NB: This email and its contents are subject to our email legal notice
which can be viewed at http://www.sars.gov.za/Email_Disclaimer.pdf
----
Hi All
Using SRDF and ADABAS to a remote site, our tests on ADABAS were
successful (verified by our DBA's).
I did however queried SAG in South Africa and they came up with the
following:
"When you run with Suspend and Resume, the physical database is in a
stable state. What is in memory is not relevant because all users will
be at
ET state. So, there should be no pending transactional units of work.
Suspend and Resume are available with V742 and above.
The issue with SRDF is that it is in the hands of the users. Depending
on the save options used and, more specifically, what is actually done
during any restore, we cannot accept SRDF as being an integrated
backup/restore mechanism for an active Adabas database.
Issues have been encountered in the past where DASD was switched off
before a buffer flush was completed.
In this case a restore was needed because we could see the start of the
buffer flush and not the end; thus, we
could not guarantee that Data Storage and Asso are still in sync because
of the way in which we write the blocks to the VTOCS.
This is an extraction of a PPRC disaster scenario.
ADA - PPRC disaster scenario question
Symptom:
We are going to use PPRC (Peer to Peer Remote Copy) technology in order
to continue with our production work in the event of a disk crash.
If a crash happens, is there anything we need to do to make sure our
mirrored Adabas database is consistent?
Resolution:
As long as the mirrored data is a consistent real time shadow of primary
data there is no problem to:
1) Start an Adabas nucleus using mirrored ASSO, DATA, and WORK after a
disk crash on primary - Adabas will perform the autorestart.
2) Repair primary data after the hardware is fixed by copying the
mirrored data (completely: ASSO, DATA and WORK).
It is most important for the DBA to understand that--ASSO, DATA, and
WORK--together make a consistent database. We have seen inconsistencies
after sites have used SNAPSHOTS of databases to restore DATA to a status
snapped half an hour ago, but leaving ASSO in its present state. Or
there was one case where ASSO and DATA were snapped after an abend in
Adabas (with pending autorestart on WORK) and were restored leaving the
database inconsistent. The DBA needs to make sure the database is
completely mirrored (all volumes of ASSO, DATA, and WORK) and that these
mirrored datasets are used completely if required.
Symptom:
Does Adabas support IBM's Peer to Peer Remote Copy (PPRC) technology?
Resolution:
PPRC - Peer to Peer Remote Copy - is a feature of the control unit
(3990-6) which should be transparent to all applications. PPRC is
supported with Adabas, and is transparent to Adabas. On the other hand,
Adabas does not exploit that feature in any way".
Kind Regards,
Sarel Swanepoel
Capacity Management
South African Revenue Services
Office: +27 (0)12 422 5033
Mobile: +27 (0)82 4927 321
Fax: +27 (0)12 422 6068
Email: [EMAIL PROTECTED]
Web Site:
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
