[REEF-1343] Fix events received in case of evaluator failure #961
Conversation
This addressed the issue by * Invoke RuntimeStop on Exception in RuntimeClock. * Fix handling of Exceptions in EvaluatorRuntime. JIRA: [REEF-1343](https://issues.apache.org/jira/browse/REEF-1343)
{ | ||
lock (_heartBeatManager) | ||
{ | ||
Logger.Log(Level.Error, string.Format(CultureInfo.InvariantCulture, "evaluator {0} failed with exception", _evaluatorId), e); | ||
Logger.Log(Level.Error, "evaluator {0} failed with exception {1}.", _evaluatorId, e); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
evaluator
-> Evaluator
Done with a first pass, mostly questions :) |
@tcNickolas would be great if you can have a pass as well. |
|
||
if (_state == State.RUNNING) | ||
{ | ||
_state = State.DONE; | ||
_heartBeatManager.OnNext(); | ||
OnException(runtimeStop.Exception); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
note to self: check existence of Exception
.
Could you please describe the root cause of the issue in JIRA? Currently JIRA has only a description of symptoms, not the cause. |
@markusweimer Addressed your comments. |
@tcNickolas sure. |
@tcNickolas updated the JIRA description with the cause. Thanks! |
There are several exceptions logged which I'm a bit concerned about.
|
@tcNickolas both
|
I see. We probably need to add a From test behavior point of view, this change looks good to me. |
} | ||
catch (Exception e) | ||
else |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Do we need to check if the sate is already DONE or any other state that might reach here?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Any other State
other than INIT
and RUNNING
are essentially DONE
.
@jwang98052 @markusweimer any other comments? |
LGTM, will test and merge. |
This addressed the issue by
JIRA:
REEF-1343