Coming with Cluster Control Monitor is a cron job setup under /etc/cron.d/cmon, this one is defaultly ran each 5 minutes.
After some problems occurs on the machine hosting Cluster Control Monitor, I've finally inspected this script and put some debug timing to see how it performs and saw that it takes sometimes more than 5 minutes to execute on a i3 2120 / 4 cores CPU !!? It takes an average of 4 to 4.30 minutes to ran, sometimes more.
This is a big problem as there is no lock mechanism on this script and so each run every 5 mins ran against the other etc ... etc ... making the machine almost unusable after an amount of time.
I tried to put the cron jobs every 10 mins rather than 5 mins with an output to a log file to check :
*/10 * * * * root test -x $CMON_BINDIR/cmon_rrd_all && $CMON_BINDIR/cmon_rrd_all $CMON_BINDIR > /var/log/cmon_cron.log 2>&1
then the script always finish before the next execution but it leads to another problem : graphs are not processes correctly because it seems to be hardened with a 5 minutes processing ...
What can we do here please ? Is it a normal behaviour or do I have to fix some things on my config to see this script ran faster ?
Please sign in to leave a comment.