if (reason instanceof CancellationException) {
return;
}
// if task is done, ignore the error
TaskInfo taskInfo = getTaskInfo();
if (taskInfo.getState().isDone()) {
return;
}
// log failure message
if (isSocketError(reason)) {
// don't print a stack for a socket error
log.warn("Error updating task %s: %s: %s", taskInfo.getTaskId(), reason.getMessage(), taskInfo.getSelf());
}
else {
log.warn(reason, "Error updating task %s: %s", taskInfo.getTaskId(), taskInfo.getSelf());
}
// remember the first 10 errors
if (errorsSinceLastSuccess.size() < 10) {
errorsSinceLastSuccess.add(reason);
}
// fail the task, if we have more than X failures in a row and more than Y seconds have passed since the last request
long errorCount = this.errorCount.incrementAndGet();
Duration timeSinceLastSuccess = Duration.nanosSince(lastSuccessfulRequest.get());
if (errorCount > maxConsecutiveErrorCount && timeSinceLastSuccess.compareTo(minErrorDuration) > 0) {
// it is weird to mark the task failed locally and then cancel the remote task, but there is no way to tell a remote task that it is failed
PrestoException exception = new PrestoException(TOO_MANY_REQUESTS_FAILED.toErrorCode(),
format("Encountered too many errors talking to a worker node. The node may have crashed or be under too much load. This is probably a transient issue, so please retry your query in a few minutes (%s - %s failures, time since last success %s)",
taskInfo.getSelf(),
errorCount,
timeSinceLastSuccess.convertToMostSuccinctTimeUnit()));
for (Throwable error : errorsSinceLastSuccess) {
exception.addSuppressed(error);
}