X-Git-Url: http://git.onelab.eu/?a=blobdiff_plain;f=automate-default.sh;h=b5508c13ffacb32f1eb29e6d0c5840478e91a375;hb=refs%2Fheads%2F1.0;hp=1aaeb59f59e9bfae32cc9346d78049e6d4909769;hpb=ff1455177461f5adf513543fd2a85289f359ff0e;p=monitor.git

diff --git a/automate-default.sh b/automate-default.sh
index 1aaeb59..b5508c1 100755
--- a/automate-default.sh
+++ b/automate-default.sh
@@ -1,12 +1,16 @@
 #!/bin/bash
 
 # NOTE: Must be an absolute path to guarantee it is read.
-source /usr/share/monitor-server/monitorconfig.py
+INSTALLPATH=/usr/share/monitor-server/
+# Generate an 'sh' style file full of variables in monitor.conf
+$INSTALLPATH/shconfig.py >  $INSTALLPATH/monitorconfig.sh
+source $INSTALLPATH/monitorconfig.sh
 cd ${MONITOR_SCRIPT_ROOT}
 set -e
 DATE=`date +%Y-%m-%d-%T`
-MONITOR_PID="$HOME/monitor/SKIP"
+MONITOR_PID="${MONITOR_SCRIPT_ROOT}/SKIP"
 
+echo "#######################################"; echo "Running Monitor at $DATE"; echo "######################################"
 echo "Performing API test"
 API=$(./testapi.py)
 if [ "$API" != "ok" ] ; then 
@@ -20,7 +24,12 @@ if [ -f $MONITOR_PID ] ; then
 		echo "KILLING Monitor"
 		PID=`cat $MONITOR_PID`
 		rm -f $MONITOR_PID
-		${MONITOR_SCRIPT_ROOT}/kill.cmd.sh $PID
+		if [ -z $PID ] ; then
+			${MONITOR_SCRIPT_ROOT}/kill.cmd.sh $PID
+			echo "done."
+		else
+			echo "No PID to be killed."
+		fi
 	else 
 		# skipping monitor
 		echo "SKIPPING Monitor"
@@ -29,7 +38,17 @@ if [ -f $MONITOR_PID ] ; then
 fi
 echo $$ > $MONITOR_PID
 
+# SETUP act_all database if it's not there.
+if [ ! -f ${MONITOR_SCRIPT_ROOT}/actallsetup.flag ]; then
+	if ! python -c 'import database; database.dbLoad("act_all")' 2>/dev/null ; then 
+		touch ${MONITOR_SCRIPT_ROOT}/actallsetup.flag
+	fi
+fi
+
+
+set +e
 AGENT=`ps ax | grep ssh-agent | grep -v grep`
+set -e
 if [ -z "$AGENT" ] ; then
         echo "starting ssh agent"
         # if no agent is running, set it up.
@@ -48,7 +67,7 @@ echo "Performing Findbad Nodes"
 rm -f ${MONITOR_DATA_ROOT}/production.findbad2.pkl
 ${MONITOR_SCRIPT_ROOT}/findbad.py --increment --cachenodes --debug=0 --dbname="findbad2" $DATE || :
 cp ${MONITOR_DATA_ROOT}/production.findbad2.pkl ${MONITOR_DATA_ROOT}/production.findbad.pkl
-ps ax | grep BatchMode | grep -v grep | awk '{print $1}' | xargs kill || :
+ps ax | grep BatchMode | grep -v grep | awk '{print $1}' | xargs -r kill || :
 
 echo "Performing Findbad PCUs"
 #########################
@@ -57,13 +76,13 @@ rm -f ${MONITOR_DATA_ROOT}/production.findbadpcus2.pkl
 ${MONITOR_SCRIPT_ROOT}/findbadpcu.py --increment --refresh --debug=0 --dbname=findbadpcus2 $DATE || :
 cp ${MONITOR_DATA_ROOT}/production.findbadpcus2.pkl ${MONITOR_DATA_ROOT}/production.findbadpcus.pkl
 # clean up stray 'locfg' processes that hang around inappropriately...
-ps ax | grep locfg | grep -v grep | awk '{print $1}' | xargs kill || :
+ps ax | grep locfg | grep -v grep | awk '{print $1}' | xargs -r kill || :
 
-echo "Generating web data"
+#echo "Generating web data"
 # badcsv.txt
-${MONITOR_SCRIPT_ROOT}/printbadcsv.py  | grep -v loading | tr -d ' ' > badcsv.txt
-cp badcsv.txt /plc/data/var/www/html/monitor/
-${MONITOR_SCRIPT_ROOT}/showlatlon.py | head -9 | awk 'BEGIN {print "<table>"} { print "<tr><td>", $0, "</td></tr>"} END{print "</table>"}'  | sed -e 's\|\</td><td>\g' > /plc/data/var/www/html/monitor/regions.html
+#${MONITOR_SCRIPT_ROOT}/printbadcsv.py  | grep -v loading | tr -d ' ' > badcsv.txt
+#cp badcsv.txt /plc/data/var/www/html/monitor/
+#${MONITOR_SCRIPT_ROOT}/showlatlon.py | head -9 | awk 'BEGIN {print "<table>"} { print "<tr><td>", $0, "</td></tr>"} END{print "</table>"}'  | sed -e 's\|\</td><td>\g' > /plc/data/var/www/html/monitor/regions.html
 
 echo "Performing uptime changes for sites, nodes, and pcus"
 ########################
@@ -76,31 +95,33 @@ echo "Converting pkl files to phpserial"
 #########################
 # 4. convert pkl to php serialize format.
 ${MONITOR_SCRIPT_ROOT}/pkl2php.py -i findbadpcus2 -o findbadpcus
-${MONITOR_SCRIPT_ROOT}/pkl2php.py -i act_all -o act_all
-${MONITOR_SCRIPT_ROOT}/pkl2php.py -i plcdb_hn2lb -o plcdb_hn2lb
+for f in act_all plcdb_hn2lb ; do
+	if [ -f ${MONITOR_DATA_ROOT}/production.$f.pkl ]; then
+		${MONITOR_SCRIPT_ROOT}/pkl2php.py -i $f -o $f
+	else
+		echo "Warning: ${MONITOR_DATA_ROOT}/production.$f.pkl does not exist."
+	fi
+done
 ${MONITOR_SCRIPT_ROOT}/pkl2php.py -i findbad -o findbadnodes
-${MONITOR_SCRIPT_ROOT}/pkl2php.py -i ad_dbTickets -o ad_dbTickets
-${MONITOR_SCRIPT_ROOT}/pkl2php.py -i idTickets -o idTickets
+#${MONITOR_SCRIPT_ROOT}/pkl2php.py -i ad_dbTickets -o ad_dbTickets
+#${MONITOR_SCRIPT_ROOT}/pkl2php.py -i idTickets -o idTickets
 
 echo "Archiving pkl files"
 #########################
 # Archive pkl files.
-for f in findbad act_all findbadpcus l_plcnodes site_persistflags node_persistflags pcu_persistflags ; do 
-	cp ${MONITOR_DATA_ROOT}/production.$f.pkl ${MONITOR_ARCHIVE_ROOT}/`date +%F-%H:%M`.production.$f.pkl
+for f in findbad act_all findbadpcus l_plcnodes site_persistflags node_persistflags pcu_persistflags ; do
+	if [ -f ${MONITOR_DATA_ROOT}/production.$f.pkl ] ; then
+		cp ${MONITOR_DATA_ROOT}/production.$f.pkl ${MONITOR_ARCHIVE_ROOT}/`date +%F-%H:%M`.production.$f.pkl
+	else
+		echo "Warning: It failed to archive ${MONITOR_DATA_ROOT}/production.$f.pkl"
+	fi
 done
 
 echo "Running grouprins on all dbg nodes"
 ############################
 # 5. Check if there are any nodes in dbg state.  Clean up afterward.
-${MONITOR_SCRIPT_ROOT}/grouprins.py --mail=1 \
-	--nodeselect 'state=DEBUG&&boot_state=(rins|dbg|boot)' \
-	--stopselect 'state=BOOT&&kernel=2.6.22.19-vs2.3.0.34.9.planetlab' \
-	--reboot || :
-${MONITOR_SCRIPT_ROOT}/findbad.py --increment --cachenodes --debug=0 --dbname="findbad" --nodeselect 'state=DEBUG&&boot_state=dbg||state=DEBUG&&boot_state=boot' || :
-
-echo "Collecting RT database dump"
-##########################
-# 6. cache the RT db locally.
-python ${MONITOR_SCRIPT_ROOT}/rt.py
+${MONITOR_SCRIPT_ROOT}/grouprins.py --mail=1 --reboot --nodeselect 'state=DOWN&&boot_state=(boot|rins|dbg|diag)' --stopselect "state=BOOT" || :
+${MONITOR_SCRIPT_ROOT}/grouprins.py --mail=1 --reboot --nodeselect 'state=DEBUG&&boot_state=(rins|dbg|boot)' --stopselect 'state=BOOT' || :
 
+cp ${MONITOR_SCRIPT_ROOT}/monitor.log ${MONITOR_ARCHIVE_ROOT}/`date +%F-%H:%M`.monitor.log
 rm -f $MONITOR_PID