ca.portal.admin

Re: Installing the R16 SP4 maint tape

Discussion created by ca.portal.admin on Jan 23, 2007
Dave,

For the SMP/E RECEIVE job, if your system is at a level earlier than SP4
(Genlevel 0603) you need to modify the job (i.e. uncomment the steps in
the job) to receive each missing service pack, e.g. if you are at SP2
(Genlevel 0502), you need to receive the maintenance for SP3 (Genlevel
0509)and SP4 (Genlevel 0603)

Restore all SYSMODS indicated by the GJG0MAPC job and restore all
USERMODS which contains JCLIN.

APPLY SOURCEID for each missing service pack e.g. (GJSP4 GJSP3).

You will need to apply the following fixes for Express Reorg in SP4:
QO81105,QO83987,QO84337,TE4C074,T5B0056

We attempted to run Express Reorg on a large database (6.5 GB). We
encountered problem in calculating the required space for the work
files. During the reload phase, we encounter a large number of E37-08
abends, even though we have specified a 15 3390-3 DASD volumes to
accommodate the data. The DASD were temporarily made available to us and
are not managed by SMS which would have made life a lot easier for space
management. We have to assign multiple volumes when defining files.
We have raised this issue with CA and they are in the process of
enhancing Express Reorg to address this problem, but unfortunately, this
will not be available until IDMS release 17.0 becomes generally
available (estimated Q1 2008).

There is no good way to accurately identify the needed work file sizes.
CA Developer's recommendation is to run with NULLFILE through the UNLOAD
step and this may give some guidelines to the size requirements of the
RELOAD datasets. The other thing would be to perform the function
against a smaller test DB and then try to extrapolate to your production
DB.
However that can be way out depending on the relative densities of the
test and production DBs and the amount of overflow processing that may
occur.

We have successfully used the NULLFILE technique to assist is
calculating the work file sizes for the unload phase, which uses a total
of 5 3390-3 volumes. But during the reload phase, the utility continues
to increase this space requirement significantly across 70 work files
without any consistency in each files sizing.

We finally managed to run Express Reorg against the production database
by using Virtual Tape for some intermediate work files after many
attempts to find sufficient work space. The use of Virtual Tape does not
impede too much on the I/O speed. In total the unload and reload steps
took some 6-8 hours for 6.5 GB database

Regards,
Paul Mak
EDS | Australia ADU | Data Engineering Capability - DBA |
Level 1, 36-46 George Street, Burwood, NSW 2134 |
Office: +61 2 9012 5434 | Mobile: +61 419 398 116 |
Fax: +61 2 9012 6707 | Email: paul.mak@eds.com |

Outcomes