check_snfs_latency

Project Notes

Usage ./check_sngfs_latency.sh -f [-w ] [-c ] [-i ] [-t ] [-p on|off] [-h] [-v] Read the statistics file written in the last 60 minutes of the FSM host. Averages PIO Read and PIO Write times and exits OK - WARNING - CRITICAL and optionally outputs performance data. The check is done on stat .csv files written by the FSM on every MetaData Controller and this script needs direct access to them. Execute checks via NRPE or SSH. CSV files are written every hour. Nagios check interval should be set to 60 for this script to work. Since journals "only" write there are two values for PIO Write HiPri in the CSV file. -f Filesystem Name Name the SNFS file system to check. Case sensitive. Mandatory. -w Waring Level (us) Set the warnining level in microseconds. Defauls to Quantum's recommendation of 500us. Optional. Warning level has to be a lower value then critical level. -c Critical Level Set the critcal level in microseconds. Defaults to 1000us. Optional. Critical level has to be a higher value then warning level. -i SNFS install path Path to SNFS installation with ./qustats directory. Optional. Defaults to /usr/cvfs -p Perfdata output on or off "off" outputs no perfdata. Default is output perfdata on -t Testrun Use the file form the argument for the check. -d Tempfile name and location Path to temporary file. Defaults to /tmp/check_snfs_latency.tmp -h Print this usage and exit. -v Print verision Info and exit. EXAMPLE ./check_sngfs_latency.sh -f cvfs01 -w 200 -c 700 -p off -i /usr/adic Example data from .csv file: # NAME,TYP,COUNT,MIN,MAX,TOT/LVL,AVG PIO Write HiPri,TIM,3199,246,7872,1275437,399 PIO Read HiPri,TIM,7189,75,4981,1332093,185