LFS-5.1-pre1 info

Alexander E. Patrakov semzx at newmail.ru
Sat Jan 31 19:52:04 PST 2004

Nathan Coulson wrote:

>if I recall, mysql only has one PID which should have been killed.
>I dont know how this will work, but does anyone think that adding
>                if [ -f "$PIDFILE" ]; then
>                        pidlist="`cat "$PIDFILE"` $pidlist";
>                fi
>to the killproc will help?  [it'll append contents of a PIDFILE to the
>$pidlist], so that the first pid that gets killed first is the one we
>wish.  and if for some reason, the pidfile got messed up, then it'll kill
>the running pids anyway.
>then change to the bottom
>                if [ $failure = 0 ]; then
>                        rm -f /var/run/$base.pid
>                        if [ -f "$PIDFILE" ]; then rm -f $PIDFILE; fi
>                fi
>[looks like I'll be doing some testing tonight]
That's bad because mysql removes the "master" pid file before all other 
pids go away.

BTW now I cannot really test because I am on an NPTL box and there is 
only one mysql process (you need to hack the ./configure script though, 
it greps for "LinuxThreads" in /usr/include/pthread.h)

Alexander E. Patrakov

More information about the lfs-dev mailing list