https://bugzilla.novell.com/show_bug.cgi?id=332530#c3 Michal Marek <mmarek@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |NEEDINFO Info Provider| |gtsm@gandalfsemporium.homelinux.com --- Comment #3 from Michal Marek <mmarek@novell.com> 2007-10-11 04:27:49 MST --- Lars Rupp> for me it looks like
mysql has problems with open connections during an update of the binary.
Good point. Graham, was it a loaded mysql server in your case? It could be that the _shutdown_ doesn't wait and then the start finds the old mysqld instance still blocking the sockets.
I think the problem is in the init script around the wait_for_socket() function. Perhaps you can encapsulate the call of mysqld and the wait_for_socket() function in a simple loop (I think two turns of this loop are enough as all problems are gone after a second start).
Hmm, interesting idea, I have to think what would happen if the script manages to start two mysqlds... Hopefully the slower one would just abort.
Or increase the time in wait_for_socket() ? Fedora waits 30 seconds f.e.
10.3 also has 30 seconds. Maybe we should do that in 10.2 as well. I'm still wondering why exactly 10.2 has the problem. It must be something with the mysql_upgrade call in the rc script, because that's eg. the only difference between 10.1 / SLES10 mysql and 10.2 mysql. Thanks for your input Lars ;-) -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.