Bug 518326

Description of problem: 

After turning on slave replication on a system, the slave creates critical relay log files in /var/run/mysqld by default, and that directory is blown away at boot time.

Version-Release number of selected component (if applicable):
mysql-server-5.0.45-7.el5


How reproducible: Easy


Steps to Reproduce:
1. Configure mysql server as a master in master/slave replication setup and add replication permissions for a DB user
2. On second slave server, add server-id setting in /etc/my.cnf (documented well) but do not add relay-log setting (which is only mentioned as optional in /var/log/mysqld.log to survive a hostname change)  
3. follow mysql docs for using "change master to" command and then "start slave"
  
Actual results:
Slave replication will work until the system is rebooted

Expected results:
Slave system should not lose its relay logs and should continue to replicate data

Additional info:
Recommendations: change default location of relay-log setting when you build mysql-server, or document in /etc/my.conf comments section that relay-log location needs to be set to a persistent directory, like /var/lib/mysql, because /var/run/mysqld is the default place mysqld uses and is not persistent.