Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel-dev, 2014-11-25

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

All times shown according to UTC.

Time Nick Message
00:28 AlexF joined #fuel-dev
00:40 AlexF joined #fuel-dev
00:58 proberts joined #fuel-dev
01:17 rmoe joined #fuel-dev
01:52 xarses joined #fuel-dev
02:09 teran joined #fuel-dev
03:01 rtau joined #fuel-dev
03:09 proberts joined #fuel-dev
03:13 LiJiansheng joined #fuel-dev
03:39 xarses joined #fuel-dev
03:41 teran joined #fuel-dev
04:08 xenolog joined #fuel-dev
04:10 akislitsky_ joined #fuel-dev
05:11 bdudko joined #fuel-dev
05:11 AlexF joined #fuel-dev
05:11 teran joined #fuel-dev
05:12 apalkina_ joined #fuel-dev
05:22 proberts joined #fuel-dev
05:38 AlexF joined #fuel-dev
06:28 monester_laptop joined #fuel-dev
06:42 Longgeek joined #fuel-dev
06:42 Longgeek joined #fuel-dev
06:55 prmtl joined #fuel-dev
07:00 teran joined #fuel-dev
07:01 Longgeek joined #fuel-dev
07:08 dklepikov joined #fuel-dev
07:14 Longgeek joined #fuel-dev
07:14 AlexF joined #fuel-dev
07:19 AlexF joined #fuel-dev
07:21 dshulyak_ joined #fuel-dev
07:23 adanin joined #fuel-dev
07:25 sambork joined #fuel-dev
07:29 barthalion how much do we care about posix compatibility of our shell scripts?
07:37 teran joined #fuel-dev
07:40 e0ne joined #fuel-dev
07:53 sbog Hi, guys. Can you help with review, please -  https://review.openstack.org/#/c/135905/ and https://review.openstack.org/#/c/136283/ ?
07:57 e0ne joined #fuel-dev
07:59 gilmeir joined #fuel-dev
08:03 e0ne joined #fuel-dev
08:09 stamak joined #fuel-dev
08:10 romcheg joined #fuel-dev
08:12 e0ne joined #fuel-dev
08:15 hyperbaba joined #fuel-dev
08:19 azemlyanov joined #fuel-dev
08:29 aviramb joined #fuel-dev
08:33 aviramb_ joined #fuel-dev
08:33 salmon_ joined #fuel-dev
08:40 holser barthalion: sure
08:40 holser barthalion, it’s not hard but time consuming ;)
08:42 dklepikov Hello.Have you  any ideas why running cron job on controller ruby hangs?
08:42 dklepikov and fuel stops see controller node online
08:51 HeOS joined #fuel-dev
08:52 breton morning
08:53 salmon_ dklepikov: by ruby, do you mean nailgun-agent ?
08:54 breton I'll annoy you with my yesterdays question again
08:54 dklepikov breton : yes
08:54 dklepikov salmon : yes
08:54 dklepikov breton : sorry
08:55 breton could you please explain me, what I need to do in order to understand, which token provider is set in puppet scripts?
08:55 breton this is keystone's init.pp http://paste.openstack.org/show/137473/
08:56 dklepikov salmon : all processes hang and not running again
08:56 breton what else I should look at?
08:56 dklepikov salmon : only after killing them.
08:57 salmon_ dklepikov: how do you know that it's caused by nailgun-agent?
08:57 breton the problem is that something thinks that PKI is used as token provider, but everything is configured for UUID tokens
08:58 dancn joined #fuel-dev
08:59 dklepikov salmon : on master node /var/log/docker-logs/nailgun/api.log  when i run script manually  fuel start to see node
08:59 dklepikov salmon : but it is only up to next check
09:02 salmon_ dklepikov: we have bug for something similar, can you download snapshot and attach it to this bug https://bugs.launchpad.net/fuel/+bug/1380786 ?
09:04 dklepikov salmon : it looks like mine. i got this error after cluster recovery
09:05 sambork joined #fuel-dev
09:05 salmon_ dklepikov: so we need your logs now :)
09:11 zynzel guys, i prepared zabbix in HA for 5.1 (code is ready) https://blueprints.launchpad.net/fuel/+spec/zabbix-ha
09:11 zynzel can anybody look at this? BP is in initial phase...
09:17 dklepikov <salmon_> it is too big, launchpad  can't upload it
09:18 salmon_ dklepikov: can you upload it in other place like google drive and give us a link?
09:18 dklepikov salmon_ : maybe i can send separate logs. Tell me what logs do you need.
09:21 salmon_ dklepikov: ok, wait
09:21 dklepikov salmon_ : check e-mail
09:22 salmon_ dklepikov: thx, I will attach required files
09:22 dklepikov salmon_ : i can send lsof from hanging processes on node
09:23 salmon_ dklepikov: sure, It may help
09:24 salmon_ dklepikov: which node hungs?
09:25 e0ne joined #fuel-dev
09:25 dklepikov salmon_ : node-9
09:26 salmon_ thx
09:26 dklepikov i attach to launchpad lsof data
09:55 salmon_ dklepikov: hmm, can you kill all this programs from ps axwu | grep nail and delete /var/lock/agent.lock
10:02 dshulyak sambork:  salmon_: evgeniyl__: ikalnitsky: check my suggestion about moving deployment serialization into background https://review.openstack.org/#/c/135323/
10:02 dshulyak i believe it can be done quite easily right now
10:04 salmon_ dshulyak: yup, background task make sense
10:06 ikalnitsky personally, i'd like to keep code as readable as possible.. and it looks like background tasks are going to help us with it.
10:07 bogdando https://bugs.launchpad.net/fuel/+bug/1396016/comments/3 do we support re-assign of primary controller? looks like this bug is related to https://bugs.launchpad.net/fuel/+bug/1390397
10:07 sambork dshulyak: agree
10:10 dshulyak bogdando: we can mark that bug as duplicate or i will assign it to myself
10:10 bogdando dshulyak, thanks
10:10 bogdando dshulyak, and please update the 5.1.1 series for original bug, it has only 6.0 target
10:12 alexbh joined #fuel-dev
10:13 dshulyak bogdando: actually it is a bit different, i dont get what is "old public vip"
10:13 dshulyak vip is assigned per cluster
10:14 dshulyak how it can be "old"
10:14 dshulyak skolekonov: ^^
10:14 bogdando dshulyak, yes, it looks unclear to me as well. But anyway if we support changing the layout of primary controller role?
10:15 bogdando if we do not yet, perhaps we should prohibit removal of primary controler then scaling
10:15 dklepikov <salmon_> done
10:16 dklepikov salmon_ : it runs again and still the same
10:17 adanin joined #fuel-dev
10:17 skolekonov dshulyak, I mean that when we remove primary controller another controller becomes primary and vip should point to it
10:17 dshulyak bogdando: afaiu we can not prohibit removal of primary controller, we dont know what controller is primary at a certain point in time
10:18 bogdando dshulyak, so, we allow changing the primary role, but there is a related bug about lowering of uuid, right?
10:19 bogdando dshulyak, looks like the bug  https://bugs.launchpad.net/fuel/+bug/1396016 should be checked thoroughly then either is the same case as #1390397 or not
10:21 dshulyak bogdando: skolekonov: yes, the root cause may not be the same, i will check logs, and update bug if we can fix it by proper orchestration
10:23 dancn joined #fuel-dev
10:27 salmon_ dklepikov: what happen when you tun full command:
10:27 salmon_ /bin/sh -c flock -w 0 -o /var/lock/agent.lock -c "/opt/nailgun/bin/agent 2>&1 | tee -a /var/log/nailgun-agent.log  | /usr/bin/logger -t nailgun-agent"
10:33 dklepikov salmon_ : killall, run a command - cli waiting for command ends, in another terminal i see all new processes
10:34 salmon_ dklepikov: does it end or hung?
10:34 dklepikov salmon_ : hung
10:43 azemlyanov joined #fuel-dev
10:49 monester_laptop joined #fuel-dev
10:54 vtzan joined #fuel-dev
11:02 ignatenko joined #fuel-dev
11:27 stamak joined #fuel-dev
11:36 dklepikov Can fuel  save config for setup controller node and apply in to another controller  node?
12:05 teran joined #fuel-dev
12:06 teran_ joined #fuel-dev
12:06 akislitsky__ joined #fuel-dev
12:07 tzn joined #fuel-dev
12:10 monester_laptop joined #fuel-dev
12:12 salmon_ dklepikov: every node has different config
12:24 bpiotrowski joined #fuel-dev
12:41 corepb joined #fuel-dev
12:44 Longgeek_ joined #fuel-dev
12:47 corepb_ joined #fuel-dev
12:47 corepb_ joined #fuel-dev
12:54 barthalion left #fuel-dev
12:58 teran joined #fuel-dev
12:58 teran_ joined #fuel-dev
13:01 bpiotrowski joined #fuel-dev
13:23 Longgeek joined #fuel-dev
13:31 omolchanov joined #fuel-dev
13:31 ikalnitsky joined #fuel-dev
13:33 tzn joined #fuel-dev
13:46 IvanBerezovskiy joined #fuel-dev
13:57 enikanorov_ joined #fuel-dev
14:23 mgariepy hello, where do we report security issue with Fuel ?
14:40 akupko joined #fuel-dev
14:45 MiroslavAnashkin mgariepy: If the issue is related to one of common Linux component or you have doubts - then here: https://bugs.launchpad.net/mos/ If you sure the issue is related to Fuel library - then here: https://bugs.launchpad.net/fuel/+bug/
14:46 MiroslavAnashkin Please note - for additional safeyt Mirantis converts all security issues to the private Launchpad bugs, visible to the proper team only.
14:50 sovsianikov joined #fuel-dev
14:51 mgariepy ok thanks.
15:00 jobewan joined #fuel-dev
15:05 AlexF joined #fuel-dev
15:11 proberts joined #fuel-dev
15:32 mihgen folks, if there are no objections on workaround  "Use CentOS 6.6 kernel in boot/install stage of nodes", I'm merging the patch… https://review.openstack.org/#/c/135383/
15:36 dshulyak mihgen: wait a min
15:37 angdraug joined #fuel-dev
15:39 dshulyak mihgen:  there is no versioning for cobbler profiles right now, so this kernel will be used for previous environments as well
15:45 dshulyak evgeniyl__: mihgen: well, maybe i missed something and if it wont break upgraded envs - it is safe to merge
15:48 evgeniyl__ dshulyak: mihgen only after testing it'll be clear if it brakes something, and yes, dshulyak is correct, all of this changes (new kernel) will be used even for old environments
15:57 fandi joined #fuel-dev
16:05 bogdando regarding https://bugs.launchpad.net/fuel/+bug/1382527/comments/4. What was an initial point in adding mounts as a conatiners, and why we could switch to host mounts now?
16:07 bogdando So, indeed, what was the point of making the same things in the different ways?
16:09 dshulyak_ joined #fuel-dev
16:15 bogdando I understand that "Direct mounts proved to be a better alternative than trying to manage symbolic links to work around LXC limitations", but it is not clear why we couldn't use them initially
16:25 mihgen dshulyak: evgeniyl__: guys, it's new kernel only for installation of new nodes
16:25 mihgen basically we are using this 6.6 kernel in target system
16:32 IvanBerezovskiy left #fuel-dev
16:34 mattymo mihgen, bogdando angdraug MOS team needs a quick merge on their spec: https://review.openstack.org/#/c/134937
16:35 mattymo questions to dmitryme if you have any
16:36 angdraug why the rush? don't we have critical 5.1.1 bugs we have to deal with first?
16:37 mihgen mattymo: merged.
16:37 mihgen angdraug: this is the thing which was requested a while ago in openstack-dev, so we agreed to merge it in 6.0
16:37 mihgen mattymo: thanks for reviewing it
16:37 mattymo angdraug, it was a request from dmitryme to get this done so they can focus on their fuel-library code without worrying about it being accepted/rejected on the spec level
16:41 angdraug I thought we were in feature freeze for 6.0...
16:41 mihgen angdraug: there was a request for exception in ML and irc meeting I believe
16:42 mihgen so there were no objections, it's pretty trivial thing which doesn't affect anything besides sahara
16:42 mattymo after looking at the code, it's not so likely to affect anything else
16:42 mattymo there's just a few code quality concerns
16:42 angdraug ok
16:47 bogdando mattymo, akislitsky_  One more comment regarding  https://bugs.launchpad.net/fuel/+bug/1382527. If we used direct mounts, what would have blocked upgrade-rollback from 6.0 to 5.1(.1) due to fuel upgrade script does not create containers for storage volumes, right?
16:48 mattymo there's 1 scenario that I know might break
16:49 mattymo but it's rare
16:49 mattymo if you started with 5.0.1, then 5.1, then 6.0, then deleted 1 of the 5.1 containers, then tried to revert
16:49 mattymo 5.1 storage container parameters in dockerctl differ from what fuel_upgrade makes
16:49 mattymo but if nothing was deleted, nobody gets upset
16:52 bogdando mattymo, well, the question is - then you do upgrade from 5.1 *with* storage containers, to 6.0 which would be w/o containers anymore, will it removes them? And how will roll back recreate them back?
16:52 mattymo nope it makes new ones
16:53 mattymo each storage container is versioned to a fuel release
16:53 mattymo and none are deleted
16:53 bogdando mattymo, but in bug comments I can see that fuel_upgrade doesn't create the new storage containers (which obvious by the bug description)
16:53 bogdando then roll back, how could it create it?
16:53 mattymo the agreement made with evgeniyl__ is that fuel_upgrade will be modified to use host mounts just like my patch
16:54 mattymo it doesn't. they still exist on the host. they're just turned off
16:54 bogdando mattymo, well, ok then, looks like not not idempotent transitions 5.1 - > 6.0 -> 5.1
16:54 mattymo the only non-idempotent part is 6.0 containers still exist (but are off) when you go back to 5.1
16:55 bogdando ok, so it should work...
16:55 bogdando please put it to the bug comments so it will be better explained
16:55 evgeniyl__ bogdando: in my comment I said that it creates, the problem was that they have different names, we decided not to use this intermidiate containers at all, mattymo's patch should solve the problem, and for consistency ikalnitsky works on patch for fuel_upgrade system
16:56 mattymo ok bogdando
16:56 mattymo evgeniyl__, ikalnitsky what's your status on your side?
16:56 bogdando evgeniyl__, thanks. looks clear now
16:58 ikalnitsky mattymo: that patch is pretty easy, but it fails and I'm debugging right now. btw, it's not mandatory. your patch will be enough to fix the bug.
16:59 mattymo ikalnitsky, it works in the sense that you'll be able to destroy and recreate containers, but it's still a bug in that we have 2 different deployment architectures in 1 product for Fuel Master
16:59 mattymo and they produce different results with potentially different bugs
16:59 mattymo parity will help a lot here
17:00 mattymo fuel_upgrade still makes its own containers
17:00 ikalnitsky mattymo: i agree.. that's a problem and it's very painful to support
17:01 mattymo ikalnitsky, I discussed with evgeniyl__ and I talked about how warpc___ worked on a patch for Fuel Library to allow us to define all container options from your yaml format
17:02 mattymo that yaml format could be parsed to generate the config file format I'm using
17:02 mattymo that way we can at least use 1 set of parameters
17:04 ikalnitsky mattymo: yep, but it's not a root of cause. ideally, we should use one runner for docker container, so both dockerctl and fuel_upgrade will be able to use it.
17:04 mattymo I want to move to another deployment tool like Docker Fig
17:05 mattymo it supports a yaml for all options
17:05 mattymo this way we can get updates for free upstream and not develop our own proprietary manager
17:06 mattymo http://www.fig.sh/ and it's written in python
17:07 mattymo for actual deployment we have dockerctl still, and I'm not sure why reconfiguration of deploy params was repeated in fuel_upgrade though
17:11 mattymo https://review.openstack.org/#/c/135383/ has 5 +1s and a +2
17:11 mattymo and still isn't merged
17:11 mattymo angdraug, bogdando mihgen ?
17:18 xarses joined #fuel-dev
17:19 izinovik joined #fuel-dev
17:21 AlexF joined #fuel-dev
17:40 e0ne joined #fuel-dev
17:40 AlexF_ joined #fuel-dev
17:46 rmoe joined #fuel-dev
17:49 xarses bookwar: mihgen teran_ not getting any downloads on http://seed.fuel-infra.org/fuelweb-iso/fuel-community-5.1.1-20-2014-11-25_01-14-45.iso.torrent
17:51 bookwar xarses: does your torrent client support webseed feature?
17:51 teran_ xarses: what the client are using?
17:51 xarses rtorrent
17:51 xarses been working fine
17:51 xarses downloaded a internal iso yesterday
17:51 teran_ xarses: afaik it doesn't support webseeds
17:52 SergK xarses: downloading now....
17:52 teran_ xarses: internaly you might catch some luck and download from other peers instead of webseeds
17:52 xarses community has a different seed format than community?
17:52 teran_ xarses: nope, there are simply less peers
17:53 teran_ xarses: I see the only peer for your ISO
17:57 xarses left #fuel-dev
17:57 xarses joined #fuel-dev
18:04 fandi joined #fuel-dev
18:04 xarses also, what's up with retracker.local:80?
18:05 salmon_ joined #fuel-dev
18:06 e0ne joined #fuel-dev
18:10 pboldin joined #fuel-dev
18:12 teran_ xarses: looks like some DNS stuff in internal network, will check, thanks
18:16 e0ne joined #fuel-dev
18:30 bookwar WARNING: fuel-web CI tests are on hold, we are cleaning nailgun leftovers from previous tests in postgres
18:32 monester_laptop joined #fuel-dev
18:43 iFiz_ joined #fuel-dev
18:46 iFiz_ hey.. anyone who could give me a hint how long it 'should' take to create an empty cinder volume on ceph (rbd)? size is 10G
18:51 xarses joined #fuel-dev
18:59 prmtl joined #fuel-dev
19:03 teran joined #fuel-dev
19:04 teran joined #fuel-dev
19:05 proberts joined #fuel-dev
19:07 xarses joined #fuel-dev
19:27 dshulyak_ joined #fuel-dev
19:29 ignatenko joined #fuel-dev
19:29 agordeev joined #fuel-dev
20:01 HeOS joined #fuel-dev
20:02 xarses joined #fuel-dev
20:07 mihgen xarses: hi, can you take a look on whether the status of the bug is correct: https://bugs.launchpad.net/fuel/+bug/1272766
20:11 xarses mihgen: updated, its fix commited
20:11 xarses committed even
20:14 mihgen xarses: cool, thx!
20:45 e0ne joined #fuel-dev
20:48 prmtl joined #fuel-dev
21:16 proberts left #fuel-dev
21:31 monester_laptop joined #fuel-dev
21:36 iFiz joined #fuel-dev
21:48 salmon_ joined #fuel-dev
22:09 adanin joined #fuel-dev
22:17 angdraug joined #fuel-dev
22:18 proberts joined #fuel-dev
22:37 prmtl joined #fuel-dev
22:41 DasHereNow joined #fuel-dev
22:42 e0ne joined #fuel-dev
23:01 enikanorov_ joined #fuel-dev
23:01 adanin joined #fuel-dev
23:01 teran joined #fuel-dev
23:03 albs joined #fuel-dev
23:27 teran_ joined #fuel-dev
23:35 DasHereNow joined #fuel-dev

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