Managed to replicate this on dev environment. Sent POST request with larger
body and killed it mid way. Nginx logs this as the 59 second HTTP 400
response. Looks like this should be something configurable.
Posted at Nginx Forum:
https://forum.nginx.org/read.php?2,262308,262388#msg-262388
___
In this case the client that closes connection is ELB with default 60
seconds. It doesn't really answer the question of why it holds the POST
requests for so long. This doesn't happen for GET requests.
Won't be able to test the debug logging output for next ~2 weeks.
Posted at Nginx Forum:
https
>>client prematurely closed connection
that means your client has closed the socket.
You can rebuilt the nginx to enable the detailed logging for debug
http://nginx.org/en/docs/debugging_log.html
2015-10-21 15:52 GMT+08:00 Pauls V :
> Had tried with reload and that didn't change output. Tried re
Had tried with reload and that didn't change output. Tried restarting Nginx
and that actually enabled the logging.
If someone is wondering then this is what I see in log now:
... client prematurely closed connection ...
Thanks,
Pauls
Posted at Nginx Forum:
https://forum.nginx.org/read.php?2,262
Hello!
On Mon, Oct 19, 2015 at 07:23:00AM -0400, Pauls V wrote:
> Problem: Nginx returns 400 on small amount of POST requests (1-5 per minute
> from 10k POST requests)
> Setup: Nginx in front of Tomcat
> Version: 1.8.0 (upgraded from 1.4.6 that had the same problem)
> Log format: log_format main
Hello,
Problem: Nginx returns 400 on small amount of POST requests (1-5 per minute
from 10k POST requests)
Setup: Nginx in front of Tomcat
Version: 1.8.0 (upgraded from 1.4.6 that had the same problem)
Log format: log_format main $remote_addr - $remote_user [$time_local]
$request $status $reques
