I attach an example of a monit script I use. Some notes: 1) Not everybody likes to automatically restart stone if it crashed (I do it). Some prefer a simply allow monit to throw a notification (email?) and then take human action (recover from tranlogs? analyze logs? recover from backup? run object audit? etc etc). 2) I use my own scripts for starting/stopping a seaside gem at a given port. 3) You must change the timeout of the FastCGI parts of the monit file to match the one in your webserver configuration timeout for FastCGI. 4) Again, I try to keep monit file in the directory of the stone and do a symlink to the monit directory in /etc. 5) I autogenerate these monit files as part of my "site building scripts" 6) You can change the start/stop program of the stone to use gsDevKit `startStone` and `stopStone` and also the pid file to point to $GS_HOME/server/stones/stoneXXX/ mmmmmmm `startStone` does not create a pid file??? else I cannot find it... 7) There could be times where you want to put gems down (maybe for a scheduled cleanup/maintenance) or even the stone (upgradeStone for example). For this cases, killing monit may not be the best path. So for all those cases, what I use is code like this: monit unmonitor -g ${GEMSTONE_NAME} >> $GEMSTONE_LOGDIR/${GEMSTONE_NAME}_backup.log 2>&1 Cheers, On Tue, Nov 24, 2015 at 6:23 PM, Jupiter Jones <[hidden email]> wrote:
_______________________________________________ Glass mailing list [hidden email] http://lists.gemtalksystems.com/mailman/listinfo/glass monitDebrisTesting.txt (4K) Download Attachment |
Free forum by Nabble | Edit this page |