# Nginx Terminates at 71.6 kilobytes

**URL:** <https://www.ruby-forum.com/t/nginx-terminates-at-71-6-kilobytes/187578>\
**Category:** NGINX\
**Created:** [April 26, 2010, 7:38am UTC](https://www.ruby-forum.com/t/nginx-terminates-at-71-6-kilobytes/187578 "2010-04-26T07:38:22Z")\
**Posts on this page:** 7\
**Page:** 1

<div class="post-metadata">

**Author:** ![Shahan\_K](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/s/f4b2a3/32.png) [@Shahan\_K](https://www.ruby-forum.com/u/Shahan_K)\
**Post date:** [April 26, 2010, 7:38am UTC](https://www.ruby-forum.com/t/nginx-terminates-at-71-6-kilobytes/187578/1 "2010-04-26T07:38:22Z")

</div>

Hi,

I am using the latest development version of nginx (0.8.36) on  
Debian Lenny x86.

For some reason, my nginx installation is not allowing  
files larger than 71.6 kilobytes to transfer. A larger file simply just  
transfers the first 71.6 kilobytes and then terminates.

Any suggestions  
on what I can do would be helpful.

Thank you,

Shahan

---

<div class="post-metadata">

**Author:** ![Shahan\_K](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/s/f4b2a3/32.png) [@Shahan\_K](https://www.ruby-forum.com/u/Shahan_K)\
**Post date:** [April 26, 2010, 8:30am UTC](https://www.ruby-forum.com/t/nginx-terminates-at-71-6-kilobytes/187578/2 "2010-04-26T08:30:27Z")

</div>

On Sun, Apr 25, 2010 at 10:08:40PM -0700, Shahan K. wrote:

> I am using the latest development version of nginx (0.8.36) on  
> Debian Lenny x86.
> 
> For some reason, my nginx installation is not allowing  
> files larger than 71.6 kilobytes to transfer. A larger file simply just  
> transfers the first 71.6 kilobytes and then terminates.
> 
> Any suggestions  
> on what I can do would be helpful.

Is anything in error\_log ?

–  
Igor S.

> **[Igor Sysoev](http://sysoev.ru/en/)**

---

<div class="post-metadata">

**Author:** ![Shahan\_K](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/s/f4b2a3/32.png) [@Shahan\_K](https://www.ruby-forum.com/u/Shahan_K)\
**Post date:** [April 27, 2010, 3:37am UTC](https://www.ruby-forum.com/t/nginx-terminates-at-71-6-kilobytes/187578/3 "2010-04-27T03:37:54Z")

</div>

The error wasn’t nginx’s fault. I the temporary cache directory ran out  
of space.

Thanks,

Shahan

On Sun, 25 Apr 2010 22:08:40 -0700, Shahan  
Khan wrote:

Hi,

I am using the latest development version of nginx  
(0.8.36) on Debian Lenny x86.

For some reason, my nginx installation is  
not allowing files larger than 71.6 kilobytes to transfer. A larger file  
simply just transfers the first 71.6 kilobytes and then terminates.

Any  
suggestions on what I can do would be helpful.

Thank you,

Shahan

---

<div class="post-metadata">

**Author:** ![Shahan\_K](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/s/f4b2a3/32.png) [@Shahan\_K](https://www.ruby-forum.com/u/Shahan_K)\
**Post date:** [April 27, 2010, 9:12am UTC](https://www.ruby-forum.com/t/nginx-terminates-at-71-6-kilobytes/187578/4 "2010-04-27T09:12:05Z")

</div>

The same thing happened to me ~1 year ago, but I didn’t reported it.  
When you run out of space, nginx only serves the first  
proxy\_buffer\_size kb from the response. This was the source of  
corruption of part of my images during a backup done through nginx.

IMO in this case nginx should answer with a 50x .

On Tue, Apr 27, 2010 at 3:36 AM, Shahan K. [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid) wrote:

> Shahan
> 
> * * *
> 
> nginx mailing list  
> [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid)  
> [mailman.nginx.org archive](http://nginx.org/mailman/listinfo/nginx)

–  
Joaquin Cuenca Abela

---

<div class="post-metadata">

**Author:** ![Shahan\_K](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/s/f4b2a3/32.png) [@Shahan\_K](https://www.ruby-forum.com/u/Shahan_K)\
**Post date:** [April 27, 2010, 6:48pm UTC](https://www.ruby-forum.com/t/nginx-terminates-at-71-6-kilobytes/187578/5 "2010-04-27T18:48:48Z")

</div>

Maybe a nginx feature that should be added is when the temporary cache  
directory runs out of space, nginx disable the proxy cache buffer and  
also  
disables gzip encoding.

Shahan

On Tue, 27 Apr 2010 09:11:42 +0200, Joaquin Cuenca Abela  
[removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid) wrote:

> > space.  
> > Lenny  
> > x86.
> > 
> > For some reason, my nginx installation is not allowing files larger  
> > than  
> > 71.6 kilobytes to transfer. A larger file simply just transfers the  
> > first

---

<div class="post-metadata">

**Author:** ![Shahan\_K](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/s/f4b2a3/32.png) [@Shahan\_K](https://www.ruby-forum.com/u/Shahan_K)\
**Post date:** [April 27, 2010, 9:12am UTC](https://www.ruby-forum.com/t/nginx-terminates-at-71-6-kilobytes/187578/6 "2010-04-27T09:12:05Z")

</div>

The same thing happened to me ~1 year ago, but I didn’t reported it.  
When you run out of space, nginx only serves the first  
proxy\_buffer\_size kb from the response. This was the source of  
corruption of part of my images during a backup done through nginx.

IMO in this case nginx shou

On Tue, Apr 27, 2010 at 3:36 AM, Shahan K. [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid) wrote:

> Shahan
> 
> * * *
> 
> nginx mailing list  
> [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid)  
> [mailman.nginx.org archive](http://nginx.org/mailman/listinfo/nginx)

–  
Joaquin Cuenca Abela

---

<div class="post-metadata">

**Author:** ![Shahan\_K](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/s/f4b2a3/32.png) [@Shahan\_K](https://www.ruby-forum.com/u/Shahan_K)\
**Post date:** [April 29, 2010, 6:43pm UTC](https://www.ruby-forum.com/t/nginx-terminates-at-71-6-kilobytes/187578/7 "2010-04-29T18:43:23Z")

</div>

On Tue, Apr 27, 2010 at 6:47 PM, Shahan K. [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid) wrote:

> Maybe a nginx feature that should be added is when the temporary cache  
> directory runs out of space, nginx disable the proxy cache buffer and also  
> disables gzip encoding.

I personally disagree, it’s quite normal (if not expected!) for software  
to  
stop working upon disk fill-up, and trying to recover smartly from this  
(and  
then maybe a few million other issues) would make it more complex, less  
mantainable and slower.

+1 for returning a 500 instead, and logging the error as critical if it  
doesn’t already.

Bye
