RE: exporting a 35G database [3024]
Posted in 2004
dbexport will be very slow, instead of using dbexport you may consider using
the HPL utility to unload and load the database, if you use multiple files
option in HPL you will not have to worry about the 2Gb limit (since a table
will be split into multiple files and the load/unload will be much more
faster). i would think that, with HPL you will be able to export/import a
35GB database in under 3-4 hours, here are few tips on how to achieve this,
1)use multiple file option in HPL to unload/load large tables
2)use EXPRESS mode to unload and load the data (don't use the binary mode,
it may not be compatible across the OS platforms)
3) capture the database schema using the dbschema utility, make sure you
specify -ss option.the schema can be used to create database tables, stored
procedures, triggers, indexes on the destination system before loading the
data
4) initially create the destination database as no-logging database, change
the logging mode after th e load.
Hope this helps.
-----Original Message-----
From: owner-informix-list@iiug.org
[mailto:owner-informix-list@iiug.org]On Behalf Of Alexey Sonkin
Sent: Wednesday, May 26, 2004 1:02 PM
To: 'KENNETH PENZA'; ids@iiug.org
Cc: 'informix-list@iiug.org'
Subject: RE: exporting a 35G database [3024]
KENNETH,
You didn't mention what platform You are going to use for 9.40.
I suppose, it is not NCR, because there is no 9.40 for NCR.
If the platform You are going to migrate to is Sun, IBM or any
other supporting files larger then 2GB You can use the following approach.
Use 9.40 'dbexport' on remote host to connect to 7.30 database to
extract data from it into flat UNL files, located on the New machine.
This must work, and You should be able to export tables into files
larger then 2GB (9.40 dbexport supports large files)
Then, after properly editing SQL script generated by dbexport, just
reconnect to the 9.40 engine and import Your data into 9.40 engine.
------------------------------------------
Alexey Sonkin
> -----Original Message-----
> From: KENNETH PENZA [mailto:kenneth.penza@gov.mt]
>
> People,
>
> We are tring to export a 35G informix 7.30 database to importing again on
> an infromix 9.4 database.
> The 7.3 database resides on a NCR Unix machine that does not support files
> larger then 2G (unfortunately we have 5 tables larger then this limit).
> Due to incompatible tape drives fitted on the machines we cannot perform
> an export to tape and import it again.
>
> What I was thinking is as follows
> create a unix pipe (mknod mypipe.1 p)
>
> In the backgroud start a process that reads data from the pipe.
> dd if=mypipe.1 | split -a 5 -b 1024m &
>
> and start the export
> dbexport -d se4rt -t mypipe.1 -s 1024000 -b 32 <<!>
> !
>
> However I fear that the split and the dbexport get out of sync and make
> the export process useless.
>
> Is there a better or cleaner way to do this ?
>
> Thanks and regards
> Kenneth
sending to informix-list
sending to informix-list