startMachine not called after previously failed uninstall-application


After several calls to install / uninstall-application, the last install-apllication call did not start any machines.

We are calling add-templates which seems to be successful, calling install-application which seems to be successful, but ESM is complaining about SomeProcessingUnitsHaveNotCompletedStateRecoveryException. No evidence of calling startMachine.

attached logs-analysis.txt which contains my initial findings.
attached full logs of ESM, GSM, REST, etc.
attached also petclinic app and template we are using.


Itai Frenkel
December 9, 2012, 9:28 AM

The ESM logs shows a SEVERE error in the cloud driver ( that caused a bean creation to fail. This could explain why the ESM state is compromised, and why the ESM "disabled" itself as a stop-gap measure.

Itai Frenkel
December 9, 2012, 10:17 AM

The GSM logs also indicate that depdepdep1.mongod was not uninstalled properly.

Itai Frenkel
December 9, 2012, 10:19 AM

The CLI logs also indicate that application depdepdep1 was not uninstalled. That explains the problem in the ESM

Itai Frenkel
December 9, 2012, 10:20 AM

Added JIRA GS-10865 that if fixed, would have allowed the ESM to keep working even when depdepdep1 was not installed succesfuly, and not uninstalled.

Itai Frenkel
December 25, 2012, 9:47 AM

Commit fixed for GS-10865. ESM now will manage other services regardless of the faulty service.
There is still GS-684 open, which means that the faulty service may not be able to uninstall - so it remains faulted.


Itai Frenkel


Meron Avigdor