Description of problem: in the container page on the cloudforms web UI, all the containers have the same name but when I run docker ps they have UUID's Version-Release number of selected component (if applicable): cfme-5.6.0.5-beta2.4.1.el7cf.x86_64 docker-1.8.2-7.el7.x86_64 atomic-openshift-sdn-ovs-3.1.1.6-4.git.32.adf8ec9.el7aos.x86_64 How reproducible: 100% Steps to Reproduce: 1. install openshift and connect it to cfme 2. create pods, containers and replicators 3. navigate to compute -> containers -> containers 4. on openshift run: docker ps Actual results: in the UI we see that all the containers have the same name when running docker ps, all the containers have their own UUIDs Expected results: if the containers have their own UUID's we should show a unique ID Additional info: [root@dafna-openshift-master01 ~]# docker ps CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES eb47cdc40572 registry.access.redhat.com/openshift3/jenkins-1-rhel7:latest "/usr/libexec/s2i/run" About an hour ago Up About an hour k8s_jenkins.c7abe241_jenkins-1-cnor8_default_b36b3a1b-12d3-11e6-b635-001a4a169777_ae4969e6 69585f2921ea registry.access.redhat.com/openshift3/jenkins-1-rhel7:latest "/usr/libexec/s2i/run" About an hour ago Up About an hour k8s_jenkins.c7abe241_jenkins-1-9vvva_default_b36ca505-12d3-11e6-b635-001a4a169777_4f7b51e1 c452bd506043 registry.access.redhat.com/openshift3/jenkins-1-rhel7:latest "/usr/libexec/s2i/run" About an hour ago Up About an hour k8s_jenkins.c7abe241_jenkins-1-f3dt2_default_b366a5f5-12d3-11e6-b635-001a4a169777_58ad15ae b0ee4eb040ea registry.access.redhat.com/openshift3/jenkins-1-rhel7:latest "/usr/libexec/s2i/run" About an hour ago Up About an hour k8s_jenkins.c7abe241_jenkins-1-2e5b3_default_b35367b6-12d3-11e6-b635-001a4a169777_7efa9515 e860c7c630d4 registry.access.redhat.com/openshift3/jenkins-1-rhel7:latest "/usr/libexec/s2i/run" About an hour ago Up About an hour k8s_jenkins.c7abe241_jenkins-1-fv6v3_default_b3947a1f-12d3-11e6-b635-001a4a169777_e0ef8865 e2b8a0a6f52e registry.access.redhat.com/openshift3/jenkins-1-rhel7:latest "/usr/libexec/s2i/run" About an hour ago Up About an hour k8s_jenkins.c7abe241_jenkins-1-mj1d7_default_b3769ffc-12d3-11e6-b635-001a4a169777_12387023 4928f00b3bf9 registry.access.redhat.com/openshift3/jenkins-1-rhel7:latest "/usr/libexec/s2i/run" About an hour ago Up About an hour k8s_jenkins.c7abe241_jenkins-1-wn2d7_default_b3934b2c-12d3-11e6-b635-001a4a169777_b94bb8f2 8db9ca5d6bfa registry.access.redhat.com/openshift3/jenkins-1-rhel7:latest "/usr/libexec/s2i/run" About an hour ago Up About an hour k8s_jenkins.c7abe241_jenkins-1-0vmv0_default_b395e731-12d3-11e6-b635-001a4a169777_35b20bae 7e10a8dae8e2 registry.access.redhat.com/openshift3/jenkins-1-rhel7:latest "/usr/libexec/s2i/run" About an hour ago Up About an hour k8s_jenkins.c7abe241_jenkins-1-sxsim_default_bac89a23-12c9-11e6-b635-001a4a169777_863937a9 8d3434a0d6d9 registry.access.redhat.com/openshift3/jenkins-1-rhel7:latest "/usr/libexec/s2i/run" About an hour ago Up About an hour k8s_jenkins.c7abe241_jenkins-1-b37gw_default_b31bfc7c-12d3-11e6-b635-001a4a169777_b1e31834 2cc1ac237850 openshift3/ose-haproxy-router:v3.1.1.6 "/usr/bin/openshift-r" About an hour ago Up About an hour k8s_router.ca73e799_router-1-5dqt3_default_ceb5c3b1-12c3-11e6-8b67-001a4a169777_3e901c38 001867477d8d openshift3/ose-pod:v3.1.1.6 "/pod" About an hour ago Up About an hour k8s_POD.e071dbf6_jenkins-1-0vmv0_default_b395e731-12d3-11e6-b635-001a4a169777_b16fb7a5 fadbcb4333f8 openshift3/ose-pod:v3.1.1.6 "/pod" About an hour ago Up About an hour k8s_POD.e071dbf6_jenkins-1-wn2d7_default_b3934b2c-12d3-11e6-b635-001a4a169777_e2f7b291 aa1015e3efc4 openshift3/ose-pod:v3.1.1.6 "/pod" About an hour ago Up About an hour k8s_POD.e071dbf6_jenkins-1-9vvva_default_b36ca505-12d3-11e6-b635-001a4a169777_71a12058 10d7a126c0c1 openshift3/ose-pod:v3.1.1.6 "/pod" About an hour ago Up About an hour k8s_POD.e071dbf6_jenkins-1-mj1d7_default_b3769ffc-12d3-11e6-b635-001a4a169777_a1295d6b 0a5c4536d9e9 openshift3/ose-pod:v3.1.1.6 "/pod" About an hour ago Up About an hour k8s_POD.e071dbf6_jenkins-1-cnor8_default_b36b3a1b-12d3-11e6-b635-001a4a169777_8dacaf4e f2a7ad1f3ef5 openshift3/ose-pod:v3.1.1.6 "/pod" About an hour ago Up About an hour k8s_POD.e071dbf6_jenkins-1-fv6v3_default_b3947a1f-12d3-11e6-b635-001a4a169777_6cd74645 95a476ac6a9d openshift3/ose-pod:v3.1.1.6 "/pod" About an hour ago Up About an hour k8s_POD.e071dbf6_jenkins-1-2e5b3_default_b35367b6-12d3-11e6-b635-001a4a169777_7b07dead a689a2e1c13b openshift3/ose-pod:v3.1.1.6 "/pod" About an hour ago Up About an hour k8s_POD.e071dbf6_jenkins-1-f3dt2_default_b366a5f5-12d3-11e6-b635-001a4a169777_3d7ba90c d6efc228d382 openshift3/ose-pod:v3.1.1.6 "/pod" About an hour ago Up About an hour k8s_POD.e071dbf6_router-1-5dqt3_default_ceb5c3b1-12c3-11e6-8b67-001a4a169777_bfc19944 ec84cac682a6 openshift3/ose-pod:v3.1.1.6 "/pod" About an hour ago Up About an hour k8s_POD.e071dbf6_jenkins-1-b37gw_default_b31bfc7c-12d3-11e6-b635-001a4a169777_c092a4fc b43ebf9289d8 openshift3/ose-pod:v3.1.1.6 "/pod" About an hour ago Up About an hour k8s_POD.e071dbf6_jenkins-1-sxsim_default_bac89a23-12c9-11e6-b635-001a4a169777_e17ec25a [root@dafna-openshift-master01 ~]#
Dafna, a note for the future: for us docker is an implementation detail (it could be any other container runtime). The API of reference is kubernetes (e.g. you should have reported "oc get pods --all-namespaces -o yaml"). For this specific issue I find that UI-wise "jenkins" is more descriptive than "b36b3a1b-12d3-11e6-b635-001a4a169777". If you want to look at the UUID you can just click on the single event and you'll have all the relevant information (event type, relevant pod, etc.). If you want to augment the information you see at first glance (e.g. uuids/types, etc.) that will reduce the number of items and scalability of the timeline UI.
(In reply to Federico Simoncelli from comment #1) > Dafna, a note for the future: for us docker is an implementation detail (it > could be any other container runtime). > The API of reference is kubernetes (e.g. you should have reported "oc get > pods --all-namespaces -o yaml"). Fair enough. > > For this specific issue I find that UI-wise "jenkins" is more descriptive > than "b36b3a1b-12d3-11e6-b635-001a4a169777". I agree, I do not think that you need to replace the name with the UUID what I want is that we show the name of the container and not just "jenkins" i.e: name: jenkins-1-0vmv0 If you create 10 containers, they get the general name + ID (as above) what you guys show is just the name (i.e name: jenkins-1-0vmv0 is cut into "jenkins" for all containers) > > If you want to look at the UUID you can just click on the single event and > you'll have all the relevant information (event type, relevant pod, etc.). clicking all of them when you are looking for a specific event is not user friendly > > If you want to augment the information you see at first glance (e.g. > uuids/types, etc.) that will reduce the number of items and scalability of > the timeline UI. I don't think the entire UUID should be shown, just the actual reported name (i.e name: jenkins-1-0vmv0). Also, the timeline is not scalable as is now any way... if you create 100 pods you do not see all events
(In reply to Dafna Ron from comment #2) > what I want is that we show the name of the container and not just "jenkins" > i.e: name: jenkins-1-0vmv0 > > If you create 10 containers, they get the general name + ID (as above) what > you guys show is just the name (i.e name: jenkins-1-0vmv0 is cut into > "jenkins" for all containers) Those (jenkins-1-0vmv0) are Pods names, not Container names. Remember that each Pod can contain multiple Containers (identified by a name). We can go with an approach of pod-name/container-name, e.g.: jenkins-1-0vmv0/jenkins but I feel that it would be truncated by the UI anyway. > I don't think the entire UUID should be shown, just the actual reported name > (i.e name: jenkins-1-0vmv0). > > Also, the timeline is not scalable as is now any way... if you create 100 > pods you do not see all events That's handled in another RFE with Dan Clarizio IIRC.
Dafna, please attach logs and screenshots.
Created attachment 1157148 [details] screen shot here is a screen shot as you can see the containers have the same name and the pods have pod_name<uuid>
Created attachment 1157162 [details] logs
(In reply to Dafna Ron from comment #6) > Created attachment 1157148 [details] > screen shot > > here is a screen shot > as you can see the containers have the same name and the pods have > pod_name<uuid> OK, personally I'd go with "<pod-name>/<container-name>" although that may not be unique enough. Another option would be "<container-name>-<short-pod-uuid>" which is less descriptive but more unique. I'd leave the docker id out of the question as it's not an API that we should expose. Also remember that the names may be subject to truncation (not in scope here).
> OK, personally I'd go with "<pod-name>/<container-name>" although that may not be unique enough. > Another option would be "<container-name>-<short-pod-uuid>" which is less descriptive but more unique. We already have a column for pod-name, what new information we get using one of this two options ?
Created attachment 1160555 [details] option a: <pod-name>/<container-name>
Created attachment 1160556 [details] option b: <container-name> - <short-uuid>
Submited upstream: https://github.com/ManageIQ/manageiq/pull/8889 Attached images of two options: option a: "<pod-name>/<container-name>" option b: "<container-name>-<short-pod-uuid>" Dafna: please review and comment which option is best, a third option we did not think of is also welcome :-)
we possibly cannot do it no because of the way the timeline is, but when we discuss changes perhaps we should suggest it. perhaps we use the pod name but we have a + next to it to expend it and see the containers? than we can use: <pod name> <short-pod-uuid> -<container-name><short uui> -<container-name><short uui> -<container-name><short uui> for now, my preference would actually be for <container-name>-<short-pod-uuid> (since it's the container's tab we expect the info on the container and we can trace it to pod) Hope it helped and not complicate it :)
> for now, my preference would actually be for <container-name>-<short-pod-uuid> Dafna it's fine :-) Updated: https://github.com/ManageIQ/manageiq/pull/8889
> perhaps we use the pod name but we have a + next to it to expend it and see the containers? Updated: https://github.com/ManageIQ/manageiq/pull/8889 It now use Pod folders, and when open show the containers.
New commit detected on ManageIQ/manageiq/master: https://github.com/ManageIQ/manageiq/commit/51ac5e0d61fc6cbc02eb67fa203c18437c070dd7 commit 51ac5e0d61fc6cbc02eb67fa203c18437c070dd7 Author: Yaacov Zamir <yzamir> AuthorDate: Mon May 23 15:00:27 2016 +0300 Commit: Yaacov Zamir <yzamir> CommitDate: Thu May 26 20:14:01 2016 +0300 Display unique container names in containers explorer tree view. Currently all Containers from all Pods are displayed in one list, this is bad because Containers from different pods can have the same name. This commit sparate each container group by it's pod, and user can recognize each container by name+pod. Bugzilla reference: https://bugzilla.redhat.com/show_bug.cgi?id=1333520 app/presenters/tree_builder_containers.rb | 25 ++++++++++++++++++++++--- spec/presenters/tree_builder_containers_spec.rb | 13 +++++++++---- 2 files changed, 31 insertions(+), 7 deletions(-)
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2016:1348