On 11/06/2012 04:01 PM, Colin Guthrie wrote:
Curious indeed. Can you pastebin/attach the initscript and give the
full status output after the failure? Cheers Col
This is nuts. For the instance I started successfully with
SKIP_REDIRECT, a straight "service oracle-xe stop" from a bash session
without SKIP_REDIRECT worked, and brought it down. A subsequent
"service oracle-xe start" worked. As did another stop/start.
So, as a result (apparently) of a successful start of oracle-xe, now
everything works again.
Here's one possibility: Before I tried the SKIP_REDIRECT, I modified the
"start" statement in the script to eliminate a ">/dev/null 2>&1" in the
hopes of getting some diagnostics. Then I did a "systemctl --system
daemon-reload". It still failed, and I didn't get any output that I
could find, but maybe the reload had something to do with the eventual
success (I can't see how).
I restored the ">/dev/null 2>&1" and did another reload, but everything
still worked. Is it possible that whatever is supposed to suck the
initscript into the systemd universe isn't quite doing the job ? That
still wouldn't explain the failure after the reload but before the
SKIP_REDIRECT, but it might have some bearing.
If you want to test this, you can get the RPM at
http://www.oracle.com/technetwork/products/express-edition/downloads/index.html.
Install it and add your own LSB headers to the initscript, run
"/etc/init.d/oracle-xe configure" and take all the defaults, and you're
good to go. As it's an RPM, it's easy and clean to get rid of when
you're done. My gut tells me that I'm not going to get this to fail
again unless I do a fresh install.
Thanks,
Frank
