57##-SS1 option 42 HA Journal Performance

We are a BPCS shop.
We are also a Mimix shop.
We were trying to reduce the time of our month end close. Disk bound.
52 disk drives of which 48 are 140GB HDD's and 4 are SSD's.

We did the CHGPF on all BPCS files to say unit(*SSD), saved the library,
deleted it, and restored it to ensure they were on the SSD's.
Month end improved, but not nearly as much as we hoped.

Then we installed 57##-SS1 option 42 HA Journal Performance. Followed
recommended instructions to use journal caching attributes, talked to
Mimix and got it all set up. Reapplied existing cume and no further PTF's
were installed therefore we could install this LPP without having to do a
disruptive IPL. Put it on a day or so before month end to allow us three
month ends to test before 70 day license key expired.
In case you stink at math try this:
STRSQL
values days('2010-06-02') - days('2010-03-30')
returns 64, which is less than 70

Ran month end. Wow! Like, freaking wow!

Process Jan, pre anything Feb, w/SSD March, w/opt 42 + SSD
CST600C 90 minutes 89 minutes 56 minutes
INV900C 115 minutes 108 minutes 52 minutes
of which
INV900 98 minutes 87 minutes 35 minutes

According to our journals the number of transactions beaten into ITH by
INV900 is 29,940,735 upon 18,975 items.

Got a lot more bang for the buck with the option 42 versus the SSD's
***IF*** you are journalling.





Rob Berendt