Cloned a system from (say) LP01 to a new LPAR LP02. both are LOCAL mode (i.e. non-SYSPLEXed).
This was achieved in the normal fashion by volume copies, dataset copies, the usual renames, manual editing etc.
RACF ICHRDSNT reflects correct RACF database names in each LPAR (in USER.LINKLIB).
LP01 IPL's straight through - no action message for RACF displayed.
LP02 IPL's OK but displays:
*ICH588A CONTINUED USE OF DATABASE
SYS1.LP02.RACFDS ON VOLUME Z2SCAT
CAN RESULT IN DATABASE CORRUPTION AND SYSTEM OUTAGE.
IF ANY SYSTEM IS USING THE DATABASE IN DATA SHARING MODE, AND ANY
OTHER SYSTEM CONCURRENTLY USES IT IN NON-DATA SHARING MODE,
DATABASE CORRUPTION WILL RESULT.
YOU ARE INITIALIZING INTO DATA SHARING MODE. PROFILE
IRRPLEX_ADCDPL , IN CLASS GXFACILI INDICATES NON-DATA SHARING MODE.
IF THE DATABASE IS NOT BEING USED BY SYSTEMS OUTSIDE OF THIS
SYSPLEX, AND IS NOT BEING USED BY SYSTEMS IN THIS SYSPLEX IN NON-DATA
SHARING MODE, THEN SPECIFY 'CONTINUE'. OTHERWISE SPECIFY 'NODATASHARE'
AND THE SYSTEM WILL INITIALIZE IN NON-DATA SHARING MODE.
*01 ICH600A VALID RESPONSES ARE 'CONTINUE' OR 'NODATASHARE'
Normally reply NODATASHARE but I understand CONTINUE would probably also be OK in the case of two LOCAL LPARs?
On both systems, I have the same classes defined under RACF -
resource class is GXFACILI
profile names IRRPLEX_ADCDPL and IRRPLEX_LOCAL with same characteristics.
To confirm that, command "SR CLASS(GXFACILI) shows
IRRPLEX_ADCDPL
IRRPLEX_LOCAL
on both systems
Now, I found on Google that RACF "self-protects" itself when the database(s) are copied in such a cloning situation.
But what can I do top safely stop the message appearing at IPL whilst bringing both LPAR's RACFs up in NONDATASHARING modes?
Ideally, of course, I would just like the IRRPLEX_LOCAL to be used.
Obviously the IRRPLEX_ADCDPL comes from an earlier time of ADCDs when the system was originally built some years ago.
But can I just delete IRRPLEX_ADCDPL leaving IRRPLEX_LOCAL on both systems?
Joined: 06 Jun 2008 Posts: 8696 Location: Dubuque, Iowa, USA
The manual explanation for ICH588A has a set of steps to take when this message appears. Have you tried the D XCF,SYSPLEX and D XCF,GROUP,IRRXCF00 commands mentioned in the manual?
D XCF,SYSPLEX
IXC336I 21.46.33 DISPLAY XCF 732
SYSPLEX LOCAL
SYSTEM TYPE SERIAL LPAR STATUS TIME SYSTEM STATUS
LP01 3907 58A7 01 08/02/2018 21:46:32 MONOPLEX MODE TM=SIMETR
SYSTEM STATUS DETECTION PARTITIONING PROTOCOL CONNECTION EXCEPTIONS:
PROTOCOL NOT APPLICABLE IN MONOPLEX MODE
D XCF,GROUP,IRRXCF00
IXC340I 21.47.50 DISPLAY XCF 734
GROUP 'IRRXCF00' IS NOT DEFINED TO THIS SYSPLEX
Which is maybe what you'd expect in a LOCAL/Monoplex mode?
On LP02 LPAR (the one with the IPL message) you get:
Code:
D XCF,SYSPLEX
IXC336I 21.51.20 DISPLAY XCF 123
SYSPLEX LOCAL
SYSTEM TYPE SERIAL LPAR STATUS TIME SYSTEM STATUS
LP02 3907 58A7 02 08/02/2018 21:51:20 MONOPLEX MODE TM=SIMETR
SYSTEM STATUS DETECTION PARTITIONING PROTOCOL CONNECTION EXCEPTIONS:
PROTOCOL NOT APPLICABLE IN MONOPLEX MODE
further interesting (?) information:
On LP01, RVARY LIST shows:
Code:
ICH15013I RACF DATABASE STATUS:
ACTIVE USE NUM VOLUME DATASET
------ --- --- ------ -------
YES PRIM 1 Z22CAT SYS1.RACF
YES BACK 1 Z22CAT SYS1.RACFSEC
ICH15020I RVARY COMMAND HAS FINISHED PROCESSING.
on LP02, RVARY LIST shows:
Code:
ICH15013I RACF DATABASE STATUS:
ACTIVE USE NUM VOLUME DATASET
------ --- --- ------ -------
YES PRIM 1 Z2SCAT SYS1.LP02.RACFDS
YES BACK 1 Z2SUSS SYS1.LP02.RACFDS.BACKUP
MEMBER LP02 IS SYSPLEX COMMUNICATIONS ENABLED & IN NON-DATA SHARING MODE.
ICH15020I RVARY COMMAND HAS FINISHED PROCESSING.
Note the extra line for LP02! But why? Where should I look?!
Coded for you - do it yourself next time
All hinges on the ICHRDSNT (RACF resident tables) module that is compiled to primarily state the RACF database names?
(Unless you have zOS v2.3 in which case you can take advantage of the new IRRPRMxx PARMLIB member)
When I setup LP02, I recompiled this module as follows:
ICHRDSNT CSECT
DC AL1(1) one RACF database
DC CL44'SYS1.LP02.RACFDS' primary RACF db
DC CL44'SYS1.LP02.RACFDS.BACKUP' backup RACF db
DC AL1(255) number of resident data blocks
DC X'8C' updates duplicates on backup db
END
/*
//* Note the last flag byte!
//* Description of flag byte bits...
//* x'1... ....' - duplicate updates to backup
//* x'.1.. ....' - duplicate statistical updates to backup
//* x'..x. ....' - reserved
//* x'...x ....' - reserved
//* x'.... 1...' - enable for sysplex communications
//* x'.... .1..' - default mode
//* 0 = non data sharing
//* 1 = data sharing
//* x'.... ..x.' - reserved
//* x'.... ...x' - reserved
//*
The X’8C’ is the important bit here?
Says “allow for SYSPLEX operations (if applicable)” and also - more importantly - "set data sharing on as default”.
I’ve recompiled this to remove the sysplex ability (we are running Local/Monoplex) and data sharing facility to use X’80’ instead, so next IPL of LP02 shouldn’t ask the question!
