# Question about UPSTREAM configuration

**URL:** <https://www.ruby-forum.com/t/question-about-upstream-configuration/228652>\
**Category:** NGINX\
**Created:** [May 1, 2013, 7:30pm UTC](https://www.ruby-forum.com/t/question-about-upstream-configuration/228652 "2013-05-01T19:30:47Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![lflacayo](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/l/8797f3/32.png) [@lflacayo](https://www.ruby-forum.com/u/lflacayo)\
**Post date:** [May 1, 2013, 7:30pm UTC](https://www.ruby-forum.com/t/question-about-upstream-configuration/228652/1 "2013-05-01T19:30:47Z")

</div>

Hello All,

Hope some one out there can help me a clear up my understanding.

If I have 5 server on an upstream configuration list, to help load  
balance  
the load across the 5 servers. Once in a while I am seen “(110:  
Connection timed out) while connecting to upstream”. Now what I am not  
clear on is that message as a result of trying all the servers on the  
upstream list or after failing on a single entry on the upstream list?  
if  
it is a single server then it should transfer the request to the next  
server  
on the list, right?

Thank you.

Luis

Posted at Nginx Forum:

> **[Question about UPSTREAM configuration...](https://forum.nginx.org/read.php?2,238770,238770#msg-238770)**
>
> Hello All, Hope some one out there can help me a clear up my understanding. If I have 5 server on an upstream configuration list, to help load balance the load across the 5 servers. Once in a while I am seen "(110: Connection timed out) while...

---

<div class="post-metadata">

**Author:** ![lflacayo](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/l/8797f3/32.png) [@lflacayo](https://www.ruby-forum.com/u/lflacayo)\
**Post date:** [May 1, 2013, 7:43pm UTC](https://www.ruby-forum.com/t/question-about-upstream-configuration/228652/2 "2013-05-01T19:43:04Z")

</div>

Hello!

On Wed, May 01, 2013 at 01:30:14PM -0400, lflacayo wrote:

> Hello All,
> 
> Hope some one out there can help me a clear up my understanding.
> 
> If I have 5 server on an upstream configuration list, to help load balance  
> the load across the 5 servers. Once in a while I am seen “(110:  
> Connection timed out) while connecting to upstream”. Now what I am not  
> clear on is that message as a result of trying all the servers on the  
> upstream list or after failing on a single entry on the upstream list?

It’s about single server (and it should print server’s address in  
the following text).

> if  
> it is a single server then it should transfer the request to the next server  
> on the list, right?

This depends on proxy\_next\_upstream directive, but usually the  
answer is yes.

–  
Maxim D.

> **[nginx: donation](http://nginx.org/en/donation.html)**

---

<div class="post-metadata">

**Author:** ![lflacayo](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/l/8797f3/32.png) [@lflacayo](https://www.ruby-forum.com/u/lflacayo)\
**Post date:** [May 2, 2013, 3:55pm UTC](https://www.ruby-forum.com/t/question-about-upstream-configuration/228652/3 "2013-05-02T15:55:05Z")

</div>

Thank you for your prompt reply. just one more … If I have 5  
servers on  
the upstream, use the IP\_HASH, is the load distribution based on a round  
robin when requests would go to the first on the list. then the second,  
the third, … Or does NGIX create some kind of distribution array.

Thank you in advance.

Luis

Posted at Nginx Forum:

> **[Re: Question about UPSTREAM configuration...](https://forum.nginx.org/read.php?2,238770,238794#msg-238794)**
>
> Hello All, Hope some one out there can help me a clear up my understanding. If I have 5 server on an upstream configuration list, to help load balance the load across the 5 servers. Once in a while I am seen "(110: Connection timed out) while...