Import...

[anandkl]

Hi Friends,
I again ran into a major issue...this time with import...

We started a import process of a schema which was around 750 MB,
Initiall i tried with index=y commit=y analyze=y...this was taking too long around 5+ Hrs..then i start with index=n commit=n and analyze=n...again this was also taking 5+ Hrs...so we decided to do only empty schema import..i.e import only empty tables..again this was taking huge time...For all the methods followed, the import would insert the tables very fast say 30 minutes..then it would stop with no error messages and the CPU utilization was close to 100%..there was no other process was running.
The schema had LOB datatypes...

Can somebody tell me why this happened and what is solution for this.

regards
anandkl

[gandolf989]

If the initial extents are large it might be allocating space. Even if you are not bringing in data it will create the initial extent. If the initial extent is 100mb it will create a 100mb extent. You should look at how mauch space is allocated in your tablespaces.

[anandkl]

Hi Gandolf,
My initial extents are just 10k,but the fact is i just imported a similar schema without data and it just finished in 30 minutes...

regards
Anandkl

[gandolf989]

10K is nothing. I would look at the autoextend on for the datafile setting. If your database is creating a lot of extents you could set a minimum extent size of 128K. It is really hard to say what is causing the import to go slow. What is the import command that you are using?How big are the tablespaces? How many extents are being created for the tables that have already been imported? Are you having problems with errors on your volumes? It could be almost anything.