# Max clients != worker processes \* worker connections

**URL:** <https://www.ruby-forum.com/t/max-clients-worker-processes-worker-connections/212711>\
**Category:** NGINX\
**Created:** [October 11, 2011, 5:02pm UTC](https://www.ruby-forum.com/t/max-clients-worker-processes-worker-connections/212711 "2011-10-11T17:02:52Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![dubstep](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/d/cdc98d/32.png) [@dubstep](https://www.ruby-forum.com/u/dubstep)\
**Post date:** [October 11, 2011, 5:02pm UTC](https://www.ruby-forum.com/t/max-clients-worker-processes-worker-connections/212711/1 "2011-10-11T17:02:52Z")

</div>

Hello,

I relise that it may have been hammered on quite a few times already,  
but I cannot understand the results I’m getting. I have a rather basic  
setup on Ubuntu Server 11.04, where nginx (run as root) spawns 2  
worker\_processes and serves a basic HTML page. Each of these processes  
should have its worker\_connections set to 8192. There’s no limit\_zone  
defined. worker\_rlimit\_nofile is set to 65536, keepalive\_timeout 5.

Verifying max load, I run ab from another server on the subnet. It works  
fine with “ab -kc 8000 -n 16000 [http://10.1.1.10/](http://10.1.1.10/)”. However, when I do  
“ab -kc 10000 -n 16000 [http://10.1.1.10/](http://10.1.1.10/)”, ab shows about 3486 failed  
requests in its results (Length: 1743, Exceptions: 1743), while nginx’s  
error.log features numerous “2011/10/11 16:49:24 [alert] 12081#0: 8192  
worker\_connections are not enough” errors.

Testing a number of settings, it seems that there’s a close to 1-1  
relationship between worker\_connections and the maximum concurrency  
parameter to ab that doesn’t produce errors. I tried setting  
worker\_processes to some high number (like 16), but it seems to have no  
effect whatsoever.

Can you please let me know why this setup might not be serving the  
“promised” 😉 worker\_processes \* worker\_connections connections? Is it  
possible that new connections are not being evenly distributed to the  
two processes? Apologies if this is some basic error on our side, we’re  
still learing (and admiring) nginx and are more used to IIS!

Kind regards,  
Dawid Ciecierski

Posted at Nginx Forum:

> **[max\_clients != worker\_processes \* worker\_connections](https://forum.nginx.org/read.php?2,216517,216517#msg-216517)**
>
> Hello, I relise that it may have been hammered on quite a few times already, but I cannot understand the results I'm getting. I have a rather basic setup on Ubuntu Server 11.04, where nginx (run as root) spawns 2 worker\_processes and serves a basic...

---

<div class="post-metadata">

**Author:** ![davidcie](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/d/b3f665/32.png) [@davidcie](https://www.ruby-forum.com/u/davidcie)\
**Post date:** [October 11, 2011, 5:38pm UTC](https://www.ruby-forum.com/t/max-clients-worker-processes-worker-connections/212711/2 "2011-10-11T17:38:43Z")

</div>

Hello!

On Tue, Oct 11, 2011 at 11:02:34AM -0400, davidcie wrote:

> fine with “ab -kc 8000 -n 16000 [http://10.1.1.10/](http://10.1.1.10/)”. However, when I do
> 
> Can you please let me know why this setup might not be serving the  
> “promised” 😉 worker\_processes \* worker\_connections connections? Is it  
> possible that new connections are not being evenly distributed to the  
> two processes? Apologies if this is some basic error on our side, we’re  
> still learing (and admiring) nginx and are more used to IIS!

With small html page it’s very likely that one worker process will  
be able to hold accept mutex exclusively. Upon aproaching  
worker\_connections limit it will stop using accept mutex, but it  
may take some time (up to accept\_mutex\_delay, default 500ms) for  
other workers to come into play. Additionally, the worker will  
still try to accept() connections until it sees other worker  
locked accept mutex (and this may take some additional time).

With real load you should get much better distribution of client  
connections between workers. In extreme use cases like the above  
you may try using

```
events {
    accept_mutex off;
    ...
}

```

to achieve better connection distribution between workers.

This implies some CPU overhead though, especially when using many  
worker processes (OS will wake up each worker on any new  
connection instead of only one holding accept mutex).

Maxim D.

---

<div class="post-metadata">

**Author:** ![davidcie](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/d/b3f665/32.png) [@davidcie](https://www.ruby-forum.com/u/davidcie)\
**Post date:** [October 11, 2011, 5:39pm UTC](https://www.ruby-forum.com/t/max-clients-worker-processes-worker-connections/212711/3 "2011-10-11T17:39:57Z")

</div>

On 10/11/11, davidcie [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid) wrote:

> Can you please let me know why this setup might not be serving the  
> “promised” 😉 worker\_processes \* worker\_connections connections?

> Is it possible that new connections are not being evenly distributed to the  
> two processes?

They aren’t supposed to be evenly distributed. Whoever gets cpu time  
first is likely to accept as many connections as possible. Unless you  
have blocking or cpu intensive requests in which case it won’t have  
enough time to call another accept but other worker will.  
And it’s actually not specific to nginx. I think any non-blocking  
server behaves like this.  
Although you can try setting accept\_mutex or something to influence  
this.

---

<div class="post-metadata">

**Author:** ![davidcie](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/d/b3f665/32.png) [@davidcie](https://www.ruby-forum.com/u/davidcie)\
**Post date:** [October 12, 2011, 10:40am UTC](https://www.ruby-forum.com/t/max-clients-worker-processes-worker-connections/212711/4 "2011-10-12T10:40:16Z")

</div>

This is an excellent explanation - thank you for such an in-depth look  
into the inner workings of nginx. I will consider sligthly decreasing  
accept\_mutex\_delay for our high traffic site, and stop worrying about  
benchmark results too much.

Thanks again, “problem” solved 🙂

Best regards,  
Dawd Ciecierski

Posted at Nginx Forum:

> **[Re: max clients != worker processes \* worker connections](https://forum.nginx.org/read.php?2,216520,216566#msg-216566)**
>
> Hello! On Tue, Oct 11, 2011 at 11:02:34AM -0400, davidcie wrote: \> Hello, \> \> I relise that it may have been hammered on quite a few times already, \> but I cannot understand the results I'm getting. I have a rather basic \> setup on Ub

---

<div class="post-metadata">

**Author:** ![davidcie](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/d/b3f665/32.png) [@davidcie](https://www.ruby-forum.com/u/davidcie)\
**Post date:** [October 11, 2011, 5:43pm UTC](https://www.ruby-forum.com/t/max-clients-worker-processes-worker-connections/212711/5 "2011-10-11T17:43:39Z")

</div>

On 10/11/11, Maxim D. [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid) wrote:  
Sorry, haven’t noticed Maxim’s answer.