Thursday, 2018-11-29

*** hongbin has quit IRC00:12
*** openstackgerrit has joined #openstack-containers00:51
openstackgerritLingxian Kong proposed openstack/magnum master: Add load balancer hook for cluster pre-deletion  https://review.openstack.org/62076100:51
*** jakeyip has quit IRC01:31
*** jakeyip has joined #openstack-containers01:35
*** ricolin has joined #openstack-containers02:21
*** hongbin has joined #openstack-containers02:28
*** ramishra has joined #openstack-containers02:53
*** rcernin has quit IRC02:58
*** ramishra has quit IRC03:04
*** ramishra has joined #openstack-containers03:06
*** rcernin has joined #openstack-containers03:38
*** chhagarw has joined #openstack-containers04:15
*** hongbin has quit IRC04:34
*** dave-mccowan has quit IRC04:34
*** dave-mccowan has joined #openstack-containers04:35
*** dave-mccowan has quit IRC04:49
*** udesale has joined #openstack-containers05:06
*** janki has joined #openstack-containers05:09
*** ykarel|away has joined #openstack-containers05:20
*** jackivanov has joined #openstack-containers05:49
*** ykarel|away has quit IRC06:02
*** ykarel|away has joined #openstack-containers06:16
*** ykarel|away is now known as ykarel|lunch06:20
*** Namrata has joined #openstack-containers06:27
*** mkrai has joined #openstack-containers06:28
NamrataHi folks, I am trying to create a kubernetes cluster which gets FAILED and I can see kube_masters in CREATE_FAILED state and I cant reach to the master node vm .06:29
Namratahow can I debug the actual failure06:29
Namratastrigazi:06:35
Namratastrigazi: can you help me with this06:35
openstackgerritJake Yip proposed openstack/magnum-tempest-plugin master: Add option for kubernetes coe  https://review.openstack.org/60188907:15
*** shrasool has quit IRC07:18
jakeyipNamrata, what does `openstack coe cluster show` says? Also, is the VM created OK?07:23
jakeyipthe VM name is <cluster_name>-xxxx-master-0 , or you can trace via the heat stack07:23
*** rcernin has quit IRC07:35
Namratajakeyip: `openstack coe cluster show` gives status `CREATE_FAILED` reason timeout and I can see the creation aborted at enable_promethus_monitoring_deployment07:42
Namratajakeyip: openstack server list shows vm but I cant ssh into it and when I ping it says host unreachable07:43
*** ykarel|lunch is now known as ykarel08:31
*** dims has quit IRC08:32
*** dims has joined #openstack-containers08:33
*** ttsiouts has joined #openstack-containers08:34
*** ttsiouts has quit IRC08:36
*** ttsiouts has joined #openstack-containers08:37
*** ttsiouts has quit IRC08:41
*** pcaruana has joined #openstack-containers08:48
*** ricolin has quit IRC09:07
jakeyipNamrata: VM is state ACTIVE?09:08
*** ttsiouts has joined #openstack-containers09:09
Namratajakeyip yes it is in ACTIVE state09:09
jakeyipNamrata: ok. what's your template like? on mine, with floating_ip_enabled = true and external_network_id = <floating>, I am able to ping.09:14
kevkoNamrata: i know what is you issue09:15
Namratakevko I am not sure but is it realted to https://bugs.launchpad.net/magnum/+bug/174436209:15
openstackLaunchpad bug 1744362 in Magnum "heat-container-agent fails to communicate with Keystone." [Medium,Confirmed]09:15
kevkoNamrata: you have to set label cert_manager_api=true when creating k8s cluster09:16
kevkoi have another issue with magnum ...some scripts are not injected to instance09:17
kevkoNamrata: your problem i was trying to solve for 4 days :D so i know what i am talking about :D09:17
kevkoNamrata: try it09:18
kevkomaybe it should be set by default ..but i am not sure ..09:18
Namratakevko this label should be added in template or while cluter creation09:25
Namrata*cluster09:25
kevkoNamrata: i was adding while creating cluster ...09:25
Namrataokay09:25
kevkoNamrata: but, i think you can add also to template ... (didn't try )09:26
kevkoNamrata: can you please help me with my problem if you will have a time ?09:26
Namratakevko I am newbie to this I am not sure I can help you09:33
Namrataand that label dosent work for me09:33
kevkoNamrata: really ?09:36
kevkoNamrata: hmm, in which step are you stuck now ?09:37
*** kaiokmo has quit IRC09:38
*** DimGR has joined #openstack-containers09:40
openstackgerritLingxian Kong proposed openstack/magnum master: Add load balancer hook for cluster pre-deletion  https://review.openstack.org/62076109:41
*** shrasool has joined #openstack-containers09:44
*** jchhatbar has joined #openstack-containers10:02
*** janki has quit IRC10:05
*** ttsiouts has quit IRC10:05
*** ttsiouts has joined #openstack-containers10:05
*** ttsiouts has quit IRC10:08
*** ttsiouts has joined #openstack-containers10:08
kevkoguys, can you someone help me with magnum ? i have issues with softwaredeployment ...some scripts are not injected into instance ..10:18
*** shrasool has quit IRC10:24
*** udesale has quit IRC10:27
*** sayalilunkad has quit IRC10:33
*** jchhatba_ has joined #openstack-containers10:37
*** lpetrut has joined #openstack-containers10:38
*** jchhatbar has quit IRC10:39
jakeyipwhat's the problem kevko?10:40
*** sayalilunkad has joined #openstack-containers11:12
*** ttsiouts has quit IRC11:24
*** ttsiouts has joined #openstack-containers11:24
*** jchhatbar has joined #openstack-containers11:30
*** ttsiouts has quit IRC11:32
*** jchhatba_ has quit IRC11:33
*** salmankhan has joined #openstack-containers11:35
kevkojakeyip: some scripts are not executed11:38
kevkojakeyip: in log i have spam of this ->   waiting for CA to be made available for certificate manager api11:38
kevkojakeyip: because script which is writing echo -e "$CA_KEY" > /etc/kubernetes/certs/ca.key  i don't know why is not executed ...any advice  ?11:39
*** mkrai has quit IRC11:40
*** jchhatbar has quit IRC11:43
jakeyiphm what's your cert_manager_type ?11:49
kevkojakeyip: barbican11:50
kevkojakeyip: i have all filled in /etc/sysconfig/heat-params , also CA_KEY ... and when i log into instance and manually echo CA_KEY to /etc/kubernetes/certs/ca.key ...everything after that is OK and cluster deployed ...11:51
jakeyipok, I had similar problem before but using x509. check cloud-init-output and cloud-init and see if you can spot the part-xxx script tha fails11:52
kevkojakeyip: it is not problem with cert manager ..problem is that script in /var/lib/heat-config/heat-config-script/ is not executed11:52
kevkojakeyip: no, all parts script ran without issue11:52
kevkojakeyip: log from cloud-init.log http://paste.openstack.org/show/736394/11:53
kevkojakeyip: but it is without issue only because i manually echoed CA_KEY into path above ...11:54
kevkojakeyip: without my manually action i have spam in log >>  waiting for CA to be made available for certificate manager api   >>>> /var/lib/cloud/instance/scripts/part-01411:55
kevkojakeyip: in that script is that comment ->  # if the certificate manager api is enabled, wait for the ca key to be handled  # by the heat container agent (required for the controller-manager)11:55
kevkojakeyip: so if i understand  correctly ..heat-container-agent should execute scripts in path /var/lib/heat-config/heat-config-script/ .... there is also script with echo CA_KEY ...i don't know why ..but this step is missing11:56
*** ttsiouts has joined #openstack-containers11:58
kevkojakeyip: how can i debug this behaviour ?11:59
jakeyipso CA_KEY is already populated in /etc/sysconfig/heat-params and CERT_MANAGER_API is True ?12:02
kevkojakeyip: yes, of course12:03
kevkojakeyip: in ps of instance i can see this >> http://paste.openstack.org/show/736396/12:04
kevkojakeyip: as u can see part-014 is running in infinity loop12:05
jakeyiphmm why is it sleeping in  /var/lib/heat-config/heat-config-script/b58e667a-6bfe-4b04-9686-ff27a89698112:05
kevkojakeyip: paste-014 is this -> http://paste.openstack.org/show/736397/12:05
*** dav_id has joined #openstack-containers12:12
*** salmankhan1 has joined #openstack-containers12:16
*** salmankhan has quit IRC12:16
*** salmankhan1 is now known as salmankhan12:16
dav_idHi, in OSA Queens 17.1.2 i try to deploy wiht Magnum, a  kubernetes Cluster, but in the Magnum Heat Template, there is a false auth_url in the Template: auth_url = https://public-ip:5000 in /etc/sysconfig/heat-params should be auth_url = https://public-ip:5000/v3 .12:20
dav_idAt the end /v3 is missing, is it a known Bug or any way to override it?12:20
dav_idI try to solve it since yesterday, help would be very appreciated.12:21
kevkodav_id: i think this value is taken from magnum.conf , isn't it ?12:21
dav_idi can show you my magnum conf, it look simular to other configs, like in cinder.conf12:22
kevkodav_id: show, use paste.openstack.org please12:22
dav_idOf course, thats my magnum.conf http://paste.openstack.org/show/736398/12:25
*** dodo_o has joined #openstack-containers12:28
kevkodav_id: did you mean /etc/os-collect-config.con ?  , because i don't have auth_url in /etc/sysconfig/heat-params12:29
*** ttsiouts has quit IRC12:30
*** ttsiouts has joined #openstack-containers12:31
*** ttsiouts has quit IRC12:32
*** ttsiouts has joined #openstack-containers12:32
dav_idHere you can have a look at /etc/sysconfig/heat-params  http://paste.openstack.org/show/736400/12:34
dav_idIn there under AUTH_URL="https://10.10.90.10:5000" the /v3 is missing, cousing the issue, i have changed i  manually, after that i works fine. But don't konw where to override it, maybe in magnum.conf?12:35
*** Nel1x has joined #openstack-containers12:38
kevkodav_id: set your auth_uri to /v3 format12:43
kevkodav_id: all auth_uri in your magnum config ..not auth_url ..but auth_uri12:43
dav_idBut auth_uri in my mangum config is already at /v3: auth_uri = http://10.10.90.9:5000/v312:45
dav_idthe auth_url is without /v3: auth_url = http://10.10.90.9:500012:45
jakeyiplooks similar to mine. another thing to check will be endpoint announced by catalog `openstack catalog list`12:47
kevkodav_id: add [clients_keystone] section with auth_uri12:48
kevkojakeyip: what about my problem ? :)12:48
jakeyipkevko: I had replied with question about sleep?12:49
kevkojakeyip: oh sorry , didn't see12:50
kevkojakeyip: echo "Waiting for Kubernetes API..."12:51
jakeyiphmm, don't think k8s can start if it's still waiting for ca info...? wonder if it's some ordering bug...12:54
kevkojakeyip: for me it sound already as a bug12:55
jakeyipcan't debug on my side, sorry, I'm not doing cert_maanger_api12:55
*** Namrata has quit IRC12:55
jakeyipguessing the heat script to populate the ca.key sound go first, then cloud-init can continue, then rest of heat-scripts ?12:56
kevkojakeyip: only issue is that script which is echoeing CA_KEY to /etc/kubernetes/certs/ca.key is not executed12:56
kevkojakeyip: yes ! exactly12:56
kevkojakeyip: but i don't have big knowledge of things ..i am greping greping greping ..see logs ..but don't know hot to make some "order" of scripts .... :/12:57
kevkojakeyip: btw, i'm using rocky version12:57
kevkojakeyip: stable12:57
jakeyipI guess a good start will be start tracing the ps that eventually launches the heat-config/ scripts, e.g. /var/lib/heat-config/hooks/script13:02
kevkojakeyip: yes, agree, now i'm creating new cluster13:03
kevkojakeyip: if you are interested , i can give you access to instance :)13:04
kevkojakeyip: or my openstack13:04
kevkojakeyip: it is only dev ..13:04
jakeyipI'll have to pass for now, it's past midnight here13:05
jakeyipzzz time :)13:05
kevkojakeyip: :D13:05
kevkojakeyip: i was wondering why noone is here always :D13:05
jakeyip:) good luck. I will look into it a bit more when I can...13:05
kevkojakeyip: thank you for your time :)13:06
jakeyipnow worries13:06
dodo_okevko: dav_id went to a meeting, i'll take over. :) setting [clients_keystone] auth_url to https://10.10.90.10:5000/v3 did not change the parameters that are deployed to the kube-master. this causes the part-011 script to fail, which creates the server.crt13:12
kevkododo_o: i said auth_uri :)13:13
kevkododo_o: no auth_url13:13
dodo_okevko: uh.. thats what i meant. it's right in the config :)13:14
*** kaiokmo has joined #openstack-containers13:31
kevkododo_o: your colleguage sent the config .. http://paste.openstack.org/show/736398/ and it isn't13:34
*** shrasool has joined #openstack-containers13:34
*** janki has joined #openstack-containers13:35
dodo_okevko: correct, but I just added [clients_keystone] auth_uri = https://10.10.90.10:5000/v3 to magnum.conf and tried to create a new kubernetes cluster13:36
*** salmankhan has quit IRC13:36
kevkododo_o: did you restart  ?13:36
dodo_okevko: yes, of course13:37
*** ttsiouts has quit IRC13:37
*** dav_id has quit IRC13:37
*** shrasool has quit IRC13:38
kevkododo_o: hmm , so don't know , maybe tcpdump and see what do you have in catalog ?13:39
kevkododo_o: maybe this will help ? https://review.openstack.org/#/c/615916/1/templates/heat.conf.j213:40
*** shrasool has joined #openstack-containers13:50
dodo_okevko: that sounds promising. i'll give it a try.13:50
dodo_okevko: thanks a lot :)13:51
kevkododo_o: but i really don't know if it will help ...which version are u using now ?13:51
dodo_okevko: we're currently on osa 17.1.2, so all the URIs in heat.conf are set to the internal ip.13:58
kevkoosa ?13:58
kevkowhat is osa ..13:58
dodo_okevko: openstack ansible :)14:00
*** zul has quit IRC14:00
*** salmankhan has joined #openstack-containers14:04
kevkododo_o: oh, nice ...our company will release our own deployment in some time14:05
kevkododo_o: salt :)14:05
*** zul has joined #openstack-containers14:16
kevkoguys ..do anyone now ..how to change the order of running deployment software ?14:27
*** ttsiouts has joined #openstack-containers14:30
*** ttsiouts has quit IRC14:47
*** ttsiouts has joined #openstack-containers14:48
*** ttsiouts has quit IRC14:48
*** ttsiouts has joined #openstack-containers14:49
*** udesale has joined #openstack-containers14:58
*** lpetrut has quit IRC14:58
*** hongbin has joined #openstack-containers15:05
*** ricolin has joined #openstack-containers15:10
*** ykarel is now known as ykarel|away15:11
*** ykarel|away has quit IRC15:15
*** zul has quit IRC15:18
*** udesale has quit IRC15:29
*** udesale has joined #openstack-containers15:29
*** ykarel|away has joined #openstack-containers15:32
*** ttsiouts has quit IRC15:32
*** ttsiouts has joined #openstack-containers15:33
*** ykarel|away is now known as ykarel15:34
*** shrasool has quit IRC15:36
*** ttsiouts has quit IRC15:37
*** itlinux has quit IRC15:38
*** ramishra has quit IRC15:39
*** ttsiouts has joined #openstack-containers15:52
*** munimeha1 has joined #openstack-containers15:53
*** udesale has quit IRC16:20
*** janki has quit IRC16:24
*** ricolin has quit IRC16:26
*** ykarel is now known as ykarel|away16:33
*** ttsiouts has quit IRC16:51
*** ttsiouts has joined #openstack-containers16:52
*** itlinux has joined #openstack-containers16:54
*** ttsiouts has quit IRC16:56
*** ykarel|away has quit IRC17:53
*** salmankhan has quit IRC18:16
*** chhagarw has quit IRC18:53
*** lpetrut has joined #openstack-containers18:58
*** chhagarw has joined #openstack-containers19:06
*** shrasool has joined #openstack-containers19:26
*** chhagarw has quit IRC19:48
*** shrasool has quit IRC19:48
*** shrasool has joined #openstack-containers19:49
*** shrasool has quit IRC19:51
*** itlinux has quit IRC19:56
*** kevko_ has joined #openstack-containers19:59
kevko_hi guys , found bug in magnum 7.0.220:00
kevko_heat-container-agent with tag rocky-stable not working20:00
kevko_heat-container-agent with tag rawhide is working20:00
*** kevko_ has quit IRC20:06
*** shrasool has joined #openstack-containers20:18
*** shrasool has quit IRC20:19
*** kevko_ has joined #openstack-containers20:35
kevko_strigazi: hi, strigazi ... could you please fix heat-container-agent:rocky-stable ? it is not working with magnum in 7.0.2, rawhide is working20:36
*** kevko_ has quit IRC20:44
flwangkevko: what's the issue?20:59
flwangwe can't just update the image only when somebody says it's not working, we need to know the root cause and fix it21:00
flwangrowhide has a multi region bug, we can't just rallback, we need to fix issue21:00
*** kevko_ has joined #openstack-containers21:01
kevko_hi, is someone here online ? :)21:01
*** ivve has quit IRC21:01
flwangkevko_: what's the your problem in heat-container-agent?21:08
kevko_flwang: it is not running scripts21:08
kevko_flwang: rawhide is running OK21:08
*** hongbin_ has joined #openstack-containers21:09
kevko_flwang: i can give you a log from both21:09
flwangwe can't rollback to rowhide21:09
*** hongbin_ has quit IRC21:09
flwangversion 'rowhide' has a multi region bug21:09
*** hongbin has quit IRC21:10
flwangi suggest using x509 as the cert manager and try again, i think there could be a bug of integration with barbican21:10
*** hongbin has joined #openstack-containers21:10
flwangkevko_: ^21:12
kevko_flwang: oh, ..so i should try with x509 and it should work ?21:13
kevko_flwang: but if i remember correct ..it was ok with certs and keys inside the instance21:14
flwangkevko_: it should, but i will continually testing the barbican case and will fix any issue if there is21:14
flwangkevko_: it definitely should work, but I don't have barbican env now, so i can't help reproduce21:15
flwangso for now, it's better that you just change the cert manager type and let's narrow the problem scope21:15
flwangif x509 works for you, then we know barbican case has bug21:15
flwangdoes that make sense for you?21:15
kevko_flwang: yes21:17
kevko_flwang: i can give you access to my test dev if you want :)21:17
flwangkevko_:  let's try x509 first, if it still doesn't work, then i'm happy to debug your instance21:18
flwangkevko_: just curious, are you going to deploy magnum in your prod?21:19
kevko_flwang: yes , we want21:19
kevko_flwang: i'm also debian maintainer for openstack packages in debian21:19
kevko_flwang: so i'm creating working packages for debian :)21:19
flwangcool21:19
flwangis your env public cloud or private?21:20
kevko_flwang: public21:20
flwangcool21:20
flwangwe (catalyst cloud) have deployed stable rocky on our prod, and it just works21:21
kevko_flwang: btw, i provided fix for python3 compatibility21:21
kevko_flwang: https://review.openstack.org/#/c/618756/21:21
flwangbut we don't have barbican, so that's why i'm not sure if there is a bug21:21
flwangkevko_: oh, i saw that patch, but i haven't got time to test, thank you for the contribution21:22
flwangi'm adding it to my list21:22
kevko_flwang: catalyst ? i was near your place in market place ...or i was on some talk maybe ...don't remember ..in Berlin ..21:23
kevko_flwang: debian is using only python3 version of all projects21:23
flwangwe don't have booth at Berlin, so you should be in our session at Berlin21:24
kevko_flwang: we switched during the queens21:24
flwangkevko_: got21:24
kevko_flwang: Failed to create certificates for Cluster: 01e0b8e8-dbcd-4263-9aed-6eb064d1326921:26
flwangcan you pls show me the log of cloud-init-output.log21:27
*** kevko__ has joined #openstack-containers21:31
kevko__flwang: do i need to do something more than switch config in magnum.conf ?21:32
*** kevko_ has quit IRC21:32
kevko__flwang: i have still --label cert_manager_api=true , should i turn off this ? ...sorry ..my knowledge of magnum is not still good ... :) just starting21:32
flwangkevko__: just remove the label, i assume default it's false21:34
flwangor set it as Flase21:34
kevko__flwang: actually , what is it ?21:34
kevko__flwang: i was unable to build cluster without it ... found in this papermail ... let me provide21:35
kevko__flwang: http://lists.openstack.org/pipermail/openstack-dev/2018-August/133721.html21:35
flwangi don't know joe's env, but we don't use it by default, and by default it's false21:37
flwangso let's try set it false or remove it21:37
kevko__flwang: ok21:39
kevko__flwang: i've created two cluster ..without and with label ..so will see21:39
kevko__flwang: will let you know when it will be completed/failed ...21:40
flwangthanks21:41
kevko__flwang: to which time you will be here ?21:42
flwangnow it's 10:42AM for me21:43
flwangi should be available in next 1 hour21:43
kevko__flwang: for me it is 10:42 PM :)21:43
kevko__flwang: ok, my env is nested openstack in openstack21:44
kevko__flwang: and currently my colleguage is doing some upgrade ...21:44
kevko__flwang: so maybe it will take some time21:44
kevko__flwang: give me 15 minutes .. i have to move to another place21:44
*** kevko__ has quit IRC21:45
flwangno rush, man21:46
flwangI'm based in NZ21:47
*** rcernin has joined #openstack-containers21:57
*** kevko_ has joined #openstack-containers22:15
kevko_flwang: i'm back22:15
flwangyes22:16
kevko_flwang: failed to create certificates22:17
kevko_flwang: let me check magnum log22:17
kevko_flwang: http://paste.openstack.org/show/736451/22:19
flwanghas the stack/master been created?22:20
kevko_flwang: i can't see heat stack22:21
kevko_flwang: so, i think this will be also **maybe** python3 related bug22:22
kevko_flwang: it is failing somewhere on the way towards heat ..22:22
kevko_flwang: let me write some debug messages in a code22:22
flwangcan you run openstack coe cluster show <id> to see the failed reason?22:23
kevko_flwang: failed to create certificates for Cluster: [ ID of cluster ]22:25
flwangcan you see any error in condutor?22:26
kevko_flwang: yes , as i said ..it is python3 issue22:29
kevko_flwang: bytes vs string ..python2 vs python322:29
flwangok, then hot fix it and try again?22:29
kevko_flwang: ValueError: A string is required in field certificate, not a bytes22:30
kevko_flwang: yes ..going to do ..22:30
kevko_flwang: but ..as i said ..now my colleguage upgrading openstack above my openstack ..so i'm loosing my connection ..because of virt routers migrating ..process of migrating neutron run now ...but i'm working via spice console :D :D :D22:31
flwangno rush22:32
kevko_flwang: maybe it will be better if i will chat you tomorrow ... you are leaving in a while , aren't you ?22:50
kevko_flwang: now i don't have good connection into instance22:51
flwangok22:55
flwangsee you next Monday :)22:55
kevko_flwang: next monday ?22:57
kevko_flwang: ok then :)22:57
kevko_flwang: i will be here tomorrow also, who can i contact with python3 issues ?22:57
flwangif it's confirmed your patch works, then we can merge it22:58
jakeyipwhere are you based kevko_ ?23:06
kevko_jakeyip: czech republic :) , but working all the time :)23:06
kevko_jakeyip: from morning to morning :)23:06
jakeyipkevko_: Hi from AU over here :)23:06
*** lpetrut has quit IRC23:07
kevko_jakeyip: did you mean australia ? :)23:07
jakeyipkevko_: yes23:07
kevko_jakeyip: oh, nice :)23:07
kevko_jakeyip: but I am originally from Slovakia :)23:07
jakeyipnice :) hope we can meet up at one of the summits one day23:10
kevko_jakeyip: i was in Berlin now :)23:13
kevko_jakeyip: as I am debian maintainer also for Openstack, maybe i will be in Brazilia on DebConf ..so maybe there ..if you will be there guys23:14
jakeyipha don't think... my travel privs doesn't extend that far :P23:22
kevko_jakeyip: have you been in berlin ?23:26
jakeyipyeah I did go to Berlin23:39
*** munimeha1 has quit IRC23:50

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