DB2 working distributed on Windoes NT Workstations

DB2 working distributed on Windoes NT Workstations

Post by Michae » Sun, 08 Sep 2002 18:16:42



Hi,

I am trying to install DB2 in a distributed environment. I have six PCs with
Windows NT in a network. They are supposed to execute a DB2 database as a
cluster. So far all my attempts have been in vain.

Does anybody know how to do that or where in the internet I can find a
tutorial about this problem? Maybe there is a website that deals with this
kind of problems?

Thanks,
Michael.

 
 
 

DB2 working distributed on Windoes NT Workstations

Post by Bob Harbu » Tue, 10 Sep 2002 01:58:14


You may wish to start with the Quick Beginnings Guide for DB2 UDB Enterprise -
Extended Edition which you can find on the IBM site:
http://www-3.ibm.com/cgi-bin/db2www/data/db2/udb/winos2unix/support/v...

Not sure what you have or have not read at this point. But that document is the
starting point.
--

Bob Harbus
IBM DB2 UDB Data Management Consulting Services
IBM Toronto Lab


> Hi,

> I am trying to install DB2 in a distributed environment. I have six PCs with
> Windows NT in a network. They are supposed to execute a DB2 database as a
> cluster. So far all my attempts have been in vain.

> Does anybody know how to do that or where in the internet I can find a
> tutorial about this problem? Maybe there is a website that deals with this
> kind of problems?

> Thanks,
> Michael.


 
 
 

DB2 working distributed on Windoes NT Workstations

Post by Mark Yudki » Tue, 10 Sep 2002 15:17:35


If by "in a cluster" you mean running a DB2 over a clustered Intel system,
you will need to install DB2 EEE. The Quick Beginnings book for that edition
contains the necessary setup information.

The Admin Guide, chapter 34 contains information on clustering for high
availability.

BTW, I would strongly recommend updating to W2K AS or, even better, W2K DC
rather than using NT4 for a tight clustering environment (i.e. one involving
the use of system interconnects).


Quote:> Hi,

> I am trying to install DB2 in a distributed environment. I have six PCs
with
> Windows NT in a network. They are supposed to execute a DB2 database as a
> cluster. So far all my attempts have been in vain.

> Does anybody know how to do that or where in the internet I can find a
> tutorial about this problem? Maybe there is a website that deals with this
> kind of problems?

> Thanks,
> Michael.

 
 
 

1. Distributed DB2 - Oracle/Oracle - Oracle/DB2 - DB2/DB2

Sorry for the strange subject line.  I am working on a system that has Oracle
running on an RS/6000 environment, which must interface with DB/2 on an OS/390
MVS environment (version 4.x).   There is an interface job running from oracle
over the Oracle Transparent Gateway for DB2.   Unfortunately, it is reading one
very important table for a little bit too long, causing a timeout on processes
trying to write to the table.  We are trying to minimize the time the read lock
is in existance.  Dirty reads and mirroring the table are not options (48
million row partitioned table)..   We have noticed that in a heterogenous
environment, i.e., Oracle /DB2 over gateway, no distributed query optimization
is possible.  In other words, if we join a 1 row Oracle table, with a 48 million
row DB2 table, an attempt is made to bring the 48 million rows over the network
to perform the join on the Oracle side.  (ugh).. therefore, joins are being
coded, which take a lot of overhead. (pass one literal value over to DB2 at a
time)..  With that background information, I am wondering the following:

If I install UDB DB2 6.1 on the Unix, RS/6000 environment, and make this entire
interface a Unix DB2 6.1 to Mainframe DB2 situation, rather than an Oracle to
DB2 situation, will smarter decisions be made by query optimization techniques,
to perform joins on the appropriate side of the distributed system.  If I have a
one row table on Unix DB2 6.1, joining with a 48 million row table on the
mainframe DB2 side, will something in DB2 be clever enough to pull the one row
over to the other side, and do the join there?

I would be interested in any answer to this question, other ideas, further
reading ideas, etc.,   The main goal is of course to hold the read lock for 20
seconds or so, as oppossed to the 2 1/2 minutes it is using now.  I am hoping
with fast join distributed optimization techniques, the time this interface runs
can be minimized to the point where it doesn't disturb any other transactions.
Just a couple of other notes.. Isolation levels cannot be set from the Oracle
side over the transparent gateway as far as I know, and in addition, I don't
even know if that's relevant, because I don't know that much about DB2 isolation
levels.   Also, one other suggestion is that we place all of the Oracle
interface tables on the DB2 side, to accomplish much the same thing..  

Any insights are welcome,

Thank you..

2. log shipping question

3. Trigger working in DB2 for NT but not in DB2 on an OS/390-machine

4. Unlogged transactions

5. DB2 on Windows NT Workstation

6. pick user group in Canada?

7. sql enterprise manger's rigester server doesn't work in a nt workstation

8. Trapping Errors in Stored Procedures

9. Does 6.5 work on NT 4.0 Workstation?

10. Help, GUI tools not working anymore - SQL7 SP3 - Workstation NT 4.0 SP6a

11. ODBC works in w98 not in NT Workstation

12. Will SQL Server work on NT Workstation?

13. does the replication works with NT workstation?