Re: CPF7091 - space for entry cannot be allocated on journal receiver

johnking@xxxxxxx wrote:

Hmmm... 1TB is what an initial reading of the manual led me to
believe also, but further excavation into the details shows there is
one more layer of complexity. The 1TB limit is good _only_ if the
journal has one of the *MAXOPT values specified on the RCVSIZOPT
parameter.

You can specify whatever you like for the receiver's THRESHOLD value,
but if the journal lacks the *MAXOPT value the system will throw
error messages at your operator as soon as the receiver's size
approaches 1.9G.

To keep this fix as simple as possible I'll try two options on the
CHGJRN:
1) set RCVSIZOPT = *RMVINTENT to reduce the number of
'internal' entries in the receiver.
2) set RCVSIZOPT = *MAXOPT1 or *MAXOPT2
to allow the 1TB receiver size

Must make sure these changes don't affect the layout of the *outfile
of DSPJRN before putting them in production. Guess I know what I'll
be doing this evening...

Not to imply testing is worthless... But I can write with effective certainty, that the size options will not impact the file layout. The layout is established by what is specified on the parameters of DSPJRN. Given no changes to the DSPJRN request, the size changes will not be manifest as differences in the output file, even if *CALC was specified or defaulted on any of the output file layout related parameters.

Even with larger size options, keeping THRESHOLD(*NONE) will leave the same condition as an eventual outcome, without having added monitoring for the warnings that the receiver is going to be full due either to the maximum sequence or maximum physical size limit. Making the journal system managed would put the onus on the system to effect the change, instead of writing an application or having an operator respond to the warnings. However with the larger maximum size option, beyond just system managed, it may be prudent to actually specify a threshold due to the larger maximum physical size.

The default for CRTJRN receiver size option has long since moved to one of the higher number *MAXOPT values, specifically to avoid the limitations of the journals that were being created versions ago. In V5R4 the newest default became *SYSDFT to effect same as RCVSIZOPT(*MAXOPT2 *RMVINTENT)
http://publib.boulder.ibm.com/infocenter/iseries/v5r4/topic/rzaki/rzakiwhatsnew.htm
http://publib.boulder.ibm.com/infocenter/iseries/v5r4/topic/rzaki/rzakisizeoptions.htm

Regards, Chuck