I changed log xinetd.conf to log to a different location. For whatever reason that allowed me to get a message onscreen just after xinetd said OK. It had a mysqld "problem." It didn't say what it was. Just this,
mysqld_safe logging to /var/log/mysqld.log
starting mysqld daemon with databases from /var/lib/mysql
I booted into the Oracle Linux Rescue disk, and at /etc/init.d and,
mv mysqld mysqld.disabled
was able to get to the desktop after that. BTW, it never did log to my new location /var/log/somelogfile
However, when I start mysqld manually with mysqld.disabled start it works. No errors. I don't know. I can't find anything useful in any log, but there are tons of them, so I'm sure I'm looking in the wrong place. It only gets hung (that I can tell) when xinetd tries to start mysqld. I can reproduce this problem by doing the reverse above..
I think the SELinux thing was because I was editing the files with the rescue cd. When I stopped doing that I didn't get the problem anymore.
You can ask me about Oracle database, but I'm afraid I do not have the slightest clue about anything to do with mysql. FWIW, SELinux can be the reason for strange issues. Anyway, the system should not be doing a SELinux policy relabel on restart. What did you do? Did you restore the system to a new drive. Did you rename or change the root user?
I suggest to see if you have a /.autorelabel file, and verify its owned by root. You can also try to disable SELinux or set it to permissive in /etc/selinux/config and reboot the system to see if helps with your mysql startup. Perhaps you have to restore the SELinux security context of your my.cnf file.
There are several articles about mysql and SELinux on the web. Perhaps the following is useful - I have no idea, it was just the first to pop up in Google.