Hi Akila
On Thu, Oct 8, 2015 at 2:51 PM, Akila Ravihansa Perera <raviha...@wso2.com>
wrote:
> Hi Thanuja,
>
> I was referring to the build fail in [1]. Please see line starting from:
>
> 2015-10-02 07:28:32 INFO
> {org.wso2.carbon.automation.extensions.servers.utils.ServerLogReader}:70 -
> [2015-10-02 07:28:32,834] ERROR
> {org.apache.stratos.autoscaler.rule.RuleTasksDelegator} - Cannot terminate
> instance
>
> I think this is a separate issue that occurs very rarely. As you noticed,
> NPE in GroupMonitor occurs when undeploying applications. The problem is
> group monitors are retrieving the status from other group monitor objects.
>
> @Reka: shall we change that logic to read the application status from
> topology or ApplicationHolder instead of reading from monitors?
>
We can't read the Application status from Topology as Application will not
contain any status rather only application instance will contain the
status. By accessing application instance, we can't identify whether
undeployment or force undeployment is in-progress. Hence we will have to
maintain these flags to ApplicationMonitor to identity the undeployment or
force undeployment. All other status are actually referring from the
Topology.
Thanks,
Reka
>
> [1] https://builds.apache.org/job/Stratos-41x-Commit-Build/68/consoleFull
>
> Thanks.
>
--
Reka Thirunavukkarasu
Senior Software Engineer,
WSO2, Inc.:http://wso2.com,
Mobile: +94776442007
