Re: trouble with keep_alive script

["Babu Koshy Konnayil" <babu at spin ad.jp>]

> 
> 
> I've gotten nocol set up and I have the individual monitors running just
> fine. When I try and use the keep_alive script I get this:
> 
> longterm 38 % ./keepalive_monitors 
> ./keepalive_monitors: test: argument expected
> 
> I've checked all of the variables in keepalive_monitors and I can't find
> anything obviously wrong.
> 
> This is on a Sun SparcServer 1000
> running Solaris 2.6
> 
> Any thoughts from folks?
> 
> Colin E. Johnson | colinj@unm.edu | http://www.unm.edu/~colinj/
> ". . .so I simply point to where I keep the soul I have, right by the
> exit, just next to the blues." --Nick Hornby _High_Fidelity_
> 

I also had this problem, but think I solved this problem by modifying the
following blocks.  

-------Original block from distribution ----------------
if [ -f ${PIDDIR}/${p}.pid ]; then
    PID=`head -1 ${PIDDIR}/${p}.pid`
   else
    PID=""
   fi
   if [ -z $PID ]; then
     pgmname=`basename $p`;
     ps ${PSFLAGS} 2>&1 |grep -v grep |egrep "$pgmname( |\$)" >/dev/null 2>&1
     #ps ${PSFLAGS} |grep -v grep |egrep "${BIN}/${p}([ ]*|\$)" >/dev/null 2>&1
   else
     ps ${PSNUMFLAGS} $PID | grep ${p} >/dev/null 2>&1
   fi
---------------end of Original block -------------------

----------- Modified block ------------------------------
 if [ -f ${PIDDIR}/${p}.pid ]; then
    PID=`head -1 ${PIDDIR}/${p}.pid`
   else
    PID="0"
   fi
   if [ $PID = "0" ]; then
     pgmname=`basename $p`;
     ps ${PSFLAGS} 2>&1 |grep -v grep |egrep "$pgmname( |\$)" >/dev/null 2>&1
     #ps ${PSFLAGS} |grep -v grep |egrep "${BIN}/${p}([ ]*|\$)" >/dev/null 2>&1
   else
     ps ${PSNUMFLAGS} $PID | grep ${p} >/dev/null 2>&1
   fi

-----------end of Modified block

Please note that the change is done on how the PID variable is assigned an initial
value if the process id is not found and how the PID variable is checked.


Regards,
Babu Koshy