I just figured it out. I found a document that explains just terms.
ThreadsPerChild is the STARTING number of threads per process.
ThreadLimit is the MAX threads that should be created by a child process.
So when calculating ServerLimit you should always use the worst case formula.
ServerLimt = MaxRequestWorkers divided by ThreadLimit .
ThreadLimit could equal ThreadsPerChild. Or ThreadLimit could be greater than ThreadsPerChild.
Thanks.
On 8/24/18, 2:46 PM, "Eric Covener" <covener@xxxxxxxxx> wrote:
On Fri, Aug 24, 2018 at 3:42 PM Campbell, Lance <lance@xxxxxxxxxxxx> wrote:
>
> Apache 2.4
>
> MPM Worker
>
> The relationship between ThreadLimit and ThreadsPerChild is not very clear. These are some questions I have.
>
>
>
> Why would you ever want ThreadLimit bigger than ThreadsPerChild?
Obscure: So you can raise ThreadsPerChild during a graceful restart
> If you do want it bigger then how do you determine how much bigger it should be?
Just by how much you'd do the above.
> If ThreadLimit is bigger than ThreadsPerChild should you then use ThreadLimit instead of ThreadsPerChild to calculate the ServerLimit?
Yes, but in practice you just don't need a ThreadLimit higher than
ThreadsPerChild.
-/-
I would say decide on MaxRequestWorkers first. Then decide on the mix
between threads and processes you want (ThreadsPerChild) which then
dictates both ThreadLimit and ServerLimit.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@xxxxxxxxxxxxxxxx
For additional commands, e-mail: users-help@xxxxxxxxxxxxxxxx
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@xxxxxxxxxxxxxxxx
For additional commands, e-mail: users-help@xxxxxxxxxxxxxxxx
|