DB2 FOR LUW / DB2FORLUW / B50 - DB2

db2_kill or failures may result in ipclean being initiated.
Db2 processes may attempt to start before ipclean completes the
clean
up of Db2 resources properly, leading to db2sycs startup to
fail.

****************************************************************
* USERS AFFECTED:                                              *
* ALL                                                          *
****************************************************************
* PROBLEM DESCRIPTION:                                         *
* See Error Description                                        *
****************************************************************
* RECOMMENDATION:                                              *
* Upgrade to Db2 11.5m6fp0 or higher                           *
****************************************************************

Possible work-around being attempted: Edit the TSA monitor
scripts to add a check for the port no longer being listened
before invoking db2gcf to start the instance. Write a script to
kill DB2, check for port still allocated and wait in a loop
until free, then restart DB2 using db2gcf and check port is
opened OK.



Testing work-around output:



Workaround added before db2gcf ?u calls in
/usr/sbin/rsct/sapolicies/db2/db2V111_start.ksh and
/usr/sbin/rsct/sapolicies/db2/hadrV111_start.ksh:



  if [[ -x $INSTHOME/sqllib/bin/db2gcf ]]; then

     ((sato=nln*2+60))



################################################################
##############################

# Defect 372: DB2 fails to open SSL and TCPIP port. Possibly
because defunct processes

#            not yet cleaned up completely by ipclean. Therefore
loop here until

#            port is no longer being used.



     loop_counter=0

     while [[ $loop_counter -lt 20 ]]; do

        port_used=$(netstat -anp|grep 1565|wc -l)

        if [ $port_used -eq 1 ]; then

           log notice "SSL Port still in use. Waiting 10
secs..."

           sleep 10

           ((loop_counter+=1))

        else

           loop_counter=100

        fi

     done



################################################################
##############################

     su - ${DB2INSTANCE} -c "$INSTHOME/sqllib/bin/db2gcf -t
$sato -u -p ${NN?} -i ${DB2INSTANCE?} -L"

     rcg=$?



  else

     log err "$LINENO: Cannot find executable
$INSTHOME/sqllib/bin/db2gcf"

     sleep 30

     rcg=1

  fi

}

****************************************************************
* USERS AFFECTED:                                              *
* ALL                                                          *
****************************************************************
* PROBLEM DESCRIPTION:                                         *
* See Error Description                                        *
****************************************************************
* RECOMMENDATION:                                              *
* Upgrade to Db2 11.5m6fp0 or higher                           *
****************************************************************

First fixed in Db2 11.5m6fp0

Date  - problem reported    :
Date  - problem closed      :
Date  - last modified       :

18.06.2021
15.07.2021
15.07.2021