I setup a userexit program and cannot get it to work correctly:

get the following errors from DB2

Time of Error: Mon Jan 12 14:15:45 2004

Parameter Count: 1
Parameters Passed:
Audit Log File:
System Call Parms:
Media Type: disk
User Exit RC: 20

> Error isolation: Action is not valid

Steps I have taken to get this far: DB2 UDB 7.2 FP6 on AIX 5.2L

# db2 get db cfg for acctview |egrep -i "log retain|user exit"
Log retain for recovery status = RECOVERY
User exit for logging status = NO
Log retain for recovery enabled (LOGRETAIN) = RECOVERY
User exit for logging enabled (USEREXIT) = OFF


# db2 "update db cfg for acctview using USEREXIT ON"
DB20000I The UPDATE DATABASE CONFIGURATION command completed successfully.

- Copy over sample user exit program:

cd /home/db2inst1/sqllib/samples/c

mkdir /home/db2inst1/work/userexit/scripts

cp HOME/sqllib/sample/c/db2uext2.cdisk /home/db2inst1/work/userexit/scripts/db2uext2.c


- make required changes to sample user exit c program:

#define ARCHIVE_PATH "/home/db2inst1/u/dbarch"
#define RETRIEVE_PATH "/home/db2inst1/u/dblog/"
#define AUDIT_ACTIVE 1 /* enable audit trail logging */
#define ERROR_ACTIVE 1 /* enable error trail logging */
#define AUDIT_ERROR_PATH "/home/db2inst1/u/dblog/"
#define AUDIT_ERROR_ATTR "a" /* append to text file */
#define BUFFER_SIZE 32 /* # of 4K pages for output buffer */



- create additional directories for user exit

/home/db2inst1/work/userexit/scripts >

# ls
db2uext2.c

/home/db2inst1/work/userexit/scripts >

# cd $HOME

/home/db2inst1 >

# mkdir u
/home/db2inst1 >

# cd u
home/db2inst1/u >

# mkdir dbarch
/home/db2inst1/u >

# cd dbarch
/home/db2inst1/u/dbarch >

# pwd
/home/db2inst1/u/dbarch
/home/db2inst1/u/dbarch >

# cd ..
/home/db2inst1/u >

# mkdir dblog
/home/db2inst1/u >

# pwd
/home/db2inst1/u
/home/db2inst1/u >

# cd dblog
/home/db2inst1/u/dblog >

/home/db2inst1/u/dblog >


# mkdir -p /MYDB/NODE0000
/home/db2inst1/u/dbarch >

- compile db2uext2.c program

cd /home/db2inst1/work >

# cc -o db2uext2 db2uext2.c

# chmod +s db2uext2

/home/db2inst1/work >

# ls
db2uext2 db2uext2.c userexit


- check to see if it works

# cd /home/db2inst1/u
/home/db2inst1/u >

# ls
dbarch dblog

review and fix errors in error file:


stop/start DB2 instance:

# db2stop
SQL1064N DB2STOP processing was successful.
db2inst1@av3: /home/db2inst1/work >

# db2start
SQL1063N DB2START processing was successful.
db2inst1@av3: /home/db2inst1/work >


try running db2uext2 again

/home/db2inst1/u/dblog >
# ls
USEREXIT.ERR

************************************************** ***************************
Time of Error: Mon Jan 12 14:15:45 2004
************************************************** ***************************
Time of Error: Mon Jan 12 14:15:45 2004

Parameter Count: 1
Parameters Passed:
Audit Log File:
System Call Parms:
Media Type: disk
User Exit RC: 20

> Error isolation: Action is not valid


************************************************** ***************************
Time of Error: Mon Jan 12 14:22:31 2004

Parameter Count: 1
Parameters Passed:
Audit Log File:
System Call Parms:
Media Type: disk
User Exit RC: 20

> Error isolation: Action is not valid

# cat USEREXIT.ERR
************************************************** ***************************
Time of Error: Mon Jan 12 14:15:45 2004

Parameter Count: 1
Parameters Passed:
Audit Log File:
System Call Parms:
Media Type: disk
User Exit RC: 20

> Error isolation: Action is not valid


************************************************** ***************************
Time of Error: Mon Jan 12 14:22:31 2004

Parameter Count: 1
Parameters Passed:
Audit Log File:
System Call Parms:
Media Type: disk
User Exit RC: 20

> Error isolation: Action is not valid


/home/db2inst1/u/dblog >
#

did this fail because I am not compiling the db2uext2.c program under the sqllib/adm directory? Do I need to do this as the AIX root superuser? Thanks!

Scott

You will have to copy the db2uext2 executable to a location where db2 can find and sqllib/misc is the best location ...

OK figured out what i did wrong and was able to recompile the db2uext2.c file into an executable.

# cc -o db2uext2 db2uext2.c
/home/db2inst1/work/userexit >

# chmod +s db2uext2
/home/db2inst1/work/userexit >

Archive Log testing:

# db2 archive log for database MYDB
DB20000I The ARCHIVE LOG command completed successfully.

/home/db2inst1/work/userexit >

So how do I know for sure that my logs are correctly being archived via the user exit program? Thanks!

Scott

You can issue an

archive log for database <dbname>

and that should close a log.

Look in the diag - if it could not archive it should say so. Also look in the archive path you should see log/s moved to this directory after the command completes.