db2/UDB replication problems (maybe ...?) - Help needed

Andy K

Hi ,

I have an existing database (on a Linux Red Hat ES3) where the
replication function exists but does not work because the target
database has been shutdown months ago .

My question is what will happens if i start the target database ?

Will there be problems with the source database or the target database
(decreasing performance or other kinds of problem )?

For info , the database is 6,7 Go (or 6.7 Go in US measure) . The
source database is on a server and the target database on an another
server (physically) . The target database is shutdown .

Any help is welcome .

Andy the beginner in DB2 .

ps1 : is there any possibilities to setup a replication processus
without the DB2 control center ? (it seems to be no after the
documents I have read but who knows ... )

ps2 : that control center seems to be quite annoying especially if the
linux server has been secured and doesn't allow graphical screening !

Nov 12 '05 #1

Jurgen Haan

Andy K wrote:

Hi ,

I have an existing database (on a Linux Red Hat ES3) where the
replication function exists but does not work because the target
database has been shutdown months ago .

My question is what will happens if i start the target database ?

Will there be problems with the source database or the target database
(decreasing performance or other kinds of problem )?
Seems to me they will be sync-ed, but I could be wrong.
Since this is what replication is intended for, there shouldn't be too
much performance problems. But to be sure you better wait for a response
from someone who actually knows what's he's talking about.
For info , the database is 6,7 Go (or 6.7 Go in US measure) . The
source database is on a server and the target database on an another
server (physically) . The target database is shutdown .

Any help is welcome .

Andy the beginner in DB2 .

ps1 : is there any possibilities to setup a replication processus
without the DB2 control center ? (it seems to be no after the
documents I have read but who knows ... )
It seems to me that everything that can be done with CC can also be done
commandline.
ps2 : that control center seems to be quite annoying especially if the
linux server has been secured and doesn't allow graphical screening !

Try to use trusted X forwarding over SSH (ssh -Y). (Need to activate X
forwarding in SSHD_CONFIG though). Works for me.

PS. I'm not that familiar with Replication, so alas, I can't help on the
other topics.

Nov 12 '05 #2

Norbert Munkel

Hi Andy,

Andy K wrote:

I have an existing database (on a Linux Red Hat ES3) where the
replication function exists but does not work because the target
database has been shutdown months ago .

My question is what will happens if i start the target database ?
Nothing bad will happen by just starting the database. If you start the
apply-process and there is remaining (not pruned) data in the cd-tables,
this data will be applied.
Will there be problems with the source database or the target database
(decreasing performance or other kinds of problem )?
as usual...it depends (amount of data to be applied, other applications
running against the database)....

Was the capture-process on the source-database running all the time?
Then you should have a reasonable amount of data in the cd-tables. If
you have stopped the capture-process you would need to have all logfiles
written by the source database since capture has been shut down. Else,
capture will not work properly and you would need to do a
(manual/automatic) full refresh.
ps1 : is there any possibilities to setup a replication processus
without the DB2 control center ? (it seems to be no after the
documents I have read but who knows ... )
1. Learn by the scripts, Replication Center creates for you. You can
easily modify them or write your own wrappers around them.
2. "SQL-Replication Guide and Reference" is the source you are looking for.
3. short list of commands you need to look up to run replication from
the command-line: asncap, asnapply, asnacmd, asnccmd, asnanalyze.
ps2 : that control center seems to be quite annoying especially if the
linux server has been secured and doesn't allow graphical screening !

I understand you are running the control center on the server where the
database resides. This is not optimal in many ways.

If the Administration Server is running on the server you should be able
to run the graphical tools on your client without any X-forwarding at all.

Yes, control center _is_ annoying like any other graphical java
interface that I have seen so far, but it is getting better from release
to release . Fortunately there are many ways to run and monitor DB2 UDB
without running a gui at all. ;-)

regards,

Norbert

Nov 12 '05 #3

Norbert Munkel

Hi again,

Norbert Munkel wrote:

Nothing bad will happen by just starting the database. If you start the

^^^^^^^^^^^

If your replication is set up to do allow automatic full refreshs, full
loads/imports of the replicated tables may occur including possible
performance issues (fullselect on the source-tables, network-activity....).

regards,

Norbert

Nov 12 '05 #4

Andy K

Hi Norbert ,

Thanks for the insight . I'm going to do my homework for DB2 (glups !)
..

Btw is there catalog view for replication ?

In Oracle (1st love) , there is a couple of dynamic views like
dba_snapshots or ALL_SNAPSHOT_REFRESH_TIMES that can be really helpful
.. Something similar with DB2 ?

Regards,
Andy

Nov 12 '05 #5

Gert van der Kooij

In article <11*********************@g44g2000cwa.googlegroups. com>,
Andy K (ak****@yahoo.com) says...

Hi Norbert ,

Thanks for the insight . I'm going to do my homework for DB2 (glups !)
.

Btw is there catalog view for replication ?

In Oracle (1st love) , there is a couple of dynamic views like
dba_snapshots or ALL_SNAPSHOT_REFRESH_TIMES that can be really helpful
. Something similar with DB2 ?

Regards,
Andy

Hi Andy,

If you want to create the replication definitions from the
commandline you can use the asnclp command. You can download the
manual from http://publibfp.boulder.ibm.com/epubs/pdf/c1894100.pdf

Nov 12 '05 #6

Andy K

Hi Gert ,

Thanks for the documentation . But asnclp is not available on "v7" for
DB2 .

Any suggestion is HIGHLY welcome ! :-D

Regards,
Andy

Gert van der Kooij a écrit :

In article <11*********************@g44g2000cwa.googlegroups. com>,
Andy K (ak****@yahoo.com) says...

Hi Norbert ,

Thanks for the insight . I'm going to do my homework for DB2 (glups !) .

Btw is there catalog view for replication ?

In Oracle (1st love) , there is a couple of dynamic views like
dba_snapshots or ALL_SNAPSHOT_REFRESH_TIMES that can be really helpful . Something similar with DB2 ?

Regards,
Andy

Hi Andy,

If you want to create the replication definitions from the
commandline you can use the asnclp command. You can download the
manual from http://publibfp.boulder.ibm.com/epubs/pdf/c1894100.pdf

Nov 12 '05 #7

Gert van der Kooij

In article <11**********************@g47g2000cwa.googlegroups .com>,
Andy K (ak****@yahoo.com) says...

Hi Andy,

If you want to create the replication definitions from the
commandline you can use the asnclp command. You can download the
manual from http://publibfp.boulder.ibm.com/epubs/pdf/c1894100.pdf

Hi Gert ,

Thanks for the documentation . But asnclp is not available on "v7" for
DB2 .

Any suggestion is HIGHLY welcome ! :-D

Regards,
Andy

Hi Andy,

I've created my own batch scripts by tracing the DJRA tool output.
The DJRA tool is written in REXX so I created my own scripts which
(after specifying all needed variables) generates replication
definitions automatically for over 300 tables.
But because DJRA isn't available on V8 anymore it's the wrong choice
if you need to use it on V8 some day. I have to 'migrate' my own
script and start using asnclp when our client migrates to V8.

Nov 12 '05 #8