If anyone was wondering what that was... basically DNS lookups were failing on the database server, and all connections to the database server were being queued with most timing out. The result was everything went very slow.
Adding skip-name-resolve to my.cnf avoids the needless DNS lookup and everything is flying again.
What sucks is that even with this grand new server farm, the single point of failure is the master database server... and that was the one thing that failed.
I'm looking forward to distributed NoSql databases to solve this shit.
**Geek. **
A good Geek; our Geek but a Geek nonetheless.
Thank you for all you have done and Maerry Christmas.
**Geek. **
A good Geek; our Geek but a Geek nonetheless.
Thank you for all you have done and Maerry Christmas.