Wayne,
If FileMaker is using the new LaunchDaemon utility - which they should - then
it wouldn't be correct to start the app using the command below, but rather
with something like terascript does:
launchctl load /path_to_daemon_plist/deamon.plist
I think you can do:
launchctl list
To see all the daemon scripts that are loaded into the launcher.
When we were moving the server to 10.6, we ran into this same problem and had
to rework the daemonization process to use launchctl, which is now - I think -
mandatory for forked processes.
It's also notable that the launchctl mechanism has the ability to auto
re-launch failed daemons. You might be able to turn this on to achieve your
goal. I think it's a setting the plist file.
If all that doesn't work, then you probably have to run the command with su,
probably by doing
sudo ./fmxdbc_listener
Note that you can edit your sudoers file so that sudo doesn't need the password
to be entered.
You should also keep hounding FileMaker. The same bug existed for ODBC
connections to the Advanced Server (Windows) and they corrected it in 11.3. I
haven't needed to restart my listener in months.
Robert
-----Original Message-----
From: Wayne Irvine [mailto:[email protected]]
Sent: Sunday, November 27, 2011 3:20 AM
To: [email protected]
Subject: Witango-Talk: Trouble launching a process
I've been batting this around a few unix boards but they are a fairly aloof lot
and I can't seem to get any answers.
I am running Terascript with FileMaker on OSX. Terascript talks to FileMaker
via JDBC and FileMaker provides a process called fmxdbc_listener that listens
for the JDBC calls and passes them on to FileMaker. fmxdbc_listener has a bug
and it's memory footprint slowly grows until it crashes. FileMaker are not
interested in this bug as I guess not enough people are using JDBC and
experiencing it. I am.
I have converted an old script I got from who knows where that used to watch
witangod, and if it crashed, launch it again. The new version can test whether
fmxdbc_listener is alive and if it is, writes to a log. If it isn't it writes
to a log and fires up the process. Works fine when triggered manually. When I
trigger it through launchd (something that apparently replaces CRON) it writes
to the log just fine but fails to start up fmxdbc_listener.
[code]
#!/bin/sh
#checking to see if fmxdbc_listener exists and writing to a variable
ps -ax |grep '/[f]mxdbc_listen' > /dev/null 2>&1
#echo "$checkfmxdbc"
if [ $? = "0" ]; then
echo "fmxdbc_listener is still Running - `date`" >> /checkfmxdbc.log
else
echo "ERROR - fmxdbc_listener is NOT Running - `date`" >>
/checkfmxdbc.log
cd /Applications/FileMaker\ Pro\ 11\ Advanced/Extensions/xDBC\
Support/fmxdbc_listener.bundle/Contents/MacOS/
./fmxdbc_listener -u wayneirvine &
fi
[/code]
Using Lingon I have tried as a 'My agents', 'User agents', 'User Daemons' and
'System Agents'. I have tried in the user account 'wayneirvine' and the root
account, with the -u argument and without.
Any one have any input on this?
Wayne Irvine
----------------------------------------
To unsubscribe from this list, please send an email to [email protected]
with "unsubscribe witango-talk" in the body.
----------------------------------------
To unsubscribe from this list, please send an email to [email protected]
with "unsubscribe witango-talk" in the body.
