This procedure describes how to setup the AIX Workload Manager (WLM) and Jtopas to store trending information to a central location via NFS.

Assumptions:

• WLM is configured and running on the AIX system to be monitored.
• An NFS directory exists for the the purpose of accepting WLM/Jtopas trending information from AIX system.
• The NFS directory is mountable by the AIX system to be monitored.
• The NFS directory is exported from one of the NIM servers.

At the time of this writing, the NIM servers in both data centers are being used as the central storage location, via NFS exported directories, for the monitoring information generated from WLM and Jtopas on each AIX system. The commands to mount the NFS directories being exported from each NIM server to an AIX system to be monitored are:

mdcapnim01:

mount mdcapnim01:/prfpmce0 /prfpmce0

ddcapnim01:

mount ddcapnim01:/prfdmce0 /prfdmce0

In the following procedure, EXAMPLE system names are used to illustrate the commands used to perform the configuration. The NIM server "mdcapnim01" is used as the EXAMPLE NFS server name, and "mdcapora01" is used as the EXAMPLE AIX system name to be monitored.

1. On the NIM server, create a shell variable to contain the system name of the AIX system to be monitored:

   mdcapnim01:

   AIXSYS="mdcapora01"    # AIX system to be Monitored
   

2. Create directories on the NFS server to store the WLM and Jtopas trending information for the AIX system to be monitored:

   mdcapnim01:

   mkdir -p /prfpmce0/${AIXSYS}/Top
   mkdir -p /prfpmce0/${AIXSYS}/wlm
   find /prfpmce0/${AIXSYS} -exec ls -ld {} \;
   

3. On the AIX system to be monitored, create shell variables to contain the system name of the AIX system to be monitored and the system name of the NIM server where the monitoring information will be stored:

   mdcapora01:

   /usr/bin/ksh93
   AIXSYS="$( hostname )"    # AIX system to be Monitored
   AIXSYS="${AIXSYS%%.*}"
   NFSSRV="mdcapnim01"    # NIM Server with NFS exported filesystem for data storage
   print ${AIXSYS}
   print ${NFSSRV}
   

4. On the AIX system to be monitored, create the NFS mount point to be automatically mounted at boot time:

   mdcapora01:

   /usr/sbin/mknfsmnt -f /prfpmce0 -d /prfpmce0 -h ${NFSSRV} -M sys -B -A -t rw -w bg -Y -X -S -j -q -g
   grep -p "/prfpmce0:" /etc/filesystems
   df
   

5. On the AIX system to be monitored, modify the "/etc/inittab" to contain the startup commands for WLM and for Jtopas:

   mdcapora01:

   rmitab ptxwlm
   rmitab ptxtrend
   mkitab "ptxwlm:2:once:/usr/bin/xmwlm -d /prfpmce0/${AIXSYS}/wlm -n xmwlm > /dev/null 2>&1 # Start xmwlm"
   mkitab "ptxtrend:2:once:/usr/bin/xmtrend -f /etc/perf/jtopas.cf -d /prfpmce0/${AIXSYS}/Top -n jtopas > /dev/null 2>&1 # Start trend"
   egrep "^ptxwlm:|^ptxtrend:" /etc/inittab
   

6. On the AIX system to be monitored, stop any existing "xmwlm" or "xmtrend" processes, then manually start the WLM and Jtopas trending processes or reboot the system and verify the processes are started. To manually start the WLM and Jtopas trending processes:

   mdcapora01:

   for PID in  $( ps -ef -F pid,args | egrep 'xmwlm|xmtrend' | grep -v grep | awk '{ print $1 }' )
   do  
       print "Killing process ${PID}"
       kill ${PID}
       sleep 2
       kill -9 ${PID}   
   done
   ps -ef -F pid,args | egrep 'xmwlm|xmtrend'
   
   nohup /usr/bin/xmwlm -d /prfpmce0/${AIXSYS}/wlm -n xmwlm > /dev/null 2>&1 &
   sleep 2
   nohup /usr/bin/xmtrend -f /etc/perf/jtopas.cf -d /prfpmce0/${AIXSYS}/Top -n jtopas > /dev/null 2>&1 &
   sleep 2
   find /prfpmce0/${AIXSYS} -exec ls -ld {} \;
   jobs
   

7. Cleanup any shell variables created during the execution of this procedure:

   mdcapora01:

   umount /prfdmce0
   /usr/sbin/rmnfsmnt -f /prfdmce0 -B
   unset AIXSYS
   unset NFSSRV
   

   mdcapnim01:

   unset AIXSYS
   

In the following procedure, EXAMPLE system names are used to illustrate the commands used to perform the configuration. The NIM server "ddcapnim01" is used as the EXAMPLE NFS server name, and "ddcadora01" is used as the EXAMPLE AIX system name to be monitored.

1. On the NIM server, create a shell variable to contain the system name of the AIX system to be monitored:

   ddcapnim01:

   AIXSYS="ddcadora01"    # AIX system to be Monitored
   

2. Create directories on the NFS server to store the WLM and Jtopas trending information for the AIX system to be monitored:

   ddcapnim01:

   mkdir -p /prfdmce0/${AIXSYS}/Top
   mkdir -p /prfdmce0/${AIXSYS}/wlm
   find /prfdmce0/${AIXSYS} -exec ls -ld {} \;
   

3. On the AIX system to be monitored, create shell variables to contain the system name of the AIX system to be monitored and the system name of the NIM server where the monitoring information will be stored:

   ddcadora01:

   AIXSYS="$( hostname )"    # AIX system to be Monitored
   AIXSYS="${AIXSYS%%.*}"
   NFSSRV="mdcapnim01"    # NIM Server with NFS exported filesystem for data storage
   print ${AIXSYS}
   print ${NFSSRV}
   

4. On the AIX system to be monitored, create the NFS mount point to be automatically mounted at boot time:

   ddcadora01:

   /usr/sbin/mknfsmnt -f /prfdmce0 -d /prfdmce0 -h ${NFSSRV} -M sys -B -A -t rw -w bg -Y -X -S -j -q -g
   grep -p "/prfdmce0:" /etc/filesystems
   df
   

5. On the AIX system to be monitored, modify the "/etc/inittab" to contain the startup commands for WLM and for Jtopas:

   ddcadora01:

   rmitab ptxwlm
   rmitab ptxtrend
   mkitab "ptxwlm:2:once:/usr/bin/xmwlm -d /prfdmce0/${AIXSYS}/wlm -n xmwlm > /dev/null 2>&1 # Start xmwlm"
   mkitab "ptxtrend:2:once:/usr/bin/xmtrend -f /etc/perf/jtopas.cf -d /prfdmce0/${AIXSYS}/Top -n jtopas > /dev/null 2>&1 # Start trend"
   egrep "^ptxwlm:|^ptxtrend:" /etc/inittab
   

6. On the AIX system to be monitored, stop any existing "xmwlm" or "xmtrend" processes, then manually start the WLM and Jtopas trending processes or reboot the system and verify the processes are started. To manually start the WLM and Jtopas trending processes:

   ddcapora01:

   for PID in  $( ps -ef -F pid,args | egrep 'xmwlm|xmtrend' | grep -v grep | awk '{ print $1 }' )
   do  
       print "Killing process ${PID}"
       kill ${PID}
       sleep 2
       kill -9 ${PID}   
   done
   ps -ef -F pid,args | egrep 'xmwlm|xmtrend'
   
   nohup /usr/bin/xmwlm -d /prfdmce0/${AIXSYS}/wlm -n xmwlm > /dev/null 2>&1 &
   sleep 2
   nohup /usr/bin/xmtrend -f /etc/perf/jtopas.cf -d /prfdmce0/${AIXSYS}/Top -n jtopas > /dev/null 2>&1 &
   sleep 2
   find /prfdmce0/${AIXSYS} -exec ls -ld {} \;
   jobs
   

7. Cleanup any shell variables created during the execution of this procedure:

   ddcadora01:

   umount /prfpmce0
   /usr/sbin/rmnfsmnt -f /prfdmce0 -B
   unset AIXSYS
   unset NFSSRV
   

   ddcapnim01:

   unset AIXSYS
   

This file last modified 02/04/09