[UIMA-4321] DUCC should not retry JPs forever when the JP framework fails

• Type: Bug
• Status: Closed
• Priority: Blocker
• Resolution: Fixed
• Fix Version/s: 2.0.0-Ducc
• Component/s: DUCC
• Assignee: Jaroslaw Cwiklik
• Reporter: Burn L. Lewis
• Created: 01/Apr/15 18:19
• Updated: 23/Jul/15 17:29
• Resolved: 23/Jul/15 17:29

Description

Job 235131 had a large string of JPs fail (when the JD OOM'd) with:
HttpWorkerThread.run() I/O exception (org.apache.commons.httpclient.NoHttpResponseException) caught when processing request: The server 192.168.3.77 failed to respond

For the short-term we should count this as a Croak (i.e. an unexpected termination that DUCC didn't request), even though it is not caused by user error, so that the users's process_failures_limit can eventually end the job.
Perhaps we need a "framework_failures_limit" in ducc.properties for errors caught in the ducc-side JP code as opposed to errors caught in user code.