Last week after upgrading to 6.1 I ran STROBJCVN on each library instead of
doing one for the entire system. It was a quick cl to submit one STROBJCVN
for each library. I submitted them into a job queue with 6 active jobs into
the *SBSD. ANZOBJCVN said it would take 33 hours to do the conversion, and
doing it this way took about 14 hours. A good savings, and then I could see
for each library if it ended abnormally. This was so much better than doing
one STROBJCVN command and having to look through a mess of information.
Pete
Pete Massiello
iTech Solutions
http://www.itechsol.com
-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx
[mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of rob@xxxxxxxxx
Sent: Tuesday, May 20, 2008 2:46 PM
To: midrange-l@xxxxxxxxxxxx
Subject: STROBJCVN
Ok, so, you've passed ANZOBJCVN, upgraded to V6R1, and now run STROBJCVN
so your users do not pay the "first touch" penalty. I think we discussed
this on a thread but I can't remember. Anyway, when you run STROBJCVN you
might want to change your system value QMAXSPLF first. In fact, due to an
issue I had, the redpiece on STROBJCVN will be upgraded to suggest this.
Apparently IBM opens a spool file for each object they convert, but rarely
(if ever) writes to the spool file. The spool file gets closed and
immediately goes to a FIN state. However, it still counts as one of the
spool files controlled by QMAXSPLF. Another suggestion is to submit a
STROBJCVN for each user library instead of one big job for *ALLUSR. I
wrote a program that determines if a library is a *ALLUSR and fires off a
STROBJCVN for that library (giving the library name as the job name, etc).
If you still have enough objects in a single library you still may have
to modify QMAXSPLF.
IBM will not change the logic to only generate the spool file if needed.
They are still working on the issue that a particular message id doesn't
"trim" one of the parameters passed to the message and thus ties up over a
whole page of blank space in the joblog in the middle of the message.
Failure to change QMAXSPLF, or break down STROBJCVN into manageable pieces
may result in erroneous messages that numerous objects would not convert.
When the only real problem was that it couldn't generate the false spool
file.
Note: Even Domino recommends you run STROBJCVN on the IBM Domino LPP
libraries after the upgrade to avoid the first touch penalty when firing
up Domino.
Gosh, I hope IBM isn't doing something real stupid like
DSPOBJD OUTPUT(*pRINT)
CPYSPLF...
DLTSPLF
But something made me think it was more of a SQL issue because of some
other message in the joblog that I can't remember.
Rob Berendt
midrange.com
