On 2/10/03 2:53 PM, "Bruce Dembecki" <bru...@liveworld.com> wrote:

I have analyzed the output from mysqld_safe - and obtained the following command by using "echo" in the mysqld_safe script rather than executing the command itself...

nohup /usr/local/mysql-standard-4.0.10-gamma-sun-solaris2.8-sparc/bin/mysqld --defaults-extra-file= /usr/local/mysql-standard-4.0.10-gamma-sun-solaris2.8-sparc/data/my.cnf --basedir=/usr/local/mysql-standard-4.0.10-gamma-sun-solaris2.8-sparc --datadir=/mysqldata --user=mysql --pid-file=/mysqldata/mysql2.pid >> /mysqldata/mysql2.err 2>&1

Clearly I have forced some line breaks in the command for easier reading and email transference...

I also know that mysqld_safe itself is called by mysql.server using commands to force it into the background, and in fact mysqld_safe continues to run after exiting SSH... Therefore as we know mysqld is a child of mysqld_safe and mysqld_safe is run in the background from the shell and mysqld_safe starts mysqld with nohup, it would appear that something in the way mysqld is compiled is preventing it from working correctly when it gets a SIGHUP it should ignore.

This of course applies only to Solaris 2.8 and 2.9 which I have used for testing here. It also may be totally wrong, but it's the best theory I've got.

Best Regards, Bruce