[jbosstools-issues] [JBoss JIRA] (JBIDE-22803) Explorer: When OS projects are created and deleted, seems Openshift explorer restores non existing values
Dmitrii Bocharov (JIRA)
issues at jboss.org
Thu Oct 6 08:38:00 EDT 2016
[ https://issues.jboss.org/browse/JBIDE-22803?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13303428#comment-13303428 ]
Dmitrii Bocharov edited comment on JBIDE-22803 at 10/6/16 8:37 AM:
-------------------------------------------------------------------
[~adietish] tested on https://github.com/openshift/origin/releases/tag/v1.3.0 - everything is fine and i could get /openshift/version.
tested in OS v3.2.1.9-1-g2265530 with openshift-dev/devel user: User "openshift-dev" cannot "get" on "/version/openshift"
Tried to watch projects on the same v3.2.1.9-1-g2265530 which doesn't support the feature. What happens:
1) {code}WatchClient.WatchEndpoint.onFailure(IOException, Response) is called with:
_err.getClass()_ = "class java.net.ProtocolException"
_err.getMessage()_ = "Expected HTTP 101 response but was '200 OK'"
_response_ = "Response{protocol=http/1.1, code=200, message=OK, url=https://10.1.2.2:8443/oapi/v1/projects?watch=true&}"
_response.request()_ = Request{method=GET, url=https://10.1.2.2:8443/oapi/v1/projects?watch=true&, tag=Request{method=GET, url=https://10.1.2.2:8443/oapi/v1/projects?watch=true&, tag=null}}{code}
2) This onFailure calls listener.error(
3) WatchManager.WatchListener.error(Throwable) is called, which restart()s the request and that's why we repeatedly catch these 3 steps.
The first question is why is WatchClient.WatchEndpoint.onFailure is called. The status is OK [with code 200] and it isn't a Failure. WatchClient.WatchEndpoint.onFailure is an implementation of okhttp3.ws.WebSocketListener and i think that it's called by OS server. So i think it's a questions for OS devs why is it failure with non-failure status.
[~jcantrill] do you have any thoughts on this?
[~jcantrill] one question is about your comment: // TODO enhance fix to only check project once in the restart() method of Watch. What are the cases where watching needs to get restarted?
was (Author: dbocharov):
[~adietish] tested on https://github.com/openshift/origin/releases/tag/v1.3.0 - everything is fine and i could get /openshift/version.
tested in OS v3.2.1.9-1-g2265530 with openshift-dev/devel user: User "openshift-dev" cannot "get" on "/version/openshift"
Tried to watch projects on the same v3.2.1.9-1-g2265530 which doesn't support the feature. What happens:
1) WatchClient.WatchEndpoint.onFailure(IOException, Response) is called with:
{quote}_err.getClass()_ = "class java.net.ProtocolException"
_err.getMessage()_ = "Expected HTTP 101 response but was '200 OK'"
_response_ = "Response{protocol=http/1.1, code=200, message=OK, url=https://10.1.2.2:8443/oapi/v1/projects?watch=true&}"
_response.request()_ = Request{method=GET, url=https://10.1.2.2:8443/oapi/v1/projects?watch=true&, tag=Request{method=GET, url=https://10.1.2.2:8443/oapi/v1/projects?watch=true&, tag=null}}{quote}
2) This onFailure calls listener.error(
3) WatchManager.WatchListener.error(Throwable) is called, which restart()s the request and that's why we repeatedly catch these 3 steps.
The first question is why is WatchClient.WatchEndpoint.onFailure is called. The status is OK [with code 200] and it isn't a Failure. WatchClient.WatchEndpoint.onFailure is an implementation of okhttp3.ws.WebSocketListener and i think that it's called by OS server. So i think it's a questions for OS devs why is it failure with non-failure status.
[~jcantrill] do you have any thoughts on this?
[~jcantrill] one question is about your comment: // TODO enhance fix to only check project once in the restart() method of Watch. What are the cases where watching needs to get restarted?
> Explorer: When OS projects are created and deleted, seems Openshift explorer restores non existing values
> ---------------------------------------------------------------------------------------------------------
>
> Key: JBIDE-22803
> URL: https://issues.jboss.org/browse/JBIDE-22803
> Project: Tools (JBoss Tools)
> Issue Type: Bug
> Components: openshift
> Affects Versions: 4.4.1.AM2
> Reporter: Jeff MAURY
> Assignee: Dmitrii Bocharov
> Labels: explorer, openshift, openshift_v3
> Fix For: 4.4.2.AM2
>
> Attachments: after application creation.png, after build finished.png, Openshift Web Console.png, screenshot-1.png
>
>
> EXEC: create an Openshift project
> EXEC: expand it
> EXEC: Create an application in this project (nodejs-example)
> ASSERT: wait for the pod to be available
> EXEC: delete the Openshit project
> EXEC: create an Openshift project (using the same name. You may have to repeat this step as you may got error that it still exists)
> EXEC: expand it
> EXEC: Create an application in this project (nodejs-example)
> EXEC: expand the nodejs-example node
> ASSERT: you should see 2 pods the old one and the build pod you just created: [^screenshot-1.png]
--
This message was sent by Atlassian JIRA
(v6.4.11#64026)
More information about the jbosstools-issues
mailing list