dasahcc commented on a change in pull request #1231:
URL: https://github.com/apache/helix/pull/1231#discussion_r469467068
##########
File path: helix-core/src/main/java/org/apache/helix/task/JobConfig.java
##########
@@ -147,10 +147,17 @@
StartTime,
/**
- * The expiration time for the job
+ * The expiration time for the job if it's completed; once the expiry is
reached and the job is
+ * completed, the job will be purged
*/
Expiry,
+ /**
+ * The expiration time for the job if it's failed or timed out; once the
expiry is reached and
+ * the job has failed or timed out, the job will be purged
+ */
+ TerminalStateExpiry,
Review comment:
@NealSun96 Even we dont setup the expiry, there is a default value for
expiry is one day. That's why I am saying the default behavior is cleaning up
successful job. That's our assumption.
We need to balance the operability and flexibility. If you really consider
flexibility, we should have cluster level config + job level. Then the logic
will be much complicated.
My suggestion is still having a global change flag instead of to
fine-grained, because we dont see any fine-grained control request from any of
our customers. And from operability perspective, it will be very hard for them
roll out since it requires code level change and redeploy all their scheduler
code. But if we support it from cluster level, it they can dynamically change
it with our latest controller deployment. That will be much simpler.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
