| *** ivve has quit IRC | 00:23 | |
| *** ivve has joined #openstack-containers | 00:35 | |
| openstackgerrit | Mohammed Naser proposed openstack/magnum master: wip: multinode conformance https://review.openstack.org/625448 | 00:46 |
|---|---|---|
| *** kothari has quit IRC | 00:49 | |
| openstackgerrit | Mohammed Naser proposed openstack/magnum master: wip: multinode conformance https://review.openstack.org/625448 | 00:49 |
| openstackgerrit | Mohammed Naser proposed openstack/magnum master: wip: multinode conformance https://review.openstack.org/625448 | 01:44 |
| *** dave-mccowan has joined #openstack-containers | 02:07 | |
| openstackgerrit | Mohammed Naser proposed openstack/magnum master: wip: multinode conformance https://review.openstack.org/625448 | 02:11 |
| *** hongbin has joined #openstack-containers | 02:11 | |
| *** hongbin has quit IRC | 02:28 | |
| openstackgerrit | Mohammed Naser proposed openstack/magnum master: wip: multinode conformance https://review.openstack.org/625448 | 02:40 |
| *** dave-mccowan has quit IRC | 02:47 | |
| *** hongbin has joined #openstack-containers | 02:50 | |
| *** lbragstad has joined #openstack-containers | 03:26 | |
| *** PagliaccisCloud has quit IRC | 03:37 | |
| *** ramishra has joined #openstack-containers | 03:42 | |
| *** ykarel has joined #openstack-containers | 03:43 | |
| *** udesale has joined #openstack-containers | 03:49 | |
| *** zufar has quit IRC | 03:50 | |
| *** itlinux has quit IRC | 03:59 | |
| *** itlinux has joined #openstack-containers | 04:29 | |
| *** itlinux has quit IRC | 04:30 | |
| *** ricolin has joined #openstack-containers | 04:31 | |
| *** hongbin has quit IRC | 04:37 | |
| *** ykarel has quit IRC | 04:40 | |
| *** PagliaccisCloud has joined #openstack-containers | 04:48 | |
| *** ykarel has joined #openstack-containers | 05:01 | |
| *** ykarel has quit IRC | 05:10 | |
| *** ykarel has joined #openstack-containers | 05:12 | |
| *** itlinux has joined #openstack-containers | 05:33 | |
| *** janki has joined #openstack-containers | 05:36 | |
| *** rcernin has joined #openstack-containers | 05:46 | |
| *** rcernin has quit IRC | 05:47 | |
| *** rcernin has joined #openstack-containers | 06:09 | |
| *** rcernin has quit IRC | 06:09 | |
| *** rcernin has joined #openstack-containers | 06:09 | |
| *** rcernin has quit IRC | 06:09 | |
| *** itlinux has quit IRC | 07:09 | |
| *** itlinux has joined #openstack-containers | 07:24 | |
| *** yolanda has joined #openstack-containers | 07:50 | |
| *** ykarel is now known as ykarel|lunch | 07:58 | |
| *** ignaziocassano1 has joined #openstack-containers | 08:00 | |
| *** pcaruana has joined #openstack-containers | 08:18 | |
| ignaziocassano1 | Please, help me on magnum under queens | 08:29 |
| ignaziocassano1 | kubernets cluster does not work | 08:30 |
| ignaziocassano1 | the cloud-init output log on master reports docker volums is created but before it reports: /var/lib/cloud/instance/scripts/part-007: line 57: /etc/etcd/etcd.conf: No such file or directory | 08:32 |
| *** ykarel|lunch is now known as ykarel | 08:39 | |
| ignaziocassano1 | Anyone could help me , please ? | 08:40 |
| strigazi | ignaziocassano1: can you show me: cluster show, cluster template show, stack show, and /var/log/cloud-init-output.log | 08:48 |
| strigazi | ignaziocassano1: From the output, I think that the node failed to pull the etcd container. | 08:49 |
| ignaziocassano1 | would you like I send the output log ? in /etc/sysconfig/heat-params proxy is configured correctly | 08:55 |
| ignaziocassano1 | ah ok | 08:55 |
| ignaziocassano1 | sorry | 08:56 |
| ignaziocassano1 | I did nod read your request | 08:56 |
| ignaziocassano1 | root@tst2-osctrl01 ~]# openstack coe cluster show rrrrrrrrrrrrrr | 08:58 |
| ignaziocassano1 | +---------------------+------------------------------------------------------------+ | 08:58 |
| ignaziocassano1 | | Field | Value | | 08:58 |
| ignaziocassano1 | +---------------------+------------------------------------------------------------+ | 08:58 |
| ignaziocassano1 | | status | CREATE_IN_PROGRESS | | 08:58 |
| ignaziocassano1 | | cluster_template_id | 3eb07fc2-dbdc-4468-9d4d-c3ebab572d2f | | 08:58 |
| ignaziocassano1 | | node_addresses | [] | | 08:58 |
| ignaziocassano1 | | uuid | a40e5cb5-0b35-47f2-b65c-d4b5a6a48d6b | | 08:58 |
| ignaziocassano1 | | stack_id | 6ae09921-dea5-40f3-a4a7-b66b3446d4f0 | | 08:58 |
| ignaziocassano1 | | status_reason | None | | 08:58 |
| ignaziocassano1 | | created_at | 2018-12-17T08:26:34+00:00 | | 08:58 |
| ignaziocassano1 | | updated_at | 2018-12-17T08:26:41+00:00 | | 08:58 |
| ignaziocassano1 | | coe_version | None | | 08:58 |
| ignaziocassano1 | | labels | {u'etcd_volume_size': u'80'} | | 08:58 |
| ignaziocassano1 | | faults | | | 08:58 |
| ignaziocassano1 | | keypair | opstkcsi | | 08:58 |
| ignaziocassano1 | | api_address | None | | 08:58 |
| ignaziocassano1 | | master_addresses | [] | | 08:58 |
| ignaziocassano1 | | create_timeout | 60 | | 08:58 |
| ignaziocassano1 | | node_count | 1 | | 08:58 |
| ignaziocassano1 | | discovery_url | https://discovery.etcd.io/0fbe42ac761147fb3e2e46bfaec92d93 | | 08:58 |
| strigazi | ignaziocassano1 paste.openstack.org | 08:58 |
| ignaziocassano1 | | master_count | 1 | | 08:58 |
| ignaziocassano1 | | container_version | None | | 08:58 |
| ignaziocassano1 | | name | rrrrrrrrrrrrrr | | 08:58 |
| ignaziocassano1 | | master_flavor_id | m1.small | | 08:58 |
| ignaziocassano1 | | flavor_id | m1.small | | 08:58 |
| ignaziocassano1 | +---------------------+------------------------------------------------------------+ | 08:58 |
| ignaziocassano1 | ok | 09:00 |
| ignaziocassano1 | sorry but it's firts time I use irc | 09:00 |
| ignaziocassano1 | Paste #737458 | 09:02 |
| ignaziocassano1 | I connected to paste.openstack.org and pasted text | 09:02 |
| ignaziocassano1 | its return the above code | 09:02 |
| ignaziocassano1 | Paste #737459 | 09:04 |
| ignaziocassano1 | Can you read my baste ? | 09:04 |
| ignaziocassano1 | Can you read my paste ? | 09:06 |
| strigazi | yes | 09:06 |
| strigazi | and /var/log/cloud-init-output.log | 09:07 |
| ignaziocassano1 | ok | 09:07 |
| ignaziocassano1 | just a moment | 09:07 |
| *** lpetrut has joined #openstack-containers | 09:09 | |
| *** ricolin has quit IRC | 09:10 | |
| ignaziocassano1 | Paste #737461 | 09:10 |
| ignaziocassano1 | googling I found other two peoples with the same issue with queens | 09:17 |
| strigazi | can you access docker.io from the node? | 09:19 |
| strigazi | it seems that the atomic pull command doesn't work | 09:19 |
| openstackgerrit | Merged openstack/magnum master: Delete Octavia loadbalancers for fedora atomic k8s driver https://review.openstack.org/497144 | 09:20 |
| ignaziocassano1 | I am going to try | 09:20 |
| strigazi | atomic install --system --system-package no --storage ostree --name server docker.io/openstackmagnum/kubernetes-apiserver:v1.11.5-1 | 09:21 |
| ignaziocassano1 | In the master bash proxy is not setted | 09:21 |
| ignaziocassano1 | so I must set it | 09:21 |
| ignaziocassano1 | If I set the proxy in bash I can | 09:23 |
| ignaziocassano1 | Probably the cloud init does not inherit proxy settings ? | 09:24 |
| ignaziocassano1 | I am going to sent new paste | 09:25 |
| *** jonaspaulo_ has joined #openstack-containers | 09:25 | |
| ignaziocassano1 | Paste #737462 | 09:26 |
| openstackgerrit | Wayne Chan proposed openstack/python-magnumclient master: Update mailinglist from dev to discuss https://review.openstack.org/625524 | 09:28 |
| openstackgerrit | Yang Le proposed openstack/magnum master: Update mailing list from dev to discuss https://review.openstack.org/625525 | 09:28 |
| strigazi | ignaziocassano1: that is the problem, proxy | 09:28 |
| ignaziocassano1 | Why it dos not inherit it ? | 09:29 |
| ignaziocassano1 | in /etc/sysconfig/heat it is a variable with proxy | 09:30 |
| ignaziocassano1 | sorry : heat-params | 09:30 |
| *** ttsiouts has joined #openstack-containers | 09:34 | |
| ignaziocassano1 | Strigazi: the proxy variables in /etc/sysconfig/heat-params are HTTP_PROXY and HTTPS_PROXY.....do you think uppercase is the problem ? | 09:36 |
| ignaziocassano1 | No, it is not. http_proxy in uppercase works. The problem is it does not inherit the variable :-( | 09:40 |
| ignaziocassano1 | Any patch, please ? | 09:41 |
| strigazi | ignaziocassano1: you can try, try with to unset lowercase, source with upper case ans try to pull again | 09:41 |
| strigazi | ignaziocassano1: you can try, try to unset lowercase, source with uppercase ans try to pull again | 09:41 |
| ignaziocassano1 | I did it and it reports: kubernetes-apiserver:v1.11.5-1 | 09:42 |
| ignaziocassano1 | server already present | 09:42 |
| ignaziocassano1 | curl works either with uppercase or lowercase | 09:43 |
| strigazi | atomic install --system --system-package no --storage ostree --name server docker.io/openstackmagnum/kubernetes-proxy:v1.11.5-1 | 09:44 |
| ignaziocassano1 | it works | 09:45 |
| ignaziocassano1 | with HTTP_PROXY and HTTPS_PROXY variables | 09:46 |
| strigazi | with uppercase? | 09:46 |
| strigazi | and lowercase? | 09:46 |
| ignaziocassano1 | yes | 09:49 |
| ignaziocassano1 | both | 09:49 |
| ignaziocassano1 | first attemps with lowercase: OK | 09:51 |
| ignaziocassano1 | second attemps with uppercase: OK | 09:51 |
| *** rtjure has quit IRC | 09:52 | |
| ignaziocassano1 | stringazi: just a moment for a coffee and I come back | 09:54 |
| openstackgerrit | Bharat Kunwar proposed openstack/magnum master: Config drive support for k8s_fedora_atomic swarm_mode https://review.openstack.org/611291 | 09:57 |
| openstackgerrit | Bharat Kunwar proposed openstack/magnum master: Config drive support for k8s_fedora_atomic swarm_mode https://review.openstack.org/611291 | 09:57 |
| ignaziocassano1 | here I am | 09:59 |
| *** PagliaccisCloud has quit IRC | 10:02 | |
| ignaziocassano1 | Strigazi: the atomic endpoint could be copied on a local site ? | 10:04 |
| ignaziocassano1 | This is a strange idea but I need it works | 10:05 |
| strigazi | you can mirror all the containers you need and use container_infra_prefix to pull from a local regisytr | 10:06 |
| strigazi | you can mirror all the containers you need and use container_infra_prefix to pull from a local regisrty | 10:06 |
| strigazi | *registry | 10:06 |
| strigazi | container_infra_prefix | 10:07 |
| strigazi | ignaziocassano1: https://docs.openstack.org/magnum/latest/user/ | 10:07 |
| ignaziocassano1 | OK | 10:09 |
| ignaziocassano1 | but if you read the cloud-init output the source of /etc/sysconfig/heat-paramas is done after the atomic command | 10:10 |
| *** lbragstad has quit IRC | 10:10 | |
| *** PagliaccisCloud has joined #openstack-containers | 10:10 | |
| ignaziocassano1 | So I think a patch is required | 10:10 |
| strigazi | what you mean? what is after? | 10:13 |
| ignaziocassano1 | If you read cloud init output it runs atomic install --storage ostree --system --system-package no --set REQUESTS_CA_BUNDLE=/etc/pki/tls/certs/ca-bundle.crt --name heat-container-agent docker.io/openstackmagnum/heat-container-agent:rawhide | 10:15 |
| ignaziocassano1 | then it runs . /etc/sysconfig/heat-params | 10:16 |
| ignaziocassano1 | So, atomic command does not get variables | 10:16 |
| ignaziocassano1 | Could you confirm it ? | 10:17 |
| strigazi | ignaziocassano1: https://github.com/openstack/magnum/blob/master/magnum/drivers/common/templates/kubernetes/fragments/configure-etcd.sh | 10:18 |
| strigazi | line 3 sources the params | 10:18 |
| strigazi | line 38 runs atomic | 10:18 |
| ignaziocassano1 | So, it should work | 10:19 |
| strigazi | yes | 10:19 |
| ignaziocassano1 | mmmm | 10:19 |
| *** PagliaccisCloud has quit IRC | 10:21 | |
| strigazi | try to put set -x in the beginning of this fules and try to create a new cluster | 10:23 |
| ignaziocassano1 | the script /var/lib/cloud/instance/scripts/part-007 runs atomic and initialize variables from /etc/sysconfig/heat-params | 10:23 |
| ignaziocassano1 | which file ? | 10:23 |
| ignaziocassano1 | Sorry I lost somesthing | 10:23 |
| strigazi | https://github.com/openstack/magnum/blob/master/magnum/drivers/common/templates/kubernetes/fragments/configure-etcd.sh | 10:23 |
| ignaziocassano1 | Sorry I lost something | 10:24 |
| ignaziocassano1 | That file is on git, where I find it on my installation ? | 10:24 |
| ignaziocassano1 | On the magnum controller ? | 10:24 |
| *** salmankhan has joined #openstack-containers | 10:25 | |
| ignaziocassano1 | Yes, it is on the magnum controller | 10:25 |
| ignaziocassano1 | Must I modify it ? | 10:26 |
| ignaziocassano1 | I modified it | 10:27 |
| ignaziocassano1 | let me run another cluster | 10:27 |
| strigazi | ok | 10:27 |
| ignaziocassano1 | it is starting | 10:34 |
| ignaziocassano1 | would you like the new cloud-init output log ? | 10:34 |
| strigazi | yes, it works now? | 10:35 |
| ignaziocassano1 | no | 10:35 |
| ignaziocassano1 | it is waiting for the atomic command output | 10:35 |
| ignaziocassano1 | probably it does not read the proxy variable | 10:36 |
| ignaziocassano1 | + mount -a | 10:36 |
| ignaziocassano1 | + chown -R etcd.etcd /var/lib/etcd | 10:36 |
| ignaziocassano1 | + chmod 755 /var/lib/etcd | 10:36 |
| ignaziocassano1 | + _prefix=docker.io/openstackmagnum/ | 10:36 |
| ignaziocassano1 | 10.102.186.116+ atomic install --system-package no --system --storage ostree --name=etcd docker.io/openstackmagnum/etcd:v3.2.7 | 10:36 |
| ignaziocassano1 | sorry | 10:37 |
| ignaziocassano1 | would you like a paste of the file ? | 10:37 |
| strigazi | ignaziocassano1: if you can paste the full output.log in paster.o.o | 10:38 |
| ignaziocassano1 | ok | 10:40 |
| ignaziocassano1 | http://paste.openstack.org/show/737472/ | 10:45 |
| ignaziocassano1 | Can you get it ? | 10:47 |
| strigazi | ignaziocassano1: doesn't work it tries to pull it but nothing | 10:51 |
| strigazi | ignaziocassano1: before atomic pull try to add: | 10:51 |
| strigazi | export HTTP_PROXY=${HTTP_PROXY} | 10:52 |
| strigazi | export HTTPS_PROXY=${HTTPS_PROXY} | 10:52 |
| *** salmankhan1 has joined #openstack-containers | 10:52 | |
| ignaziocassano1 | OK | 10:53 |
| ignaziocassano1 | I am going to modify the file and run another cluster | 10:53 |
| *** salmankhan has quit IRC | 10:54 | |
| *** salmankhan1 is now known as salmankhan | 10:54 | |
| strigazi | brtknr: can you test occm with cinder volumes? | 10:54 |
| ignaziocassano1 | I inser also an echo instruction to veryfy the content of the variable | 10:55 |
| brtknr | strigazi: we do not have support for cinder volumes on our cluster... there was a patch to support it but we managed to make do with manila | 10:55 |
| strigazi | brtknr: we use manial as well at CERN. Other sites offer cinder, like catalyst. | 10:56 |
| strigazi | brtknr: I couldn't mount cinder volumes with occm :( | 10:56 |
| brtknr | strigazi: anything noteworthy in the logs? | 10:57 |
| brtknr | occm logs* | 10:57 |
| brtknr | strigazi: are you using cinder-csi-plugin or cinder-flexvolume? | 10:59 |
| strigazi | brtknr: nothing in the logs | 11:00 |
| strigazi | brtknr: no, I just moved from the intree provider to external CPO. no other changes | 11:00 |
| brtknr | are you able to create/mount cinder volumes outside of magnum? | 11:00 |
| brtknr | outside of kubernetes* | 11:01 |
| ignaziocassano1 | ehiiiiii atomic is downloading | 11:01 |
| strigazi | ignaziocassano1: the values need to exported then | 11:01 |
| strigazi | brtknr: I'm doing this https://docs.openstack.org/magnum/latest/user/#using-cinder-in-kubernetes | 11:02 |
| ignaziocassano1 | Yes I exported them | 11:02 |
| ignaziocassano1 | now it is in : configuring kubernetes (master) | 11:03 |
| brtknr | strigazi: you may need to do this instead: https://github.com/kubernetes/cloud-provider-openstack/blob/master/examples/cinder-csi-plugin/nginx.yaml | 11:04 |
| ignaziocassano1 | it is always there | 11:04 |
| brtknr | strigazi: or this: https://github.com/kubernetes/cloud-provider-openstack/blob/master/examples/cinder-flexvolume/nginx.yaml | 11:05 |
| strigazi | ignaziocassano1: you need to do the same in all places where atomic is used | 11:05 |
| brtknr | strigazi: the way you are doing it looks like an old API | 11:05 |
| strigazi | brtknr: i guess so | 11:06 |
| brtknr | strigazi: lemme know how it goes :) | 11:06 |
| brtknr | strigazi: the second one looks closer to the current API | 11:07 |
| ignaziocassano1 | Strigazi: do you know the name of the script I must modify ? | 11:08 |
| brtknr | strigazi: we may be able to remove the `--volume-driver cinder` option altogether | 11:08 |
| ignaziocassano1 | configure-kubernetes-master.sh onfigure-kubernetes-minion.sh.... ? | 11:09 |
| *** udesale has quit IRC | 11:10 | |
| strigazi | ignaziocassano1: and start-heat-container-agent | 11:10 |
| ignaziocassano1 | ok | 11:17 |
| ignaziocassano1 | I modified them | 11:17 |
| ignaziocassano1 | I am going to create another cluster | 11:17 |
| ignaziocassano1 | with your name in honor of your help | 11:18 |
| ignaziocassano1 | all atomic install went fine but now the output reports: | 11:25 |
| ignaziocassano1 | Waiting for Kubernetes API... | 11:25 |
| ignaziocassano1 | + curl --silent http://127.0.0.1:8080/version | 11:25 |
| ignaziocassano1 | nothing on 8080 port | 11:28 |
| ignaziocassano1 | it continues with curl sleep and curl on 8080 | 11:30 |
| ignaziocassano1 | in the cloudinit I read Failed to enable unit: Unit file kube-apiserver.service does not exist. | 11:32 |
| ignaziocassano1 | Failed to start kube-apiserver.service: Unit kube-apiserver.service not found. | 11:32 |
| ignaziocassano1 | activating service kube-controller-manager | 11:32 |
| ignaziocassano1 | Failed to enable unit: Unit file kube-controller-manager.service does not exist. | 11:32 |
| ignaziocassano1 | Failed to start kube-controller-manager.service: Unit kube-controller-manager.service not found. | 11:32 |
| ignaziocassano1 | activating service kube-scheduler | 11:32 |
| ignaziocassano1 | Failed to enable unit: Unit file kube-scheduler.service does not exist. | 11:32 |
| ignaziocassano1 | Failed to start kube-scheduler.service: Unit kube-scheduler.service not found | 11:32 |
| *** PagliaccisCloud has joined #openstack-containers | 11:33 | |
| ignaziocassano1 | OPSsss... it was my error in modifying script for proxy | 11:35 |
| ignaziocassano1 | I try again...sorry | 11:35 |
| ignaziocassano1 | it finisched the master and it is starting minion | 11:47 |
| ignaziocassano1 | strigazi ? | 11:47 |
| strigazi | I'll be back, if it stats the minion, it should be good | 11:51 |
| ignaziocassano1 | minion seems to loop here: | 11:51 |
| ignaziocassano1 | curl -sf --cacert /etc/flanneld/certs/ca.crt --cert /etc/flanneld/certs/proxy.crt --key /etc/flanneld/certs/proxy.key 'https://10.0.0.10:2379/v2/keys/atomic.io/network/config?quorum=false&recursive=false&sorted=false' | 11:51 |
| ignaziocassano1 | + echo 'Waiting for flannel configuration in etcd...' | 11:51 |
| ignaziocassano1 | Waiting for flannel configuration in etcd... | 11:51 |
| ignaziocassano1 | + sleep 5 | 11:51 |
| strigazi | ignaziocassano1: check the flannel-service.sh if it has atomic, but you are almost there | 11:51 |
| ignaziocassano1 | on minion ? | 11:52 |
| strigazi | ignaziocassano1: do you pass --docker-volume-size in the CT? try without --docker-volume-size and --dockerstorage-driver overlay2 | 11:52 |
| ignaziocassano1 | ahhh | 11:52 |
| *** rtjure has joined #openstack-containers | 11:53 | |
| ignaziocassano1 | when I create a template I must give a flavor grater than the flavor I pass to cluster, otherwhise the docker volume is not created | 11:54 |
| ignaziocassano1 | and an error is given in the cloud init because no space is found on atomic volume group :-( | 11:54 |
| ignaziocassano1 | Anycase | 11:55 |
| ignaziocassano1 | I pass a docker volume-size | 11:55 |
| ignaziocassano1 | I am going without docker-volume-size and I check flannel-service.sh | 11:56 |
| brtknr | ignaziocassano1: You're in a for a win :) | 11:56 |
| ignaziocassano1 | I hope | 11:57 |
| ignaziocassano1 | Where I fount the flannel-service.sh ? | 11:57 |
| ignaziocassano1 | Where I find the flannel-service.sh ? | 11:58 |
| brtknr | check /var/log/cloud-init.log | 12:00 |
| brtknr | its one of the startup scripts.... but it will be called something different | 12:00 |
| brtknr | like 001.sh or something | 12:01 |
| ignaziocassano1 | OK | 12:01 |
| ignaziocassano1 | first I am tryng without docker volume size ....ok ? | 12:01 |
| brtknr | ignaziocassano1: sure... | 12:01 |
| ignaziocassano1 | and overlay2 | 12:02 |
| ignaziocassano1 | I think flannel is generate by write-network-config.sh | 12:08 |
| ignaziocassano1 | any case....master has finisched | 12:10 |
| ignaziocassano1 | minion ist starting | 12:10 |
| ignaziocassano1 | minion is starting | 12:11 |
| ignaziocassano1 | also in this case minion loops | 12:13 |
| ignaziocassano1 | curl -sf --cacert /etc/flanneld/certs/ca.crt --cert /etc/flanneld/certs/proxy.crt --key /etc/flanneld/certs/proxy.key 'https://10.0.0.18:2379/v2/keys/atomic.io/network/config?quorum=false&recursive=false&sorted=false' | 12:13 |
| ignaziocassano1 | + echo 'Waiting for flannel configuration in etcd...' | 12:13 |
| ignaziocassano1 | Waiting for flannel configuration in etcd... | 12:13 |
| ignaziocassano1 | + sleep 5 | 12:13 |
| ignaziocassano1 | in the minion cloud init output I found atomic install --storage ostree --system --system-package=no --name=flanneld docker.io/openstackmagnum/flannel:v0.9.0 | 12:17 |
| ignaziocassano1 | but it worked | 12:17 |
| strigazi | 1. you can try calico | 12:21 |
| strigazi | show me your cluster template | 12:22 |
| ignaziocassano1 | I think curl i blocked because it does not have noproxy | 12:22 |
| ignaziocassano1 | 10.0.0.18:2379 should be the master address. Is correct ? | 12:22 |
| *** PagliaccisCloud has quit IRC | 12:23 | |
| strigazi | 10.0.0.18:2379 should be master | 12:23 |
| ignaziocassano1 | OK | 12:23 |
| ignaziocassano1 | If I thhr on minion from bash the curl return exit status 22 with the proxy | 12:24 |
| ignaziocassano1 | without it returns 0 | 12:24 |
| ignaziocassano1 | If I try on minion from bash the curl return exit status 22 with the proxy | 12:24 |
| ignaziocassano1 | without it returns 0 | 12:25 |
| ignaziocassano1 | could be the problem ? | 12:25 |
| ignaziocassano1 | No I think not | 12:27 |
| ignaziocassano1 | minion is running /var/lib/cloud/instance/scripts/part-008 | 12:27 |
| strigazi | maybe you must unset the proxy for this? | 12:27 |
| ignaziocassano1 | and in that script all variables are correctly | 12:27 |
| ignaziocassano1 | it runs the curl and goes in loop | 12:28 |
| strigazi | maybe after the atomic cmd you need to unset *_PROXY | 12:29 |
| brtknr | ignaziocassano1: why do you have {u'etcd_volume_size': u'80'}? | 12:30 |
| *** janki has quit IRC | 12:31 | |
| ignaziocassano1 | I passed it with labels | 12:31 |
| ignaziocassano1 | I can remove it and restart | 12:31 |
| ignaziocassano1 | In my mind I thought etcd need a volume | 12:32 |
| ignaziocassano1 | Do I remove it and restart ? | 12:32 |
| ignaziocassano1 | as far as the proxy is concerned I try to modify /var/lib/cloud/instance/scripts/part-008 on the fly and run it | 12:33 |
| ignaziocassano1 | ok ? | 12:33 |
| strigazi | I don't know by heart what part-008 is | 12:33 |
| ignaziocassano1 | it starts the curl that generates the loop | 12:35 |
| ignaziocassano1 | Unsetting the proxy before curl it forks | 12:35 |
| ignaziocassano1 | But I must serch on the magnum server where it is implemented and modify it | 12:36 |
| ignaziocassano1 | are you agree ? | 12:36 |
| ignaziocassano1 | it is in kubernetes minion shell script | 12:38 |
| *** udesale has joined #openstack-containers | 12:39 | |
| ignaziocassano1 | I also will delete the label for etcd_volume_size | 12:40 |
| ignaziocassano1 | OK ? | 12:41 |
| ignaziocassano1 | started new cluster | 12:46 |
| ignaziocassano1 | meanwhile I need a coffee | 12:46 |
| *** lpetrut has quit IRC | 12:51 | |
| ignaziocassano1 | here I am | 12:51 |
| ignaziocassano1 | many coffees , no food ;-( | 12:51 |
| ignaziocassano1 | now the cloudinit on minion works fine | 12:53 |
| ignaziocassano1 | I erased the proxy where it runs curl to get information from master | 12:54 |
| ignaziocassano1 | the heat stack finished successfully | 12:54 |
| ignaziocassano1 | strigazi have you abbandoned me ? | 12:55 |
| strigazi | ignaziocassano1: I'm here, so CREATE_COMPLETE? | 12:57 |
| *** janki has joined #openstack-containers | 12:59 | |
| ignaziocassano1 | yes | 13:04 |
| ignaziocassano1 | now I am going on with the documentation with kubectl command | 13:04 |
| ignaziocassano1 | I do not know kubernetes , so I must install kubeclt on my workstation | 13:05 |
| *** jonaspaulo_ has quit IRC | 13:10 | |
| strigazi | ignaziocassano1: see here step three for cluster config https://docs.openstack.org/magnum/latest/install/launch-instance.html#provision-a-kubernetes-cluster-and-create-a-deployment | 13:12 |
| *** Bhujay has joined #openstack-containers | 13:15 | |
| ignaziocassano1 | kubectl -n kube-system get po | 13:17 |
| ignaziocassano1 | pods is forbidden: User "Magnum User" cannot list pods in the namespace "kube-system" | 13:18 |
| ignaziocassano1 | :-( | 13:18 |
| strigazi | you have and older magnum client | 13:18 |
| strigazi | ignaziocassano1: try to upgrade you client | 13:18 |
| ignaziocassano1 | ok | 13:18 |
| strigazi | ignaziocassano1: https://docs.openstack.org/releasenotes/magnum/queens.html#upgrade-notes | 13:19 |
| ignaziocassano1 | which version I need ? | 13:22 |
| ignaziocassano1 | I installed lated on my ubuntu 16.04 | 13:22 |
| ignaziocassano1 | 1.13 | 13:22 |
| ignaziocassano1 | Ahhh | 13:23 |
| ignaziocassano1 | I need olso the magnum client | 13:23 |
| ignaziocassano1 | ok | 13:23 |
| *** dave-mccowan has joined #openstack-containers | 13:25 | |
| *** Bhujay has quit IRC | 13:33 | |
| *** tbarron has joined #openstack-containers | 13:35 | |
| *** dave-mccowan has quit IRC | 13:42 | |
| *** pcaruana has quit IRC | 13:50 | |
| ignaziocassano1 | I installed magnum client 2.8.0 | 13:58 |
| *** lbragstad has joined #openstack-containers | 13:58 | |
| ignaziocassano1 | but the command kubectl -n kube-system get po gives Error from server (Forbidden): pods is forbidden: User "Magnum User" cannot list pods in the namespace "kube-system" | 13:59 |
| *** lbragstad has quit IRC | 13:59 | |
| *** ttsiouts has quit IRC | 14:00 | |
| *** ttsiouts has joined #openstack-containers | 14:00 | |
| *** ttsiouts has quit IRC | 14:04 | |
| *** ttsiouts has joined #openstack-containers | 14:04 | |
| *** lbragstad has joined #openstack-containers | 14:04 | |
| *** pcaruana has joined #openstack-containers | 14:05 | |
| brtknr | did you do `openstack coe cluster config <cluster-name>`? | 14:06 |
| brtknr | ignaziocassano1: | 14:06 |
| strigazi | ignaziocassano1 you need 2.9.0 | 14:07 |
| *** dave-mccowan has joined #openstack-containers | 14:07 | |
| brtknr | strigazi: any luck with cinder? | 14:08 |
| brtknr | strigazi: with the two new approaches? | 14:09 |
| ignaziocassano1 | Yes | 14:10 |
| ignaziocassano1 | I do | 14:10 |
| *** dave-mccowan has quit IRC | 14:11 | |
| ignaziocassano1 | $(openstack coe cluster config strkub --dir ~/clusters/strkub) | 14:12 |
| ignaziocassano1 | strkub is the name of the cluster | 14:12 |
| strigazi | brtknr: didn't try yet | 14:12 |
| strigazi | ignaziocassano1: magnum --version must be >= 2.9.0 | 14:13 |
| strigazi | or pip freeze | grep magnum | 14:13 |
| ignaziocassano1 | aahhh | 14:13 |
| ignaziocassano1 | but queens gives 2.8.0 | 14:13 |
| ignaziocassano1 | How can I install 2.9.0 | 14:13 |
| ignaziocassano1 | ? | 14:14 |
| ignaziocassano1 | on ubuntu 16 ? | 14:14 |
| brtknr | pip install python-magnumclient==2.9.0? | 14:14 |
| ignaziocassano1 | OK | 14:14 |
| brtknr | pip install python-magnumclient>=2.9.0 -U | 14:16 |
| ignaziocassano1 | Do I need a particular version of kubectl | 14:18 |
| ignaziocassano1 | ? | 14:18 |
| ignaziocassano1 | $(openstack coe cluster config strkub --dir ~/clusters/strkub) | 14:19 |
| ignaziocassano1 | export KUBECONFIG=/root/clusters/strkub/config | 14:19 |
| ignaziocassano1 | kubectl -n kube-system get po | 14:20 |
| ignaziocassano1 | No resources found. | 14:20 |
| ignaziocassano1 | my config under /root/clusters/strkub is: | 14:23 |
| strigazi | ignaziocassano1 works | 14:23 |
| ignaziocassano1 | apiVersion: v1 | 14:23 |
| ignaziocassano1 | clusters: | 14:23 |
| ignaziocassano1 | - cluster: | 14:23 |
| ignaziocassano1 | certificate-authority: /root/clusters/strkub/ca.pem | 14:23 |
| ignaziocassano1 | server: https://10.102.186.107:6443 | 14:23 |
| ignaziocassano1 | name: strkub | 14:23 |
| ignaziocassano1 | contexts: | 14:23 |
| ignaziocassano1 | - context: | 14:23 |
| ignaziocassano1 | cluster: strkub | 14:23 |
| ignaziocassano1 | user: admin | 14:23 |
| ignaziocassano1 | name: default | 14:23 |
| ignaziocassano1 | current-context: default | 14:23 |
| ignaziocassano1 | kind: Config | 14:23 |
| ignaziocassano1 | preferences: {} | 14:23 |
| ignaziocassano1 | users: | 14:23 |
| ignaziocassano1 | - name: admin | 14:23 |
| ignaziocassano1 | user: | 14:23 |
| ignaziocassano1 | client-certificate: /root/clusters/strkub/cert.pem | 14:23 |
| ignaziocassano1 | client-key: /root/clusters/strkub/key.pem | 14:23 |
| strigazi | kubectl get all --all-namespaces | 14:24 |
| ignaziocassano1 | yessss | 14:24 |
| ignaziocassano1 | it works | 14:24 |
| ignaziocassano1 | documentation reposrd the get po should returns somthing | 14:25 |
| ignaziocassano1 | documentation reposrd the get po should returns somsthing | 14:25 |
| ignaziocassano1 | documentation reposrd the get po should returns something | 14:25 |
| ignaziocassano1 | but it does not return anything | 14:25 |
| ignaziocassano1 | kubectl get all --all-namespaces returns a lot of things | 14:26 |
| ignaziocassano1 | Please, two things and sorry I am disturbing a lot | 14:28 |
| ignaziocassano1 | I would be sure it works | 14:28 |
| ignaziocassano1 | I must study kubernetes and swarm | 14:29 |
| ignaziocassano1 | Any axample to deploy a container and check if it works ? | 14:29 |
| ignaziocassano1 | I run kubectl run nginx --image=nginx --replicas=5 | 14:30 |
| ignaziocassano1 | it worked | 14:30 |
| ignaziocassano1 | but now where I get the address where ngunx is deployed ? Is the master node on port 80 ? | 14:31 |
| ignaziocassano1 | 2 thing: modifications we made should be object of a patch ? | 14:36 |
| ignaziocassano1 | Are you memeber ov magnum developement team ? | 14:37 |
| ignaziocassano1 | Are you memeber of magnum developement team ? | 14:37 |
| ignaziocassano1 | Or I must do something to report changes we made for obtaining patches ? I know aother persons who reported same issues | 14:39 |
| strigazi | you can create a story in storyboard.openstack.org for magnum and add an entry here: | 14:40 |
| strigazi | https://wiki.openstack.org/wiki/Meetings/Containers#Agenda_for_2018-12-18_2100_UTC | 14:41 |
| strigazi | yes, I'm the PTL of the magnum project | 14:42 |
| strigazi | you can access a service in your cluster with kubernetes services https://kubernetes.io/docs/concepts/services-networking/service/ | 14:43 |
| *** munimeha1 has joined #openstack-containers | 14:59 | |
| ignaziocassano1 | OK | 15:00 |
| ignaziocassano1 | Firstly I will call my collegue who knows kubernets for testing. I it works fine I will post on storyboard | 15:03 |
| ignaziocassano1 | Firstly I will call my collegue who knows kubernets for testing. If it works fine I will post on storyboard | 15:04 |
| strigazi | ignaziocassano1: cool, in which organization do you work? | 15:07 |
| ignaziocassano1 | I work in Italy : www.csipiemonte.it | 15:07 |
| *** spiette has quit IRC | 15:08 | |
| ignaziocassano1 | The site is also in english . We work for local government departments of Piemonte. Piemonte is a region in north-west of italy | 15:09 |
| ignaziocassano1 | Is there any documentation that explain which ports must be opened between magnum external networks a and controllers ? | 15:10 |
| *** spiette has joined #openstack-containers | 15:10 | |
| ignaziocassano1 | I saw port 5000 for keystone and some ports for heat . It could be useful writing some documentation because many people got errors on this | 15:11 |
| ignaziocassano1 | If you read emailing list operators, Zufar Dhiyaulhaq is becoming crazy for this. I talked with him on yesterday | 15:13 |
| strigazi | https://docs.openstack.org/magnum/latest/install/install.html | 15:13 |
| strigazi | the compute instances must be able to access, keystone, magnum (for kubernetes nova, cinder, octavia/neutron if you want to use the LoadBalancer service type) | 15:15 |
| strigazi | compute instance == the virtual machines that compose the cluster | 15:16 |
| *** hongbin has joined #openstack-containers | 15:17 | |
| ignaziocassano1 | Sorry, I read: | 15:17 |
| ignaziocassano1 | 15:17 | |
| ignaziocassano1 | Important | 15:17 |
| ignaziocassano1 | Magnum creates clusters of compute instances on the Compute service (nova). These instances must have basic Internet connectivity and must be able to reach magnum’s API server. Make sure that the Compute and Network services are configured accordingly. | 15:17 |
| strigazi | The configuration you need to do depends on your deployment ^^ | 15:19 |
| ignaziocassano1 | yes, bu basically compute instances need to access heat for wait condition (for example) | 15:20 |
| strigazi | cluster nodes need to issue token, talk the magnum API, talk to nova (when using the cloud provider), talk to cinder if you want to mount volumes, pull images from docker.io or an internal to the organization registry and heat for the wait handle and software deployments | 15:20 |
| ignaziocassano1 | Now it is more clear | 15:21 |
| ignaziocassano1 | strigazi I have no works to say thank you for your help. You was ver very patient with me and my poor english | 15:24 |
| ignaziocassano1 | strigazi I have no words to say thank you for your help. You was ver very patient with me and my poor english | 15:24 |
| strigazi | you are welcome | 15:25 |
| ignaziocassano1 | where are you from ? | 15:25 |
| strigazi | I work at cern.ch | 15:26 |
| *** ykarel is now known as ykarel|away | 15:26 | |
| ignaziocassano1 | So your skill is impressive | 15:27 |
| ignaziocassano1 | Thank you again. | 15:27 |
| ignaziocassano1 | bye | 15:29 |
| *** pcaruana has quit IRC | 15:29 | |
| *** ignaziocassano1 has quit IRC | 15:30 | |
| strigazi | cheers | 15:30 |
| *** ykarel|away has quit IRC | 15:35 | |
| *** gp-csi has joined #openstack-containers | 15:38 | |
| *** itlinux has quit IRC | 15:40 | |
| *** gp-csi has quit IRC | 15:40 | |
| *** pcaruana has joined #openstack-containers | 15:51 | |
| *** ykarel|away has joined #openstack-containers | 15:52 | |
| *** gp-csi has joined #openstack-containers | 15:56 | |
| *** udesale has quit IRC | 15:57 | |
| *** gp-csi has quit IRC | 15:58 | |
| *** gp-csi has joined #openstack-containers | 16:02 | |
| *** janki has quit IRC | 16:05 | |
| *** ykarel|away is now known as ykarel | 16:12 | |
| *** Bhujay has joined #openstack-containers | 16:16 | |
| *** pcaruana has quit IRC | 16:28 | |
| *** gp-csi has quit IRC | 16:30 | |
| *** itlinux has joined #openstack-containers | 16:40 | |
| colin- | seems you made a big impression strigazi :) | 16:41 |
| strigazi | colin-: xD | 16:45 |
| strigazi | colin-: do you run by any change cinder csi? | 16:46 |
| *** ttsiouts has quit IRC | 16:56 | |
| *** Bhujay has quit IRC | 16:56 | |
| *** ttsiouts has joined #openstack-containers | 16:56 | |
| *** ttsiouts has quit IRC | 17:01 | |
| colin- | i know we have used it, not sure if the template we're deploying today uses it for anything | 17:03 |
| colin- | why do you ask? | 17:03 |
| dims | strigazi : not sure if you are following along this work - https://github.com/kubernetes-sigs/cluster-api-provider-openstack (also see https://github.com/kubernetes-sigs/cluster-api) | 17:11 |
| *** ramishra has quit IRC | 17:12 | |
| strigazi | dims: I do and we intend to use/try it soon. I'm following the development for the aws one too. | 17:16 |
| strigazi | colin-: moving to openstack-ccm is not a drop-in replacement for cinder :( lbaas works great but cinder doesn't | 17:17 |
| colin- | gotcha | 17:19 |
| *** flwang1 has joined #openstack-containers | 17:19 | |
| flwang1 | strigazi: around? | 17:19 |
| strigazi | flwang1: not for long, tell me | 17:20 |
| flwang1 | strigazi: i just saw your message on slack, do you think we should just give up the effort on CPO+Cinder and go for csi? | 17:20 |
| dims | strigazi : lxkong just got it working and will help update some of the doc around ccm + cinder | 17:21 |
| strigazi | flwang1: I'm interested as a developer of magnum. Undortunately as a user we sold to cephfs and manila | 17:22 |
| flwang1 | dims: that's good. i asked lxkong took a look yesterday. but my question is | 17:23 |
| strigazi | flwang1: it looks good, I was comparing it with csi-cephsfs | 17:23 |
| flwang1 | are we going to support it ? | 17:23 |
| dims | flwang1 : yes, we (kubernetes/cloud-provider-openstack folks) are going to support it | 17:24 |
| flwang1 | as strigazi said, if csi is the trend, what's is our goal to maintain it? | 17:24 |
| strigazi | flwang1: that are missing are, pass the openstack_ca, use the trustee, move things to kube-system, everything seems to be runnning in the default ns | 17:24 |
| dims | flwang1 : it's the trend, but it's not ready yet. (and it has to work with things like cephfs and manila like strigazi pointed out) | 17:25 |
| flwang1 | dims: cool, if that's the long term goal. we (catalyst cloud) would like to help | 17:25 |
| dims | great! | 17:25 |
| strigazi | dims: FYI, patch to use the external ccm: https://review.openstack.org/#/c/577477/ | 17:26 |
| strigazi | flwang1: I have to go see you later | 17:27 |
| flwang1 | strigazi: have a good night | 17:27 |
| flwang1 | strigazi: i will test lingxian's fix and add comments on the magnum patch | 17:27 |
| strigazi | flwang1: oh, I left you comment at the patch for k8s-keystone-auth | 17:27 |
| flwang1 | strigazi: great, thanks | 17:28 |
| strigazi | flwang1: we can do a followup patch for csi-cinder if you want | 17:28 |
| strigazi | flwang1: see you | 17:28 |
| flwang1 | dims: did lxkong figure out any fix we need for CPO for cinder? | 17:28 |
| flwang1 | strigazi: re csi-cinder, we should | 17:28 |
| dims | strigazi : ack, i got to step out for a bit will review when i get back. in general the tip missing in the doc was that kube-controller-manager's --external-cloud-volume-plugin should be set to `openstack` and the config file has to be "/etc/kubernetes/cloud-config" (exactly the same path!!) and has to be available to kubelet and kube-controller-manager | 17:28 |
| dims | flwang1 : yes ... ^ | 17:28 |
| dims | (`--cloud-provider` set to external for kubelet/kube-controller-manager and kube-apiserver needless to say) | 17:29 |
| flwang1 | dims: no code change? | 17:29 |
| dims | nope | 17:30 |
| flwang1 | dims: cool | 17:30 |
| dims | flwang1 : we have had the test suites running for a while, just missing docs | 17:30 |
| flwang1 | dims: very nice | 17:30 |
| dims | back in a bit ... | 17:30 |
| flwang1 | dims: ok, i have another question for you after you back | 17:31 |
| flwang1 | dims: as for this issue https://github.com/kubernetes/cloud-provider-openstack/issues/387 just wanna confirm with you about the binary name since Magnum are going to use it for keystone auth | 17:34 |
| *** ykarel is now known as ykarel|away | 17:37 | |
| *** PagliaccisCloud has joined #openstack-containers | 17:50 | |
| *** PagliaccisCloud has quit IRC | 17:55 | |
| dims | flwang1 : replied in issue | 18:02 |
| *** PagliaccisCloud has joined #openstack-containers | 18:03 | |
| flwang1 | dims: thanks. i tested with kubectl 1.8.x which doesn't work with old way, and with kubectl 1.11.x it worked for me with the client-keystone-auth cmd | 18:04 |
| flwang1 | since i'd like to auto generate the kubeconfig for magnum k8s cluster for keystone auth, so i'd like to confirm the correct name of the keystone auth client binary | 18:05 |
| dims | flwang1 : i believe the name will remain, unless the kubectl stops accepting the kubeconfig with the full path of binary under "exec" | 18:06 |
| flwang1 | dims: ok, so the convention is not very strict, right? | 18:06 |
| flwang1 | i'm pretty sure the full name/path does work | 18:07 |
| dims | i belive the convention is for finding things that augment the kubectl command line (that inject new command line params etc) | 18:07 |
| flwang1 | dims: cool, all good. thanks for all the help | 18:08 |
| mordred | ooh look, it's flwang1 and dims | 18:08 |
| * dims waves to mordred | 18:09 | |
| * mordred wavs to dims | 18:09 | |
| *** salmankhan has quit IRC | 18:09 | |
| flwang1 | mordred: waves from New Zealand | 18:09 |
| flwang1 | mordred: did you forget to merge my ansible patch for Magnum? ;) | 18:09 |
| mordred | flwang1, dims: SO - I have a 'fun' edge case I hit with magnum | 18:09 |
| mordred | flwang1: uhoh. probably so, yes | 18:10 |
| mordred | flwang1: have a link handy? | 18:10 |
| flwang1 | mordred: i'm listening... | 18:10 |
| flwang1 | mordred: wait a sec | 18:10 |
| mordred | dims, flwang1: https://github.com/rook/rook/issues/2371 is the issue we filed with rook and has *way* more info than you'd ever want ... | 18:11 |
| mordred | but the tl;dr is | 18:11 |
| mordred | the flexvolume driver from rook that mounts cephfs shares winds up mounting them in the kubelet container and they don't propagate up to the host, so the pods where the filesystem is supposed to be mounted don't actually see the cephfs filesystem | 18:12 |
| flwang1 | https://github.com/ansible/ansible/pull/44686 | 18:12 |
| mordred | even though the /var/lib/kubelet is mounted rshared - and the tempfs filesystems mounted for secrets DO propagate back up to the host | 18:12 |
| flwang1 | i saw there is a last comment asking change 2.7 to 2.8, anything else i need to change? | 18:12 |
| mordred | flwang1: no - I think that's it | 18:13 |
| flwang1 | mordred: ok, i will do it today | 18:13 |
| flwang1 | cheers | 18:13 |
| flwang1 | mordred: as for your rook question, i saw you asked days ago, still haven't got the answer? | 18:14 |
| mordred | no - haven't figured it out or found any workarounds :( | 18:14 |
| dims | mordred : "smells" like a implementation problem in the flex volume provider | 18:14 |
| mordred | dims: yeah - I pushed this up fairly blind: https://github.com/rook/rook/pull/2376 | 18:14 |
| mordred | dims: based on a similar 'smell' feeling - since the tempfs volumes are behaving properly, it seems liek the kubelet can mount things that are seen on the host | 18:15 |
| mordred | dims: but I don't know enough about flexvolume land or the various magics involved - so I keep hoping someone smart will say something magical :) | 18:15 |
| mordred | flwang1: ping me when you have that updated and I'll leave a shipit - sorry for the delay there | 18:18 |
| flwang1 | mordred: will do it today, thanks | 18:20 |
| flwang1 | mordred: flexvolume is out of my knowledge base, can't help, sorry | 18:21 |
| *** corvus has joined #openstack-containers | 18:21 | |
| mordred | kk. I mostly am not sure if it's an atomic thing, a magnum thing, a rook thing, or an unfortunate combo of them | 18:22 |
| dims | mordred : what version of k8s and rook? | 18:23 |
| flwang1 | mordred: are you using master Magnum? | 18:24 |
| mordred | dims: latest of rook - 0.9.0 - and basically just followed their getting started guide: https://www.nytimes.com/2018/12/17/science/donald-knuth-computers-algorithms-programming.html and https://rook.io/docs/rook/v0.9/ceph-filesystem.html | 18:24 |
| mordred | we're using the magnum that's deployed in vexxhost | 18:24 |
| mordred | corvus: remember off the top of your head the k8s version it deployed? | 18:25 |
| *** ykarel|away has quit IRC | 18:25 | |
| corvus | mordred: yes, we used kube_tag to specify it: kube_tag=v1.11.5-1 | 18:26 |
| dims | thanks corvus | 18:26 |
| dims | mordred : i see some updates to the mount code in k8s that rook imports, but rook is still vendoring in 1.11.3 ... | 18:26 |
| mordred | dims: oh goodie - so if it is a rook issue, it could also need an update to the vendored code | 18:27 |
| dims | probably worth going through https://github.com/kubernetes/kubernetes/commits/master/pkg/util/mount/mount.go and related files.. | 18:28 |
| dims | does your patch work? | 18:28 |
| dims | assumption from before seems to be if /var/lib/kubelet has correct magic then things get visible - https://github.com/kubernetes/kubernetes/pull/45724 | 18:29 |
| mordred | dims: I have no clue - I'm not sure how to test it - I'm guessing i'd need to build it, and then put the resulting binary down into the flexvolume paths right? | 18:29 |
| mordred | dims: yeah - and I verified in the magnum containerized kubelet that it's marking /var/lib/kubelet as rshared - so that part *should* be good | 18:30 |
| dims | mordred : since you are using v1.11.5-1 k8s and and k8s in rook in 1.11.3 ... don't think that may be a good idea yet | 18:30 |
| mordred | dims: oh - so that pull is not in 1.11.5? | 18:31 |
| dims | don't see it - https://github.com/kubernetes/kubernetes/compare/v1.11.3...v1.11.5 | 18:33 |
| dims | so ... back in a few hours. need to take young one to dentist | 18:33 |
| mordred | dims: awesome. thanks | 18:33 |
| flwang1 | v1.11.5-1 is the highest version we're publising IIRC | 18:38 |
| flwang1 | just in case, if you need any change of the k8s config, you can just ssh into the vm and change it, restart the k8s component to verify | 18:38 |
| *** flwang1 has quit IRC | 18:48 | |
| *** PagliaccisCloud has quit IRC | 19:44 | |
| *** yolanda has quit IRC | 21:18 | |
| *** rcernin has joined #openstack-containers | 21:28 | |
| *** hongbin has quit IRC | 22:29 | |
| *** itlinux has quit IRC | 22:42 | |
| *** munimeha1 has quit IRC | 23:20 | |
| *** itlinux has joined #openstack-containers | 23:34 | |
| *** absubram has joined #openstack-containers | 23:53 | |
| *** PagliaccisCloud has joined #openstack-containers | 23:57 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!