Strange bugs - want to hire experienced web developer

[o hai guyz]

I have 2 strange issues on my site that I've been trying to solve for a couple of weeks but haven't had any luck. It's an image-sharing site similar to Tumblr or a web version of Instagram.

(1) Site runs EXTREMELY slow any time there are over about 150 people on it simultaneously (load time jumps from the usual ~1 sec to 8+ sec). This is on a $250/month dedicated server.

(2) McAfee has it flagged as a "phishing site" and Facebook has it flagged as "unsafe." This is a legitimate site with no hidden phishing scripts or anything even remotely related to that type of content. I'm not sure exactly what's causing this but I assume there's some strange pattern in the code that's causing it to be mistakenly flagged.

If you'd like to take a look please PM me your skype and/or email. I'm more than willing to pay a decent rate for this.

 

[OsamaBinBBQ]

1a) Move to AWS, you're paying way too much for your level of traffic. Scaling will be easier and even if your site isn't optimized for HA, you can scale vertically and switch over to larger instances when you're under heavy load.

1b) There are many reasons for slowing down under load, but some times it is purely because the server isn't configured to allow more than a certain amount of connections at a time. If you're not using a CDN already, use one. cloudflare.com is free and takes only minutes to setup and deploy. If you're moving to or already on AWS, check out cloudfront. Although serving static assets is the simplest and fastest thing a server can do, if you're reaching your concurrent request limit, moving these assets to a CDN will free probably 90% of your connections.

1c) Look into tuning your http server. If you're using Apache, setting MaxRequestsPerChild to something in the thousands (say 5000) rather than 0, is a good way to automatically and periodically dump a child thread that may have been leaking memory (which will increase load).

1d) If your content is dynamically generated but doesn't change very often, look into using a proxy cache like Squid (squid : Optimising Web Delivery).

2) Use the money you saved by moving to AWS and spend it on shitty McAfee services to find out why they think you're phishing.

 

[(O_o)]

1c) Look into tuning your http server. If you're using Apache, setting MaxRequestsPerChild to something in the thousands (say 5000) rather than 0, is a good way to automatically and periodically dump a child thread that may have been leaking memory (which will increase load).

this 1000%

if its the site i'm thinking of, you might need to do this. I had to do it on a site of mine and it fixed my lag as soon as I did this. I remember the default it was set to was really low.

 

[AMcDermott]

Agreed with the suggestions above, especially AWS.

1c) Look into tuning your http server. If you're using Apache, setting MaxRequestsPerChild to something in the thousands (say 5000) rather than 0, is a good way to automatically and periodically dump a child thread that may have been leaking memory (which will increase load).

If you are using Apache consider moving to Litespeed.