Thursday, 2018-07-19

*** eernst has quit IRC00:02
*** dklyle has quit IRC00:02
*** fuentess has quit IRC00:47
*** annabelleB has quit IRC00:51
*** annabelleB has joined #kata-dev00:53
*** annabelleB has quit IRC00:53
*** sameo_ has joined #kata-dev01:38
*** sameo has quit IRC01:38
*** lamego has quit IRC01:38
*** lamego has joined #kata-dev01:42
*** dlw has joined #kata-dev02:05
*** lsm5 has quit IRC02:07
*** lsm5 has joined #kata-dev02:09
*** dlw1 has joined #kata-dev02:13
*** dlw has quit IRC02:13
*** dlw1 is now known as dlw02:13
*** lsm5 has quit IRC02:16
*** lsm5 has joined #kata-dev02:16
*** zerocoolback has joined #kata-dev02:31
*** dlw1 has joined #kata-dev03:30
*** dlw has quit IRC03:32
*** dlw1 is now known as dlw03:32
*** lpetrut has joined #kata-dev03:52
*** sjas_ has joined #kata-dev04:09
*** sjas has quit IRC04:11
*** zerocoolback has quit IRC04:23
*** zerocoolback has joined #kata-dev04:31
*** lpetrut has quit IRC04:36
*** zerocoolback has quit IRC05:01
*** zerocoolback has joined #kata-dev05:03
*** zerocoolback has quit IRC05:21
*** zerocool_ has joined #kata-dev05:21
*** fdegir has quit IRC05:29
*** din has quit IRC05:29
*** fdegir has joined #kata-dev05:29
*** din has joined #kata-dev05:30
*** din has quit IRC05:32
*** din has joined #kata-dev05:32
*** din has quit IRC05:32
*** din has joined #kata-dev05:32
*** lpetrut has joined #kata-dev06:05
*** lpetrut has quit IRC06:16
*** lpetrut has joined #kata-dev06:16
*** sameo_ has quit IRC06:46
*** zerocoolback has joined #kata-dev07:29
*** zerocool_ has quit IRC07:31
*** lpetrut has quit IRC07:40
*** sameo has joined #kata-dev07:49
*** davidgiluk has joined #kata-dev08:03
*** gwhaley has joined #kata-dev08:15
*** sjas_ is now known as sjas08:20
kata-irc-bot1<kaly.xin> @salvador.fuentes @graham.whaley Hi, We have a ARM server at hand, we now plan to make this server as a Jenkins slave machine first and we will add test cases bit by bit after that. We know the basic mechanism about the Jenkins setup, but it is still case by case, depending on the configuration of the jenkins master. So could you help us to know what's the first step? for example, the information which the master needs(IP, keys, etc),08:55
kata-irc-bot1or the plugin needs to be installed on the slave? or the credential things? We want the current master to know the slave and can configure and schedule basic tasks on the slave. Thanks a lot!08:55
*** lpetrut has joined #kata-dev08:57
*** lpetrut has quit IRC09:18
*** lpetrut has joined #kata-dev09:18
gwhaleyHi @kaly.xin. So you'd like the current Jenkins master we have to schedule on your slave, yes? OK, I guess that is possible as long as our Azure based master can see your slave machine somehow.09:41
gwhaleycurrently I believe we use an azure node pool for the slave allocation - for your physical server I think we are going to have to change/modify our master setup to add that builder as a single build resource and then09:42
gwhaleywe will have to give it a label I suspect and then create a new build job that targets slaves with that label.09:42
gwhaleywe don't currently do any 'matrix' style builds, so I don't think we can tell each build job to build both x86 and arm off a single job (Jenkins can do that, but I don't think we are set up that way at present).09:43
gwhaley@salvador.fuentes knows more about how our Azure QA CI is set up though. If he does not get back to you later today (timezones...), then a comment on or a new Issue on github is probably going to be the best way09:44
gwhaleyto work out the details.09:44
*** changcheng has quit IRC10:03
*** changcheng has joined #kata-dev10:08
davidgiluksomeone the other day had a guest crash in a nested setup; I've just got an oops in a nest - I'd like to see if it's the same one - anyone remember where the backtrace from that last case was?10:39
gwhaley@davidgiluk - the only one I remember is that we have trouble running certain distros nested (iirc, maybe it was Ubuntu18 or F28)... we have seen this before, and we were seeing it on the vexxhost CI systems - let me10:43
davidgilukgwhaley: This was F28 on a RHEL host10:43
davidgilukit's in native_write_cr410:43
gwhaleysee if I can find a link - iirc I posted a similar link that was on GCE or similar10:47
gwhaleyyeah, we think it is some odd nested VM kernel incompatability between host and guest...10:47
davidgilukit tends to happen when the L2 exits I think10:47
gwhaleydavidgiluk - I think this is the relevant thread: https://github.com/kata-containers/tests/issues/41810:49
* gwhaley searched for native_write_cr4 in the kata-containers github top level project ;-)10:49
davidgilukit's in the kernel10:50
davidgilukwrite10:50
davidgilukright, so my backtrace looks just the same, I'll have a chat to some KVM people10:50
gwhaleythat'd be great - it'll be good to nail this one for the upstream. There is a link in that thread to a kernel bugzilla where it may have been seen on GCE btw...10:51
gwhaleyI see you are in sunny Manchester btw davidgiluk - heh, I'm sat over here in Leeds ;-)10:52
davidgilukgwhaley: Ah nice10:54
davidgilukI see Daivd Hildenbrand replied to that kernel bz; he's on our team10:55
*** gwhaley has quit IRC10:59
*** dlw has quit IRC11:05
*** zerocoolback has quit IRC11:52
*** zerocool_ has joined #kata-dev11:52
*** devimc has joined #kata-dev11:58
*** zerocoolback has joined #kata-dev12:00
*** zerocool_ has quit IRC12:04
* davidgiluk thinks disabling pcid is fixing it12:16
*** gwhaley has joined #kata-dev12:43
*** fuentess has joined #kata-dev13:28
*** eernst has joined #kata-dev14:06
*** eernst has quit IRC14:38
kata-irc-bot1<salvador.fuentes> Hi @kaly.xin, @graham.whaley, sure, I can help you with that. I'll figure out what information we need and I'll get back with you :slightly_smiling_face:14:57
gwhaleythanks @salvador.fuentes. I suspect one key bit of info is if we have some sort of secure connection between the machines - I'm presuming we'll need maybe an SSH setup to launch the jenkins agent for instance.15:02
gwhaleyIf the build machine happens to be deeply buried in some proxy-land, then there is a way to config Jenkins to do the reverse I believe - to have the slave call back/out to the server15:02
gwhaleyso, some of this will come down to the build machine setup - the Azure machines are publicly visible.15:03
*** sameo has quit IRC15:04
*** dklyle has joined #kata-dev15:23
*** annabelleB has joined #kata-dev15:27
kata-irc-bot1<salvador.fuentes> @mnaser, clarkb: Hi guys, I was wondering if I could get access to a vexxhost machine, we are having issues with cri-o and would like to reproduce them.15:56
clarkbsalvador.fuentes I can help with that. You need this on ubuntu xenial?15:56
kata-irc-bot1<salvador.fuentes> clarkb: yes please15:56
kata-irc-bot1<yonatan.gefen> Are there any (negative) implications for choosing "bridged" over "macvtap" for internetworking_model in configuration.toml?15:57
clarkbsalvador.fuentes I've asked zuul to hold the node for the next xenial job that fails15:59
clarkbspecifically on the proxy repo (I've rechecked my PR there)15:59
kata-irc-bot1<salvador.fuentes> clarkb: great, thank you :slightly_smiling_face:15:59
*** annabelleB has quit IRC16:04
*** annabelleB has joined #kata-dev16:16
clarkbsalvador.fuentes: root@162.253.55.14616:26
clarkblet me know when you are done with it and I will get it cleaned up16:26
kata-irc-bot1<salvador.fuentes> clarkb: thanks16:26
*** JerryLiu has joined #kata-dev16:27
*** changcheng has quit IRC16:27
*** eernst has joined #kata-dev17:19
*** zerocoolback has quit IRC17:19
*** eernst has quit IRC17:32
*** eernst has joined #kata-dev17:33
*** sgordon has joined #kata-dev17:37
*** JerryLiu has quit IRC17:54
*** JerryLiu has joined #kata-dev18:00
*** gwhaley has quit IRC18:09
*** davidgiluk has quit IRC19:19
*** lpetrut has quit IRC19:39
*** annabelleB has quit IRC19:42
*** fiddletwix has quit IRC19:47
*** fiddletwix has joined #kata-dev19:47
*** annabelleB has joined #kata-dev19:56
*** devimc has quit IRC20:10
*** dklyle has quit IRC22:18
*** lamego has left #kata-dev22:27
*** eernst has quit IRC22:32
*** eernst has joined #kata-dev22:34
*** annabelleB has quit IRC22:51
*** dklyle has joined #kata-dev23:00
kata-irc-bot1<salvador.fuentes> clarkb: hi, I will still be using the machine tomorrow morning. I was able to fix cri-o, but now cri-containerd and a k8s test seem to fail23:05
clarkbno problem23:06
kata-irc-bot1<salvador.fuentes> @mnaser, clarkb: do you know if something changed on vexxhost cloud? I ask because I am not seeing this errors when running locally or on an azure machine23:06
clarkbI am not aware of any changes, we do update the images daily though, but I woudl expect if you udpated your packges and rebooted locally you'd see the same23:07
kata-irc-bot1<salvador.fuentes> clarkb: ok, thanks, yeah I think I have updated packages, weird...23:09
*** eernst has quit IRC23:15
*** dklyle has quit IRC23:17
*** dklyle has joined #kata-dev23:50
*** fiddletwix has quit IRC23:58
*** fiddletwix has joined #kata-dev23:59

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!