On 10/31/13 9:11 AM, Luis Rodriguez wrote:
On Thu, Oct 31, 2013 at 11:16 AM, Darryl Freinkelwrote:
CPF3220 "trying to do a RMVM on QTEMP/QSQLTEMP1"
A quick "DSPMSGD CPF3220" seems to imply that your file needs either
a COMMIT or a ROLLBACK...Check your job log.
All just FWiW\FYI only:
That message also, and firstly, mentions that the member is /opened/
[implied... in that same job, else the error would have been for an
inability to allocate the member]. If the issue were due to CmtCtl
being active with pending changes, then the replacement values for &4
and &5 of the msg CPF3220 would indicate the Commit Definition ID for
that work. Just reading the second level text of the message should be
sufficient; i.e. unlikely any need to review more from the joblog.
Almost surely the second level text suggested the member &3 [as
target of the RMVM] "is either currently open or is closed with records
that are not committed for commitment definition *N." where the "*N" is
replacement data for &4, and "The commitment definition identifier is
X'00000000000000000000'." where the hex data is replacement data for &5.
That result would imply the issue is not related to Commitment
Control, just the open; noting: a member could be closed, but still
pending work, just as the message alludes [because if open, that would
have been detected first], but the replacement variables will still get
set if prior implicit or explicit STRCMTCTL scoped to that job had
And a file would have to be journaled to be under isolation. Yet
that temporary source file QSQLTEMP1, that gets created by the SQL
pre-compiler, is unlikely to be journaled and is even less likely to be
opened under isolation.