By using this site, you agree to our updated Privacy Policy and our Terms of Use. Manage your Cookies Settings.
437,924 Members | 1,686 Online
Bytes IT Community
+ Ask a Question
Need help? Post your question and get tips & solutions from a community of 437,924 IT Pros & Developers. It's quick & easy.

V9 Load from Cursor and multiple DB's

P: n/a
I am looking for comments on experience using a Load from Cursor
across multiple db's databases.
I have a multi-terrabyte database across many partitions that includes
a large table (1 Tb+). The system also contains UTF-8 and LOB data.
I am about to refresh the existing platform - going from V8.2 to V9
and leveraging new hardware. I an staying with SuSe Linux at 64-bit.
Does anyone have experience using this?

My initial plan was to export from OLD_DB in parallel to a local file,
then ftp to NEW_DB, followed by load. I am thinking now that the new
capabilities of Load from CURSOR DATABASE will provide reasonable
performance and ease of implementation. Not sure about bogging down
the coordinator on OLD_DB.

Are there any known limitations that I should be aware of ? How about
relative performance of these two methodologies?

Is it ok that NEW_DB will be V9 and the OLD_DB will be V8.2?

Thanks for your input.

Mar 6 '07 #1
Share this Question
Share on Google+
5 Replies


P: n/a
On Mar 6, 12:18 pm, "mike_dba" <michaelaaldr...@yahoo.comwrote:
I am looking for comments on experience using a Load from Cursor
across multiple db's databases.
I have a multi-terrabyte database across many partitions that includes
a large table (1 Tb+). The system also contains UTF-8 and LOB data.
I am about to refresh the existing platform - going from V8.2 to V9
and leveraging new hardware. I an staying with SuSe Linux at 64-bit.
Does anyone have experience using this?

My initial plan was to export from OLD_DB in parallel to a local file,
then ftp to NEW_DB, followed by load. I am thinking now that the new
capabilities of Load from CURSOR DATABASE will provide reasonable
performance and ease of implementation. Not sure about bogging down
the coordinator on OLD_DB.

Are there any known limitations that I should be aware of ? How about
relative performance of these two methodologies?

Is it ok that NEW_DB will be V9 and the OLD_DB will be V8.2?

Thanks for your input.
Mike:

Yes, we use Cross Load a lot and are very pleased with it. Not to
mentioned the DASD saved to store the staging files and the efforts to
transfer them. If the load failed, it is simple, just declare the
cursor and restart the load again.
The version of the DB does not matter, V9 can reference V8 DB.

If your network is not an issue, expect 10-15% impact on elapsed time
vs loading locally.

ps: It is extremely useful when you use Cross Load with Information
Integrator from a 3rd party Relational DB or MF DB2.

Regards,
Eric

Mar 7 '07 #2

P: n/a
On 3/6/2007 at 7:06 PM, in message
<11**********************@h3g2000cwc.googlegroups. com>, Eric
K<rd****@yahoo.comwrote:
On Mar 6, 12:18 pm, "mike_dba" <michaelaaldr...@yahoo.comwrote:
>I am looking for comments on experience using a Load from Cursor
across multiple db's databases.
I have a multi-terrabyte database across many partitions that includes
a large table (1 Tb+). The system also contains UTF-8 and LOB data.
I am about to refresh the existing platform - going from V8.2 to V9
and leveraging new hardware. I an staying with SuSe Linux at 64-bit.
Does anyone have experience using this?

My initial plan was to export from OLD_DB in parallel to a local file,
then ftp to NEW_DB, followed by load. I am thinking now that the new
capabilities of Load from CURSOR DATABASE will provide reasonable
performance and ease of implementation. Not sure about bogging down
the coordinator on OLD_DB.

Are there any known limitations that I should be aware of ? How about
relative performance of these two methodologies?

Is it ok that NEW_DB will be V9 and the OLD_DB will be V8.2?

Thanks for your input.

Mike:

Yes, we use Cross Load a lot and are very pleased with it. Not to
mentioned the DASD saved to store the staging files and the efforts to
transfer them. If the load failed, it is simple, just declare the
cursor and restart the load again.
The version of the DB does not matter, V9 can reference V8 DB.

If your network is not an issue, expect 10-15% impact on elapsed time
vs loading locally.

ps: It is extremely useful when you use Cross Load with Information
Integrator from a 3rd party Relational DB or MF DB2.
Is this referring to the item in the "What's New" for V9 manual under "Load
from cursor with remote fetch"? Looks interesting!

My V9 database is not up at the moment so I can't test it. Anyway, I found
the reference to it in the Data Movement Utilities and Guide Reference, but
it doesn't look like SQL Reference, Vol. 2 for DECLARE CURSOR has been
updated with the new options. Or am I just blind?

Thanks,
Frank
---
Frank Swarbrick
Senior Developer/Analyst - Mainframe Applications
FirstBank Data Corporation - Lakewood, CO USA
Mar 7 '07 #3

P: n/a
On Mar 7, 11:56 am, "Frank Swarbrick" <Frank.Swarbr...@efirstbank.com>
wrote:
On 3/6/2007 at 7:06 PM, in message
<1173233178.574549.227...@h3g2000cwc.googlegroups. com>, Eric

K<rdb...@yahoo.comwrote:
On Mar 6, 12:18 pm, "mike_dba" <michaelaaldr...@yahoo.comwrote:
I am looking for comments on experience using a Load from Cursor
across multiple db's databases.
I have a multi-terrabyte database across many partitions that includes
a large table (1 Tb+). The system also contains UTF-8 and LOB data.
I am about to refresh the existing platform - going from V8.2 to V9
and leveraging new hardware. I an staying with SuSe Linux at 64-bit.
Does anyone have experience using this?
My initial plan was to export from OLD_DB in parallel to a local file,
then ftp to NEW_DB, followed by load. I am thinking now that the new
capabilities of Load from CURSOR DATABASE will provide reasonable
performance and ease of implementation. Not sure about bogging down
the coordinator on OLD_DB.
Are there any known limitations that I should be aware of ? How about
relative performance of these two methodologies?
Is it ok that NEW_DB will be V9 and the OLD_DB will be V8.2?
Thanks for your input.
Mike:
Yes, we use Cross Load a lot and are very pleased with it. Not to
mentioned the DASD saved to store the staging files and the efforts to
transfer them. If the load failed, it is simple, just declare the
cursor and restart the load again.
The version of the DB does not matter, V9 can reference V8 DB.
If your network is not an issue, expect 10-15% impact on elapsed time
vs loading locally.
ps: It is extremely useful when you use Cross Load with Information
Integrator from a 3rd party Relational DB or MF DB2.

Is this referring to the item in the "What's New" for V9 manual under "Load
from cursor with remote fetch"? Looks interesting!

My V9 database is not up at the moment so I can't test it. Anyway, I found
the reference to it in the Data Movement Utilities and Guide Reference, but
it doesn't look like SQL Reference, Vol. 2 for DECLARE CURSOR has been
updated with the new options. Or am I just blind?

Thanks,
Frank

---
Frank Swarbrick
Senior Developer/Analyst - Mainframe Applications
FirstBank Data Corporation - Lakewood, CO USA- Hide quoted text -

- Show quoted text -
Yes, that is where we saw it as well - "What's New". We have done
some testing and it works nicely. But we tested only with small
files. Our big tables and LOB and UTF-8 data gives us pause. Perhaps
a refined cursor declaration (say by time period) and performing it
iteratively might provide better restartability for us. Network
latency is always a variable as Eric pointed out.

Mar 7 '07 #4

P: n/a
Ian
mike_dba wrote:
I am looking for comments on experience using a Load from Cursor
across multiple db's databases.
I have a multi-terrabyte database across many partitions that includes
a large table (1 Tb+). The system also contains UTF-8 and LOB data.
I am about to refresh the existing platform - going from V8.2 to V9
and leveraging new hardware. I an staying with SuSe Linux at 64-bit.
Does anyone have experience using this?
Sorry if this is a silly question, but why don't you just backup the
database and then restore it on the new servers? RESTORE will do the
migration for you, automatically.

Obviously this requires that you have enough space to hold the all of
the backup images, but this will likely be faster than export/load.

Mar 7 '07 #5

P: n/a
On Mar 7, 4:44 pm, Ian <ianb...@mobileaudio.comwrote:
mike_dba wrote:
I am looking for comments on experience using a Load from Cursor
across multiple db's databases.
I have a multi-terrabyte database across many partitions that includes
a large table (1 Tb+). The system also contains UTF-8 and LOB data.
I am about to refresh the existing platform - going from V8.2 to V9
and leveraging new hardware. I an staying with SuSe Linux at 64-bit.
Does anyone have experience using this?

Sorry if this is a silly question, but why don't you just backup the
database and then restore it on the new servers? RESTORE will do the
migration for you, automatically.

Obviously this requires that you have enough space to hold the all of
the backup images, but this will likely be faster than export/load.
Not a silly question at all. The target system has numerous changes -
# partitons, DDL changes for things like # containers, MDC, etc.

Mar 7 '07 #6

This discussion thread is closed

Replies have been disabled for this discussion.