style95 commented on code in PR #5267:
URL: https://github.com/apache/openwhisk/pull/5267#discussion_r916400121
##########
core/scheduler/src/main/scala/org/apache/openwhisk/core/scheduler/queue/QueueManager.scala:
##########
@@ -397,7 +400,7 @@ class QueueManager(
logging.warn(
this,
s"[${msg.activationId}] the activation message has not been scheduled
for ${queueManagerConfig.maxSchedulingTime.toSeconds} sec")
- completeErrorActivation(msg, "The activation has not been processed")
+ completeErrorActivation(msg, "The activation has not been processed: too
old activation is arrived.")
Review Comment:
Not sure this effectively describes the case.
This is the case that activations arrived after the max scheduling wait time.
For example, this can happen when there was a failure in Kafka so
activations could not be delivered and just stored in it. When Kafka becomes
available again, it will start delivering activations.
But if it took so much time to restore Kafka such as 1 hour, it will send
too old(1 hour-old) activations.
Also, if there were many activations stored in Kafka before the failure, it
would cause a thundering herd by sending them all at the same time. So we
complete them with an error.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
