[jbosstools-issues] [JBoss JIRA] (JBIDE-22609) Provided wrong credentials, CDK server adapter says "started" even though it is not.

[Rob Stryker (JIRA)]

    [ https://issues.jboss.org/browse/JBIDE-22609?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13263882#comment-13263882 ] 

Rob Stryker edited comment on JBIDE-22609 at 7/11/16 1:48 PM:
--------------------------------------------------------------

I'm (sadly) going to close the issue, because the following options all don't work:

1) We don't parse output, and parsing output is unreliable anyway, since output may change

2) We cannot use the exit code of the vagrant up command. If we did, our server adapter would be marked as 'stopped' when vagrant and the various vm providers will insist the vm is "running".  This will cause even more problems. 

3) Hard-coding a check against our expected IP won't work, because the user may have actually intentionally changed their IP address for their own use case. 

Basically, our only option here is to update the error messaging and let the user figure it out.  I've updated messaging as follows:

{code}
     String msg = NLS.bind("The CDK VM is up and running, but OpenShift is unreachable at url {0}. " + 
     "The VM may not have been registered successfully. Please check your console output for more information", url);
{code}


was (Author: rob.stryker):
I'm (sadly) going to close the issue, because the following options all don't work:

1) We don't parse output, and parsing output is unreliable anyway, since output may change

2) We cannot use the exit code of the vagrant up command. If we did, our server adapter would be marked as 'stopped' when vagrant and the various vm providers will insist the vm is "running".  This will cause even more problems. 

3) Hard-coding a check against our expected IP won't work, because the user may have actually intentionally changed their IP address for their own use case. 

Basically, our only option here is to update the error messaging and let the user figure it out.  I've updated messaging as follows:

+       String msg = NLS.bind("The CDK VM is up and running, but OpenShift is unreachable at url {0}. " + 
+       "The VM may not have been registered successfully. Please check your console output for more information", url);


> Provided wrong credentials, CDK server adapter says "started" even though it is not.
> ------------------------------------------------------------------------------------
>
>                 Key: JBIDE-22609
>                 URL: https://issues.jboss.org/browse/JBIDE-22609
>             Project: Tools (JBoss Tools)
>          Issue Type: Bug
>          Components: cdk
>    Affects Versions: 4.4.0.Final
>         Environment: Devstudio 10.0.0.GA (B33), CDK 2.1.RC4, Windows 10
>            Reporter: Radim Hopp
>            Assignee: Rob Stryker
>            Priority: Critical
>             Fix For: 4.4.1.AM1
>
>
> When CDK startup fails when it is provided wrong credentials (it fails on vagrant-registration), CDK server adapter sets its state to "started" and OS3 connection is created (even though OS3 is not running in vagrant) with some weird IP (10.0.2.15). Also error pops up saying "The CDK VM is up and running, but OpenShift is unreachable at url https://10.0.2.15:8443/healthz/ready". I really don't know where this IP comes from. I was able to reproduce it on Fedora using libvirt box and the IP was something like (192.168.x.x)



--
This message was sent by Atlassian JIRA
(v6.4.11#64026)