Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
Home -> Community -> Usenet -> c.d.o.server -> Re: Re-organising a database?
Am Mon, 14 Apr 2003 02:19:22 -0700 schrieb Tim Kearsley:
> Hi all,
>
> Environment is:
>
> Oracle 8.1.7 on AIX 4.3.3 on RS6000 SP hardware
>
> I've got a legacy database, about 30 Gbytes in size, which I would
> very much like to re-organise. Primarily I would like to implement
> LMTs and to put the datafiles into a structure which is fairly close
> to OFA.
>
> The obvious solution I suppose is to create the new database and
> export/import from the existing one. My problem is the time this
> appears to be taking. A trial took some two days by the time I had
> imported, re-created indices and solved a few problems with invalid
> packages. The database is pretty heavily used and it is going to be
> hard for me to sell a weekend's downtime to the users.
>
> Are there any other techniques I could use to reduce this downtime? I
> admit that I haven't tried tuning the export/import process. At
> present the export is done to a Unix pipe which compresses the dump
> file on the fly. Likewise the trial import was done by uncompressing
> the dump file on the fly. Is it likely to be substantially faster to
> export to multiple files rather than compressing (I am facing an OS
> filesize limit)?
>
> Any thoughts or ideas very welcome.
>
> Thank you.
>
> Tim Kearsley
> Database Manager
> Milton Keynes Council
Hi Tim
i will introduce a modified export/import scheme which is much faster. The
idea is to split the import in two chunks. First import data, packages and
create the referential inegrity items. Then (beeing online again) create inedexes in bulk/batch
fashion. The procedure goes as follows:
0. Read "How to stop defragmenting and start living: The definitive word on fragmentation". You get that at otn.oracle.com
The whole process executes much faster, but needs more planning and preparation.
Juergen Received on Mon Apr 14 2003 - 07:53:21 CDT