Optimize SO for nginx

Peter Booth peter_booth at me.com
Sat Apr 20 06:05:39 UTC 2019


Where is your upstream? Where is your pho executing? Do you have a CDN?

There’s three parts to this:
1 fix the bad OS defaults:
If you are using RHEL 6 this would mean:
Enabling tuned
Disabling THP
Increasing vm.min_free_kbytes
Reducing swappiness
2 generic web server specific configuration
Increasing # ephemeral ports
Adjusting ulimit
3 tuning specific to your workload.
UDP and TCP buffet size - should match BDP
NIC tuning - irq coalescing, 10G specific tuning, see CDN, melanox,redhat, hp suggestions for lie latency tuning.

That’s a start.

Peter



Sent from my iPhone

> On Apr 19, 2019, at 6:37 PM, Rick Gutierrez <xserverlinux at gmail.com> wrote:
> 
> Hi, some recommendation to optimize a server of 4gb ram and 4 core for reverse proxy, with ssl activated, brotli and centos 7 SO,  when I speak of optimizing is to modify the sysctl.conf
> 
> I need to support 12k in a couple of applications with php and asp.net
> 
> some experience advice?
> -- 
> rickygm
> 
> http://gnuforever.homelinux.com
> _______________________________________________
> nginx mailing list
> nginx at nginx.org
> http://mailman.nginx.org/mailman/listinfo/nginx
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.nginx.org/pipermail/nginx/attachments/20190420/99f8d982/attachment.html>


More information about the nginx mailing list