if that's a sync query, there is a get|post to create the "job" and a redirect to a url that runs it; the http GET after the redirect runs the job (QUEUED -> EXECUTING) and if that times out you can't retry it because it already started (and you lost your connection to the output stream)... so that error is caused by retrying the GET (not supported)
Retry should start from the top by issuing another query.
Observed error on long queries: WARNING:RetrySession:Resending request in 30s ... ERROR:: unexpected exception: java.lang.RuntimeException: ca.nrc.cadc.uws.server.JobPhaseException: cannot execute job i6yog5u4yga33kl7 when phase = EXECUTING
if that's a sync query, there is a get|post to create the "job" and a redirect to a url that runs it; the http GET after the redirect runs the job (QUEUED -> EXECUTING) and if that times out you can't retry it because it already started (and you lost your connection to the output stream)... so that error is caused by retrying the GET (not supported)
Retry should start from the top by issuing another query.
Observed error on long queries:
WARNING:RetrySession:Resending request in 30s ... ERROR:: unexpected exception: java.lang.RuntimeException: ca.nrc.cadc.uws.server.JobPhaseException: cannot execute job i6yog5u4yga33kl7 when phase = EXECUTING