Description of problem: There is .schedd_address(contains invalid address, not cleaned after scheduler crash) file which has higher priority than scheduler name(information from collector). condor_q has wrong output. condor_q -name `condor_config_val SCHEDD_NAME` has right output. Version-Release number of selected component (if applicable): How reproducible: 100% Steps to Reproduce: 1. set up pool with HA Scheduler + RH HA 2. run "i=1; while true; do echo -n "$i..."; ./y.sh ; i=$(($i+1));done" on all machines with potential HA Schedd 3. wait till condor_q start to get wrong output Actual results: Condor_q doesn't work right when there is .schedd_address with wrong address. Expected results: condor_q will output information from all available schedulers, so it will consider all schedulers from collector. Additional info: $ cat y.sh #!/bin/bash #wait till scheduler is up while [ "0$(ps aux | grep condor_schedd | grep -v grep | wc -l)" -eq "0" ]; do sleep 1 done sleep 10 killall -9 condor_schedd #wait till scheduler is up while [ "0$(ps aux | grep condor_schedd | grep -v grep | wc -l)" -eq "0" ]; do sleep 1 done sleep 60 PID=$(pidof condor_schedd) if [ -n "$PID" ]; then SAVED_PID=$(cat /var/run/condor/condor_schedd.pid) if [ "0$PID" -eq "0$SAVED_PID" ]; then echo "OK...... $PID == $SAVED_PID" else echo "ERROR... $PID != $SAVED_PID" fi fi