> -----Original Message-----
> From: IBM Mainframe Discussion List On Behalf Of Gerry Anstey
>
> Anyone any ideas about this:
>
> We have a VSAM KSDS defined to COBOL as VARYING from 586 to
> 2500. The file is updated by writing new records or rewriting
> old ones. We occasionally get a File status 44 when Writing a
> new record We have checked the record length and confirm that
> it is 586, we also use displays to DISPLAY LENGTH OF so that
> we can confirm that the program at run time has the correct
> record length. Anyone any ideas as to any "features" or
> anything we may be encountering here? We're at a bit of a
> loss as to where to look next.
>
> The problem is spurious in the sense that when it happens
> it's a different record being written but other records of
> the minumum length have already been written in the same run.
> However once a record fails it seems to always fail so we
> have been able to recreate in test so any suggestions as to
> debugging ideas would be welcome.
>
> Note the RPL is RC=8 and Reason 6C which is consistant with status 44.
Does this VSAM KSDS have an alternate index with non-unique keys that is
part of the UPGRADE set? If so, consider this (from the Macro Instructions
for Data Sets manual in the DFSMS bookshelf):
===== Begin paste ======
RECLEN was not sufficient to contain the new alternate index key pointer.
With non-unique UPGRADE AIX's, the record is automatically increased in size
each time a record is added to the base cluster and this can cause an
incorrect RECLEN. Make sure the maximum RECORDSIZE on the alternate index is
large enough for all base pointers it must contain.
===== End paste =======
I.e., the maximum recordsize of your AIX might be too small.
-jc-
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
