I'm running Linux SuSE 7.3 Pro.  I'm trying to get the Zope server to start at boot-time.  I found the 'jules' document on www.zope.org that gave steps to do so with 6.3, but after some modifications it still isn't working.  As 'root' I followed the instructions given in the link http://www.zope.org/Members/jules/SuSE-6.3_html for setting up Zope as a daemon in Suse Linux.  I made all the necessary adjustments for directory differences between 6.3 and 7.3 (e.g. - 'init.d' is a sub-dir off of '/etc' in 7.3 instead of being a sub-dir off of '/sbin' as it is in 6.3, according to 'jules' document).  I changed the symbolic link files to be in rc5.d not rc2.d since SuSE 7.3 uses RUNLEVEL 5 not 2.  I changed the number portions of the symbolic link files from the hard-coded '22' given in 'jules' document to '24' which was the next available daemon number.  (If you look in rc5.d you'll see that all the 'K$' and 'S$' scripts are sequential.  On my system '24' was the next available number).  So I have 'K24zope' and 'S24zope' in my rc5.d directory.  Lastly, if you make the 'START_ZOPE=yes' addition to rc.config, you still have to jump into 'yast2', and using the rc-config-editor module,  change START_ZOPE to 'yes'.  This will result in SuSEconfig running and setting things up properly.  If you don't, then you will find out that even though the rc.config line is 'START_ZOPE=yes', if you print out the value of the command line argument to 'zope' ($1) you'll see that it is set to 'no'.  After doing the 'yast2' operation it will be properly set to 'yes'.

 

I then tested my sample script in an xterm window.  Given the proper command line argument (start/stop/status), it started Zope successfully, stopped it, and gave a checkproc status all perfectly.  When run at bootup however, after boot completes, if I look in the boot message log I see the 'zope' script startup messages, but none of the other messages from SuSE that used to follow the initialization of all daemon's.  I think this is causing SuSE to kill the 'zope' process due to a timeout because when I get fully logged in, 'zope' is not in the kpm's process list at all.  What's also strange is that I do have to run 'zope stop' to clear a lock-flag that 'lock_file.py' put in place, and then I can run 'zope start' again and it loads fine. 

 

Just to be very clear, it's the same behavior that occurs when you run 'zope start' from an xterm window.  The script never returns until you shut down the zope server.  I would suspect that this kind of behavior doesnt' sit well with the bootup scripts.

So what do I have to do to have 'zope start' exit properly so it can run as a daemon?

thx