anilgursel commented on pull request #371:
URL: https://github.com/apache/tomcat/pull/371#issuecomment-718799285
I updated the PR to have it as `getMaxConnections() / 2`. I agree it is
just a heuristic and not clean. We can potentially make it configurable?
If my understanding is correct, one reason to schedule the `accept` is to
prevent blocking the thread. The reason behind blocking (via `LimitLatch`) is
to have a hard limit on the `maxConnections`. Is there any possibility to
relax this a bit? Something like:
```java
else if (getConnectionCount() < getMaxConnections()) {
increment();
serverSock.accept(null, this);
}
```
With this, there is possibility of number of connections to go over
`maxConnections` little bit (which I think the expectation can be managed in
documentation). Actually, I believe there will only be one thread doing
`accept` logic at any given time. Because, a new `accept` is scheduled only
when the previous one is `completed`. Is this true? If so, it would not go
beyond maxConnections.
----------------------------------------------------------------
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:
us...@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org
For additional commands, e-mail: dev-h...@tomcat.apache.org
