Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel-dev, 2014-09-11

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

All times shown according to UTC.

Time Nick Message
00:30 rmoe joined #fuel-dev
00:51 tzn joined #fuel-dev
01:08 apodrepniy joined #fuel-dev
01:16 xarses joined #fuel-dev
01:52 tzn joined #fuel-dev
02:03 teran joined #fuel-dev
02:53 tzn joined #fuel-dev
02:54 eshumkher joined #fuel-dev
03:04 teran joined #fuel-dev
03:22 kevinbenton joined #fuel-dev
03:23 kevinbenton hello, can i get some more feedback on https://review.openstack.org/#/c/118267/
03:23 kevinbenton it seems to fix the migration failure issues we ran into with 3-server cluster setups
03:29 Longgeek joined #fuel-dev
03:32 Longgeek_ joined #fuel-dev
03:44 [HeOS] joined #fuel-dev
03:45 meow-nofer_ joined #fuel-dev
03:53 tzn joined #fuel-dev
04:19 prmtl joined #fuel-dev
04:21 teran joined #fuel-dev
04:54 prmtl joined #fuel-dev
04:54 jaypipes joined #fuel-dev
04:55 tzn joined #fuel-dev
05:12 prmtl joined #fuel-dev
05:22 dshulyak_ joined #fuel-dev
05:36 teran joined #fuel-dev
05:52 tzn joined #fuel-dev
06:18 Longgeek joined #fuel-dev
06:19 prmtl joined #fuel-dev
06:24 salmon_ joined #fuel-dev
06:30 sambork joined #fuel-dev
06:33 tldr joined #fuel-dev
06:40 apodrepniy joined #fuel-dev
06:40 albs joined #fuel-dev
06:41 tuvenen joined #fuel-dev
06:52 Longgeek joined #fuel-dev
06:59 teran joined #fuel-dev
07:01 pasquier-s joined #fuel-dev
07:02 tlbr_ joined #fuel-dev
07:06 akislitsky joined #fuel-dev
07:16 artem_panchenko guys, latest BVT on Ubuntu (5.1) failed due to failed cluster deployment, but there are no errors in astute's and puppet logs on nodes. Also cluster passes OSTF. Can anybody look at that environment? I will provide an access to it.
07:21 artem_panchenko aglarendil, warpc__ ^^
07:22 warpc__ artem_panchenko: hi! i can look
07:24 artem_panchenko warpc__:  I sent you access info in private message
07:32 e0ne joined #fuel-dev
07:39 tzn joined #fuel-dev
07:58 tzn joined #fuel-dev
08:05 [HeOS] joined #fuel-dev
08:08 EeeBotu New bug #1368068 in fuel: "Can't install OS on OpenCompute HW, SOL isn't working when FB activated" [Undecided, New] https://bugs.launchpad.net/bugs/1368068
08:38 Ch00k joined #fuel-dev
08:58 tldr joined #fuel-dev
09:01 e0ne joined #fuel-dev
09:06 akasatkin mihgen: https://bugs.launchpad.net/fuel/+bug/1360230 - me and Denis D. reproduced this. About 100% reproducible on some conditions in VirtualBox only. I left a comment in ticket.
09:08 EeeBotu New bug #1368103 in fuel: "[upgarde] Environment update failed: 'Cs_resource[openstack-heat-engine]/ensure (err): change from present to absent failed: execution expired'" [Undecided, New] https://bugs.launchpad.net/bugs/1368103
09:08 EeeBotu New bug #1368100 in fuel: "Disable 5.0.2 for new envs after upgrade master node to 5.1" [Critical, New] https://bugs.launchpad.net/bugs/1368100
09:11 tzn joined #fuel-dev
09:15 asyriy joined #fuel-dev
09:20 apodrepniy joined #fuel-dev
09:22 a_teem_ joined #fuel-dev
09:27 ikalnitsky joined #fuel-dev
09:44 adanin joined #fuel-dev
09:49 teran joined #fuel-dev
09:56 albs joined #fuel-dev
09:56 asyriy joined #fuel-dev
09:56 apodrepniy joined #fuel-dev
10:10 teran joined #fuel-dev
10:12 teran_ joined #fuel-dev
10:15 evgeniyl_ tuvenen: Hi, are you here? I've read your email about nailgun part of "ssl in fuel" feature, and you mentioned there 'docker CA', I'm not a deployment engineer, so, could you please clarify what is 'docker CA'?
10:19 tuvenen evgeniyl_, Hi. Yes sorry it is not clear in the mail. Docker CA is a docker container that is able to respond to a CSR (certificate signing request). So this container will act as the authority.
10:20 tuvenen So if the fuel master needs to signed some certificates it can ask to this authority instead of self-signed them.
10:20 tuvenen It is not available right now
10:22 ikalnitsky Hi guys! I'm working on this issue https://bugs.launchpad.net/fuel/+bug/1367001 and  I have one question - should we add restrictions for HA only, or it should also affect simple mode? Could someone help me? aglarendil, mattymo ?
10:23 aglarendil haha
10:23 aglarendil ikalnitsky: you MUST not allow user to remove the ONLY controller
10:28 msemenov joined #fuel-dev
10:32 teran joined #fuel-dev
10:38 EeeBotu New bug #1368141 in fuel: "[Nailgun] Cannot assign role to node" [Undecided, New] https://bugs.launchpad.net/bugs/1368141
10:38 EeeBotu New bug #1368137 in fuel: "check for missing packages when building deb mirror" [High, Confirmed] https://bugs.launchpad.net/bugs/1368137
10:42 evgeniyl_ tuvenen: so, is it a some service in a docker container, or some command line interface for certificates signing?
10:45 tuvenen evgeniyl_, it is not decided yet if it will be a service or something else. I just mentionned it as a future work and I think that it can goes in the ca-management blueprint.
10:45 tuvenen evgeniyl_, Right now for SSL OSt endpoints and fuel ssl we will use the self signed certificate only. We will keep it simple for now.
10:47 tuvenen evgeniyl_, and it is not sure that we will use a docker container as authority (even if the idea sounds good ;)
10:47 evgeniyl_ tuvenen: ok, got it, thanks.
10:47 tuvenen evgeniyl_, you're welcome.
10:52 dshulyak_ joined #fuel-dev
11:01 apalkina new bug for zabbix https://bugs.launchpad.net/fuel/+bug/1368151
11:01 mihgen aglarendil: did you see this? https://bugs.launchpad.net/fuel/+bug/1360230 it was reproduced by akasatkin
11:01 mihgen akasatkin: let's get his env and nail it down
11:03 teran joined #fuel-dev
11:04 teran_ joined #fuel-dev
11:10 e0ne joined #fuel-dev
11:11 EeeBotu New bug #1368151 in fuel: "ERR: Could not start Service[mysql] for zabbix node" [Critical, New] https://bugs.launchpad.net/bugs/1368151
11:11 EeeBotu New bug #1368150 in fuel: "[juno] missed murano packages" [Undecided, New] https://bugs.launchpad.net/bugs/1368150
11:19 f13o joined #fuel-dev
11:35 IvanBerezovskiy joined #fuel-dev
11:50 apodrepniy_ joined #fuel-dev
11:50 akasatkin mihgen: aglarendil: Denis Dmitriev now deals with this bug. He controls envs.
11:52 teran joined #fuel-dev
11:54 enikanorov_ joined #fuel-dev
11:54 msemenov_ joined #fuel-dev
11:58 mihgen akasatkin: ddmitriev Very good. This bug scares me, as I still remember another bug which was only reproducable on your env, and finally appeared to be a huge complicated issue. Let's troubleshoot and find root cause
12:00 pasquier-s joined #fuel-dev
12:08 aglarendil akasatkin: ddmitriev guys, the problem is that we see it only with slow environments. it is just compute node timing out upload of a file. that's all
12:09 aglarendil akasatkin: ddmitriev and actually the problem is that instead of just copying data inside ceph we are downloading and then uploading files again
12:10 aglarendil akasatkin: ddmitriev there is an issue in nova about it, which angdraug promised to add into bug comments
12:13 taj joined #fuel-dev
12:14 ddmitriev aglarendil: it is no slow environment. a snapshot task that started yesterday has not finished yet (more than 24 hours).   There is no CPU load, IO load, memory lack or any other lack of resources.
12:14 aglarendil xenolog: ^^
12:15 ddmitriev This issue is reproduced only on VBOX and only with ephemeral storage enabled.  Now testing with slightly different config
12:16 xenolog Den, could you provide me acces to the you environment?
12:17 ikalnitsky aglarendil / alex_didenko : please review and add +1 if you're ok https://review.openstack.org/#/c/120741/2
12:17 prmtl joined #fuel-dev
12:17 xenolog ddmitriev, could you provide me acces to the you environment?
12:38 EeeBotu New bug #1368206 in fuel: "Wrong statement about CPU utilization if Neutron with GRE segmentation is used." [Low, Confirmed] https://bugs.launchpad.net/bugs/1368206
13:37 sambork please reviews: https://review.openstack.org/#/c/120444/
13:38 mgariepy joined #fuel-dev
13:42 e0ne joined #fuel-dev
13:42 asyriy joined #fuel-dev
13:42 albs joined #fuel-dev
13:43 jobewan joined #fuel-dev
13:46 prmtl_ joined #fuel-dev
13:53 apodrepniy__ joined #fuel-dev
13:59 youellet joined #fuel-dev
14:00 pasquier-s_ joined #fuel-dev
14:04 eshumkher joined #fuel-dev
14:05 Longgeek joined #fuel-dev
14:37 mihgen nurla: aglarendil: vkramskikh: do we have a task anywhere to make patching in experimental feature group?
14:42 Ch00k_ joined #fuel-dev
14:43 EeeBotu New bug #1368263 in fuel: "[juno] node boots in grub rescue state" [Medium, Confirmed] https://bugs.launchpad.net/bugs/1368263
14:55 mihgen nurla: can you please confirm https://bugs.launchpad.net/fuel/+bug/1368275?
14:55 mihgen and we need someone from UI team or Python team to make it happen
14:55 mihgen I don't know which team should take it
14:56 nurla mihgen: done
14:59 angdraug joined #fuel-dev
15:07 teran joined #fuel-dev
15:12 tuvenen_ joined #fuel-dev
15:13 pasquier-s__ joined #fuel-dev
15:20 pasquier-s_ joined #fuel-dev
15:24 eshumkher joined #fuel-dev
15:26 EeeBotu New bug #1368285 in fuel: "Necessity to add a new nsx's automated tests" [Undecided, New] https://bugs.launchpad.net/bugs/1368285
15:26 EeeBotu New bug #1368275 in fuel: "Patching of OpenStack should be experimental feature" [Critical, Confirmed] https://bugs.launchpad.net/bugs/1368275
15:26 asyriy joined #fuel-dev
15:33 ikalnitsky joined #fuel-dev
15:33 xarses joined #fuel-dev
15:34 apodrepniy joined #fuel-dev
15:42 ikalnitsky Could someone from library take a look https://review.openstack.org/#/c/120741/ , please?
15:42 ikalnitsky aglarendil: ^
15:43 IvanBerezovskiy left #fuel-dev
15:43 eshumkher joined #fuel-dev
15:44 mihgen tzn: https://bugs.launchpad.net/fuel/+bug/1367234 merged only to master, can you backport to stable/5.1 pls
15:47 eshumkher joined #fuel-dev
15:48 tzn Hmm, I thought holser did it, but i might have misunderstand him
15:48 tzn Will do
15:54 tzn Done, please merge
15:55 sambork joined #fuel-dev
15:55 sambork1 joined #fuel-dev
15:56 tzn Also to 5.0
15:57 Ch00k joined #fuel-dev
15:59 mihgen #openstack-meeting-alt time
16:07 eshumkher joined #fuel-dev
16:14 prmtl joined #fuel-dev
16:28 eshumkher joined #fuel-dev
16:31 eshumkher joined #fuel-dev
16:32 tdubyk joined #fuel-dev
16:46 dshulyak_ joined #fuel-dev
17:13 prmtl joined #fuel-dev
17:18 xarses joined #fuel-dev
17:21 rmoe joined #fuel-dev
17:38 dshulyak_ joined #fuel-dev
17:42 eshumkher joined #fuel-dev
17:56 tzn joined #fuel-dev
18:12 e0ne joined #fuel-dev
18:22 xarses joined #fuel-dev
18:26 teran joined #fuel-dev
19:02 eshumkher joined #fuel-dev
19:09 sbog joined #fuel-dev
19:09 teran joined #fuel-dev
19:09 aglarendil joined #fuel-dev
19:18 sbog joined #fuel-dev
19:18 aglarendil joined #fuel-dev
19:19 angdraug joined #fuel-dev
19:23 eshumkher joined #fuel-dev
19:27 sbog joined #fuel-dev
19:27 aglarendil joined #fuel-dev
19:32 eshumkher joined #fuel-dev
19:36 pasquier-s joined #fuel-dev
19:36 sbog joined #fuel-dev
19:36 teran joined #fuel-dev
19:37 aglarendil joined #fuel-dev
19:37 youellet2 joined #fuel-dev
19:39 ikalnitsky_ joined #fuel-dev
19:40 jkirnosova joined #fuel-dev
19:40 jobewan joined #fuel-dev
19:40 brain461 joined #fuel-dev
19:43 pasquier-s joined #fuel-dev
19:44 nurla joined #fuel-dev
19:45 MorAle joined #fuel-dev
19:45 sbog joined #fuel-dev
19:46 aglarendil joined #fuel-dev
20:00 teran joined #fuel-dev
20:22 HeOS joined #fuel-dev
20:38 adanin joined #fuel-dev
20:39 angdraug joined #fuel-dev
20:45 e0ne joined #fuel-dev
21:00 enikanorov_ joined #fuel-dev
21:25 mihgen xarses: hi
21:26 mihgen xarses: coming across your presentation.. which says that we use multicast for crm communication
21:26 mihgen is it true?
21:28 xarses I will double check, but last I did, yes it does
21:33 mihgen xarses: please double check… the thing is that multicast is often prohibited on switches by net admins
21:33 mihgen and we wanted to have multicast check actually as part of net-checker
21:34 mihgen it is even done by dshulyak, but as I was informed that we still don't run multicast, I've instructed to disable the check
21:34 xarses we disabled the check because the check didn't work right
21:34 mihgen https://blueprints.launchpad.net/fuel/+spec/multicast-group-verification this one
21:35 mihgen really? I thought dshulyak managed it to work finally
21:36 mihgen actually if it's multicast, it's not the only bad thing that we don't have checks; but also that we would use hardcoded network
21:36 mihgen which may easily intersect with user's other net
21:41 xarses mike it was changed to unicast only in https://github.com/stackforge/fuel-library/commit/faa0ecd9beacc9327076b8a29ba99df3ad27f6b3
21:43 xarses mihgen: ^
21:43 mihgen xarses: ok, good to know it's unicast
21:44 xarses I fixed the slide
21:44 mihgen then I have question how nodes are being added to the HA cluster
21:44 mihgen for example, we have 2 nodes initially deployed, then add 3rd one
21:44 mihgen how the first 2 update their config to know about 3rd?
21:45 salmon_ joined #fuel-dev
21:47 xarses we should avoid 2 -> 3
21:47 xarses 1 -> 3 should be fine
21:47 xarses controller rolls are re-deployed when any controller is provisioned
21:47 mihgen xarses: oh ok, then it's fine
21:48 xarses the only gap is that we don't update rabbit_hosts on computes
21:48 xarses everything else uses the vip
21:48 mihgen 2->3 can't be avoided if we lose one node, and then replace it
21:48 xarses correct, but I'm told that works fine
21:48 xarses im currently running 1 -> 3
21:49 mihgen xarses: > that we don't update rabbit_hosts on computes - this sounds like a real issue to me, unless we give same IPs for replaced controllers
21:50 xarses I'm more worried about only having one node in the list
21:51 xarses It's something we need to address in 6.0
21:52 mihgen xarses: what?? one node? not even vip?
21:52 mihgen I can't believe - we would see failing tests if so
21:53 bookwar joined #fuel-dev
21:54 xarses if we did destructive tests and scaling in the same test, sure
21:55 xarses we don't use the vip for services that support rabbit_hosts
21:55 xarses because the vip made HA worse
21:55 mihgen wait. if you see one server in such rabbit_hosts, then obviously when server with such IP is down, services won't work for sure
21:55 mihgen you don't need to play with scaling, you need just to shut down one controller with that IP
21:56 xarses rabbit_hosts will be a list of contollers at deployment time of that role
21:56 xarses so if I deploy one controller in HA mode, there will only be one item in it
21:57 xarses if i deploy 3 controllers, it will have 3 items in it
21:57 xarses but if i deploy computes when there is only one controller
21:57 xarses and then add more controllers later
21:57 xarses and the compute role (or rabbit_hosts its self) isn't updated later, then it will still have the first controller in the list
21:58 mihgen angdraug: FYI looks like we all gonna be on Monday, so you can start with scheduling of sessions since monday
21:59 mihgen angdraug: let's start with something easy / introductory first )
21:59 angdraug excellent! I will
22:00 mihgen xarses: nice. I see. Will you report it?
22:00 mihgen I think it should go to 6.0
22:00 angdraug we'll start with a long boring keynote to help everyone cope with their jet lag :)
22:00 xarses mihgen: sure
22:00 mihgen I would be pretty scared to make redeploy of computes in 5.1
22:00 xarses angdraug: lol
22:00 mihgen angdraug: )
22:08 apodrepniy_ joined #fuel-dev
22:08 asyriy joined #fuel-dev
22:18 xarses mihgen: https://bugs.launchpad.net/fuel/+bug/1368445
22:20 mihgen xarses: could you please add a bit more info there why you think it's on fuel-python? because serializer should give info to redeploy computes as well when controller is added/replaced?
22:20 xarses its either fuel-python or fuel-astute
22:20 xarses possibly both
22:21 xarses we only have to update rabbit_hosts, we dont have to run the full compute role if we don't want to
22:25 mihgen xarses: but you'd need to restart nova-compute
22:26 xarses ya, but we could run a manifest like we do to update the /etc/hosts file
22:27 mihgen xarses: that's what aglarendil is gonna say we need granular deployments for, and I'll agree :)
22:27 xarses we also need to restart / reload neutron-ovs-agent
22:27 mihgen xarses: what's that for?
22:28 mihgen is it also uses same rabbitmq_hosts ?
22:28 xarses its the neutron agent on compute node
22:28 xarses yes
22:28 apodrepniy_ joined #fuel-dev
22:28 xarses neutron agent-ist
22:28 xarses neutron agent-list
22:29 mihgen nice. we should add it to bug report too..
22:29 mihgen i've just updated description with reference to IRC logs
22:31 xarses oh, i just did the same
22:38 kat_pimenova joined #fuel-dev
22:38 xarses ok, put your note back in
22:39 jkirnosova joined #fuel-dev
22:39 brain461 joined #fuel-dev
22:48 EeeBotu New bug #1368445 in fuel: "need to update rabbit_hosts on non-controller nodes after deploying controllers" [High, Confirmed] https://bugs.launchpad.net/bugs/1368445
22:57 * xarses cheers
22:57 xarses mihgen: hcf now, so goto bed
22:57 xarses =)
22:58 mihgen xarses: looks like adrenalin is in blood because of it) can't make myself
22:59 mihgen perhaps need some highland whiskey from the frige .. )
23:05 mihgen xarses: angdraug Do we have anything in our docs about split brain?
23:05 mihgen we don't handle it still, so we need to documented how it may happen and what to do in those cases
23:17 xarses mihgen: not sure, will need angdraug to reply
23:40 angdraug no, we don't have anything about split brain recovery in the docs
23:40 angdraug http://docs.mirantis.com/openstack/fuel/master/operations.html#how-to-troubleshoot-corosync-pacemaker
23:41 angdraug that's all we got, and it only tells you how to detect split brian, not how to recover from it

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