Adam Harrison wrote:
Our database cannot keep up and we’re trying to figure out why.
Symptoms:
- Nagios is reporting too many processes, 100+ waiting to execute
- Thousands of slow queries over 1 sec
The facts:
- We have a cluster of four mail servers running SpamAssassin
(v3.2.4), and they all use one MySQL(v4.1.22) server (RHEL4.6) running
InnoDB, which replicates to a back-up server in case the primary one
goes down.
- We are doing about 200 writes a second.
We have a similar configuration and we turned off bayes auto learning on
all but two of the machines.
The other 3 machines just use bayes to score.
Regards,
Rick
