Github user gtenev commented on the pull request:
https://github.com/apache/trafficserver/pull/496#issuecomment-192317642
Sure, will remove the `malloc` failure check. We could increase the limit
much higher or remove it altogether.
With the previous code in the really misconfigured corner case described in
TS-4161 in my environment the traffic_manager pumped non-stop 6628 messages
1986000B total for 1.85s until it overflowed the stack and crashed.
It kept enqueuing messages but never got a chance to process them. Since
the current implementation has no flow control it seemed like a good idea to
put a limit to give a chance to the processing to catch up (if possible) or to
at least pause the ever-growing allocation.
But this was a misconfigured corner case, wonder how the "heaviest normal"
usage can be measured to set a reasonable limit or if I should remove it
altogether (open to suggestions).
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
