check_lustre_client

Check Lustre client health status

Usage

check_lustre_client

No command line parameters. This plugin has to be executed locally on the Lustre client, possibly via NRPE.

Detection relies largely on Lustre internal health_check. Unfortunately, Lustre can be seriously b0rk3d while health_check says "healthy".

#!/bin/bash OK=0 #WARNING=1 CRITICAL=2 #UNKNOWN=3 RET=$OK MSG= # 1. does lustre proc directory exists ? PROC_DIR=/proc/fs/lustre [ -d $PROC_DIR ] || { MSG="$PROC_DIR doesn't exist " ; RET=$CRITICAL ; } # 2. is lustre client healty ? HEALTH=$PROC_DIR/health_check grep -q healthy $HEALTH || { MSG="${MSG}lustre health status is: $(<$HEALTH) " ; RET=$CRITICAL } # 3. is it mounted ? grep -q lustre /proc/mounts || { MSG="${MSG}lustre not mounted" ; RET=$CRITICAL } # more checks would require to enable cmdline arguments passing to nrpe AND would risk to block [ $RET -eq $OK ] && MSG="lustre client OK" echo $MSG exit $RET
download code