Checking pingdom logs show's my forum was already down for 6 to 8 hours.
What I did was restart the whole server right away, and fixed the error.
Now I started on debugging what was the root cause of the error.
I didn't think of viewing the page source of the error that might show's more info (my bad ).
The only changes I made yesterday was enable ngx_pagespeed which use my Memcached.
So I raise up my Memcached memory to 1024mb (thinking that maybe the cause of the problem).
Since I'm using Centminmod on my Dedicated server, viewing /var/log/mysqld.log show this:
This the latest content I have: http://pastebin.com/9GJJghvU
Reading the log files, looks like @Chris Deeming addon "xf_addon_update_check" one of the issue right there?
And also show's I'm lacking of buffer pool so I change it from 4GB to 5GB.
Any suggestion for my problem?
Summary of the changes I made that I think fix this problem:
- max_connections from 100 to 300
- memcached memory from 650mb to 1024mb
- innodb_buffer_poll_size from 4GB to 5GB