Recently v621 to v632 adabas migrate and raw to filesystem, and batch processes have increased the time of execution, which is critical.
The slowness is observed when there are many updates.
I read in the documentation that must be enabled with the parameter in the ADANUC unbuffered.
which is the best option:
ALL | CLEAR | The default
What is your recommendation?
Angel Mendoza O
As the documentation says
so you can try UNBUFFERED=ALL, but it usually does not help a lot to tweak just (that) one parameter,
the whole set of ADANUC parameters will have to be reviewed.
Which OS are we talking about here ?
export NUCPLG NUCCLG
$ADADIR/$ADAVERS/adanuc << END_OF_PARAMS >> $DATAB/db004/NUC.LOG 2>&1 &
Are the batch processes update-centric (many buffer-flushes ?) or did read-centric processes slow down as well ?
They are processes update-centric (many buffer-flushes takes a long time)
Did you increase the LBP when upgrading from ADA 6.2 to 6.3, or did you use lbp=2560m with 6.2 as well ?