cobbler..
18:28 e0ne joined #fuel-dev
18:34 tuve joined #fuel-dev
18:36 brain461 joined #fuel-dev
18:37 odyssey4me ogelbukh: that sounds like a bug
18:38 odyssey4me (please check if one exists, otherwise register it)
18:38 odyssey4me (anyone) if I've changed the db model, how do I prepare the migration script?
18:42 odyssey4me it appears that alembic is being used
18:47 odyssey4me nvm - worked it out
18:57 tuve joined #fuel-dev
19:25 tuve_ joined #fuel-dev
20:06 e0ne joined #fuel-dev
20:12 e0ne joined #fuel-dev
20:20 e0ne joined #fuel-dev
20:26 tuve joined #fuel-dev
20:29 e0ne joined #fuel-dev
20:32 teran joined #fuel-dev
20:46 aleksandr_null Hi Guys,
20:47 aleksandr_null Hope all of you have a good day :)
20:48 aleksandr_null My question is how long it should take to provision ubuntu+ha+ceph-osd (only controllers) ?
20:50 MiroslavAnashkin 0.5h/controller
21:16 tuve joined #fuel-dev
21:38 enikanorov_ joined #fuel-dev
21:44 xarses angdraug: https://bugs.launchpad.net/fuel/+bug/1305852 ??
21:45 teran joined #fuel-dev
22:07 e0ne joined #fuel-dev
22:07 teran_ joined #fuel-dev
22:09 e0ne_ joined #fuel-dev
22:37 xarses angdraug: poke https://bugs.launchpad.net/fuel/+bug/1330610
22:50 rmoe angdraug: xarses: https://review.openstack.org/#/c/102375/
22:51 xarses +1
23:04 Guest4877 joined #fuel-dev
23:07 xarses angdraug: rmoe https://review.openstack.org/102381
23:09 e0ne joined #fuel-dev
23:17 angdraug https://bugs.launchpad.net/fuel/+bug/1322102/comments/14
23:21 xarses angdraug: rmoe ^^ new rev
23:21 xarses angdraug: wow
23:37 xarses angdraug: rmoe https://review.openstack.org/102387
23:43 jaypipes joined #fuel-dev

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