We use redirected restore to refresh test from prod on different AIX 5.2
servers. The databases have the same number, size and names of cooked
containers. The only difference is the fully qualified path contains a
different "SID": eg. "/db2/PRD/data1/..." versus "/db2/TST/data1/...".
(Please note as an SAP installation I cannot change this layout).
We use "set tablespace container" statements to redefine the fully
qualified container names. During the restore the existing containers are
dropped and new containers are created. This process takes a long time.
My question is, can we somehow avoid re-creating these containers. For
example, what would happen if we just created a symbolic link PRD->TST so
that the restore would find the existing containers? (I've done this
successfully using Informix raw devices.)
Or, do the containers need to be re-created for a valid restore?
Thanks,
Chris
--
Message posted via http://www.dbmonster.com 7 6151
CCC via DBMonster.com wrote: We use redirected restore to refresh test from prod on different AIX
5.2 servers. The databases have the same number, size and names of cooked containers. The only difference is the fully qualified path contains
a different "SID": eg. "/db2/PRD/data1/..." versus
"/db2/TST/data1/...". (Please note as an SAP installation I cannot change this layout).
We use "set tablespace container" statements to redefine the fully qualified container names. During the restore the existing containers
are dropped and new containers are created. This process takes a long
time. My question is, can we somehow avoid re-creating these containers.
For example, what would happen if we just created a symbolic link
PRD->TST so that the restore would find the existing containers? (I've done this successfully using Informix raw devices.)
Or, do the containers need to be re-created for a valid restore?
Thanks, Chris
-- Message posted via http://www.dbmonster.com
UDB does not allow links for restores. As a matter of fact, you cannot
even create a database using containers that are links.
"set tablespace containers...." command simply re-defines containers.
When you do a "restore db DBName continue", this is when the restore
process actually starts. Restore is still a restore and takes its sweet
time regardless -- whether it be regular or redirected restore. If UDB
were to allow links, I don't think you will see any significant
difference in restore time.
BTW, I agree with you -- Informix does have this flexibility.
Keith Ponnapalli
IBM Certified Advanced Database Administrator - DB2 UDB V8.1 for
Linux,UNIX, and Windows
INFORMIX Certified Database Administrator
Thank you. So "links" aren't a valid workaround. Maybe you could educate
me a bit. (I'm new to DB2, but do have Informix & Oracle experience).
When I simply restore a database (not "redirected"), I don't believe it
recreates the containers; it just starts restoring the data into the
existing containers. However, the redirected restore does rebuild DMS
containers. I can watch them being deleted and regenerated. I assume this
is a result of the "set tablespace containers". The transfer of data from
tape doesn't begin until all the containers are rebuilt. In fact we had to
increase TSM timeout settings to allow for this finish.
Am I doing something wrong here?
Chris
--
Message posted via http://www.dbmonster.com
"CCC via DBMonster.com" <fo***@nospam.DBMonster.com> wrote in message
news:08******************************@DBMonster.co m... We use redirected restore to refresh test from prod on different AIX 5.2 servers. The databases have the same number, size and names of cooked containers. The only difference is the fully qualified path contains a different "SID": eg. "/db2/PRD/data1/..." versus "/db2/TST/data1/...". (Please note as an SAP installation I cannot change this layout).
We use "set tablespace container" statements to redefine the fully qualified container names. During the restore the existing containers are dropped and new containers are created. This process takes a long time.
My question is, can we somehow avoid re-creating these containers. For example, what would happen if we just created a symbolic link PRD->TST so that the restore would find the existing containers? (I've done this successfully using Informix raw devices.)
==============
Yes, you can use symbolic link, you can also use redirected restore, both
should work in your scenario. Or, do the containers need to be re-created for a valid restore?
===========================
DB2 restore will check if the tbspaces/containers are existed, if not, db2
will build them first, then restore the data. Thanks, Chris
-- Message posted via http://www.dbmonster.com
Please help me clarify a couple of questions:
Because I am cloning a source onto a target which has a different
DB2INSTANCE, database name, host, etc, I assume I must use a "redirected
restore". Just because I have links that make the target filesystem "look"
like the source system, I cannot do a regular restore. Is this correct?
Is the deletion and recreation of the exisiting containers I observe during
the redirected restore a result of the "set tablespace containers"
statements?
If this is true, given links that make the filesystems look the same, can I
do the redirected restore without using the "set tablespace constainers"
statements?
This would save me 2-3 hours on my 300GB database.
Thanks
Chris
--
Message posted via http://www.dbmonster.com
CCC via DBMonster.com wrote: Please help me clarify a couple of questions:
Because I am cloning a source onto a target which has a different DB2INSTANCE, database name, host, etc, I assume I must use a
"redirected restore". Just because I have links that make the target filesystem
"look" like the source system, I cannot do a regular restore. Is this
correct? Is the deletion and recreation of the exisiting containers I observe
during the redirected restore a result of the "set tablespace containers" statements?
If this is true, given links that make the filesystems look the same,
can I do the redirected restore without using the "set tablespace
constainers" statements?
This would save me 2-3 hours on my 300GB database.
Thanks Chris
-- Message posted via http://www.dbmonster.com
I guess I goofed on this one. Links will work. You can use 'restore
from ...into targetDB ... ' , so it doesn't matter if it it's in
another instance/db.
I'd make sure links are supported/recommeneded by both IBM/SAP just in
case if something were to go wrong.
K
"CCC via DBMonster.com" <fo***@DBMonster.com> wrote in message
news:57******************************@DBMonster.co m... Please help me clarify a couple of questions:
Because I am cloning a source onto a target which has a different DB2INSTANCE, database name, host, etc, I assume I must use a "redirected restore". Just because I have links that make the target filesystem "look" like the source system, I cannot do a regular restore. Is this correct?
============================
No, that isn't quite correct. If you have two physical boxes, two physical
instances/databases, the regular restore still can work very well.
You can build all the containers on your target and create links and makes
them exactly look like the source server.
You can run the restore, just like you run the restore on the source server.
db2 restore db source_db ... into target_db ...
You don't must run redirect restore in order to clone a database onto
another physical database.
In fact you can even restore into the target from an online backup image
which taken from the source server, if both target and source server used
the same db2 udb version, and you can run the rollforward to the end of
logs. Is the deletion and recreation of the exisiting containers I observe
during the redirected restore a result of the "set tablespace containers" statements?
I did redirect restore before. But never observe the whole process. I can't
understand why the restore need to delete something.
When you restore a database to the different place, you'd better build all
the necessary containers, instead of letting the restore process to build
them, which will impact the restore performance. If this is true, given links that make the filesystems look the same, can
I do the redirected restore without using the "set tablespace constainers" statements?
Again, for your case, you can just run a restore, w/o redirect. This would save me 2-3 hours on my 300GB database.
Thanks Chris
-- Message posted via http://www.dbmonster.com
Another option you might want to consider is the use of raw partitions
(i.e. tablespace containers of type device) since they don't need to be
'created' in a similar fashion to file based containers. A create
tablespace or set tablespace containers command takes no time on these
types of containers. The last time I tried it, it was possible to
convert from file to device and vice versa with a redirected restore.
regards
tim This thread has been closed and replies have been disabled. Please start a new discussion. Similar topics
by: Vikrant |
last post by:
Friends,
Due an application (old) install program problem, Under AIX I could
only create DB2 instance, I could also catalog it. Its complex...
|
by: Stanley Sinclair |
last post by:
About to create a table which will "include" a BLOB. Am not sure how
large to make the container and the tablespace.
What I see says that BLOB...
|
by: GL |
last post by:
I am running DB2 8.1.1 on AIX 5.1
Having a problem with a redirected restore.
Once into the restore continue phase, I immediately get the...
|
by: Jack |
last post by:
Is it possible to find out what the tablespace definitions are of
database you are about to build using a redirected restore?
What I am wanting...
|
by: Chris |
last post by:
I have searched this group for answers and tried the responses.
I am trying to Use an full online backup from our production server and
apply it...
|
by: dbagirltx |
last post by:
We have done some testing with mixed and forgotten results. So I'm
hoping that asking here can clarify some issues for us. Right now we do
one...
|
by: africantract |
last post by:
Hi There, We have an issue on our db2 restores. We do database dumps
every night, and restore them to development servers every morning.
This works...
|
by: Lew |
last post by:
Hi,
I have a database that I am restoring to another database in the same
instance using a redirected restore. It seems to work fine but what...
|
by: mccuenpat |
last post by:
Currently trying to perform a redirected restore of DB2 database using Netbackup in AIX environment. The restore command completes successfully but...
|
by: tammygombez |
last post by:
Hey fellow JavaFX developers,
I'm currently working on a project that involves using a ComboBox in JavaFX, and I've run into a bit of an issue....
|
by: concettolabs |
last post by:
In today's business world, businesses are increasingly turning to PowerApps to develop custom business applications. PowerApps is a powerful tool...
|
by: better678 |
last post by:
Question:
Discuss your understanding of the Java platform. Is the statement "Java is interpreted" correct?
Answer:
Java is an object-oriented...
|
by: teenabhardwaj |
last post by:
How would one discover a valid source for learning news, comfort, and help for engineering designs? Covering through piles of books takes a lot of...
|
by: CD Tom |
last post by:
This happens in runtime 2013 and 2016. When a report is run and then closed a toolbar shows up and the only way to get it to go away is to right...
|
by: CD Tom |
last post by:
This only shows up in access runtime. When a user select a report from my report menu when they close the report they get a menu I've called Add-ins...
|
by: jalbright99669 |
last post by:
Am having a bit of a time with URL Rewrite. I need to incorporate http to https redirect with a reverse proxy. I have the URL Rewrite rules made...
|
by: Matthew3360 |
last post by:
Hi there. I have been struggling to find out how to use a variable as my location in my header redirect function.
Here is my code.
...
|
by: Matthew3360 |
last post by:
Hi, I have a python app that i want to be able to get variables from a php page on my webserver. My python app is on my computer. How would I make it...
| |