New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
what do you mean?
Indeed, the term "C10k problem" is from the 1990s.
5000 clients with mpm_prefork might be doable with a big enough server. It would be a horrendous use of resources but might be ok as a stopgap, launch first and fix later. You'll have to increase ulimits and maybe some kernel parameters. I remember hitting some snags at around 1500 connections with python threads.
Well prefork "died" 8-9 years ago.. or at least it should. I am talking about mpm_event or mpm_worker at least (with mod_fcgid)
OP says mpm_prefork in original post. Is it actually removed from apache? I thought maybe some application setups depend on it. I used it sometimes at low concurrency and it made life simpler. At high concurrency of course it is crazy.
see
And with Cloudflare, you want to optimise your connection between CF and your origin by using better performing SSL ciphers via ECDSA based SSL certificates. Guide I wrote for my Centmin Mod users at https://community.centminmod.com/threads/improving-cloudflare-connections-to-origin-server-use-ecdsa-ssl-certs.14817/
No, it's still there.. it's just not "industry standard" anymore. Sure it can probably host your 100-visits/month blog out of Raspberry Pi without issues
Based on my previous post and mod_fcgid recommendation (which I actually use mostly), I mainly use HHVM for PHP for high traffic needs now-days.
As soon as people realise it's not web server that is bottleneck but rather backend (PHP or.. ? ), wars nginx-apache would stop. IMHO php-fpm sucks balls on high traffic - low hardware scenarios. That is why it's only good for people having 100-visits blogs and people with budget (ie. AWS logic - scale vertically )