Hello!

On Thu, Apr 11, 2019 at 08:45:46AM -0400, guy1976 wrote:

> I'm getting random error code 500 when serving from cache, here is an
> example:
>
> 1554735637.102 10.210.44.205 57635 "GET
> kCache/.../seg-101341803-s32-v1-a1.ts HTTP/1.1" 830 "" "Mozilla/5.0 (Windows
> NT 10.0; WOW64; Trident/7.0; rv:11.0) like Gecko" 200 897748 0.052 HIT -
> https
> 1554735766.426 10.210.184.156 57413 "GET
> kCache/.../seg-101341803-s32-v1-a1.ts HTTP/1.1" 602 "" "Mozilla/5.0 (Windows
> NT 10.0; WOW64; Trident/7.0; rv:11.0) like Gecko" 500 0 0.000 HIT - https
>
> Unfortunately nginx error log doesn't show anything, any idea why?

From the logs provided I would assume that the 500 error is caused
by a memory allocation failure at some early stage when sending a
cached response. These are expected to be logged at the "emerg"
level though, and are likely to appear in the error log.

If you don't see anything in error log, first of all I would
recommend to make sure you have error_log properly configured.
Unfortunately, often people configure nginx to log errors to
/dev/null, or configure very high logging level. Consider "info"
level to see everything nginx writes to logs.

Also, make sure there are no interference with different
error_log's. As nginx allows configuring error_log on
per-top-level-module, per-server and even per-location basis, it
might be easy to overwrite things at different configuration
levels. Usually it is a good idea to configure error_log at
global level, and comment out everything else.

If this still doesn't work, consider configuring debugging log to
see what happens. See http://nginx.org/en/docs/debugging_log.html
for details.

Also, you may want to upgrade (nginx 1.8.x is not supported for
several years) and make sure you are not using 3rd party modules
(many of these can cause arbitrary problems).

--
Maxim Dounin
http://mdounin.ru/
_______________________________________________
nginx mailing list
nginx@nginx.org
http://mailman.nginx.org/mailman/listinfo/nginx