Hi,

We are trying to take table dumps using ttaad4226m000 session, but it is quite slow, will it be faster if we take the table dumps directly in oracle ?
if it is better in oracle, can anybody suggest whats needs to be done in oracle ?

thank you in advance,

Hi,

Just trying to share my exp... taking seq. dump through baan is simple n straight forward, however it has few disadvantages like :

1) it places a lock on table, so if u r exporting a heavy table, and at the same time user does some update on that very same table, he would get a fatal error "cant lock table"... however i heard recently that in LN they have changed this functionality a bit, not sure about it though since i haven't seen the LN screen yet and i am not sure of ur baan version.

2) seq. dump is slow however one can play around with db_resource to make it fast... search for db_resource on baanboard

3) You loose data consistency like if u r exporting complete co. it might export first free no. first and by the time it finishes off taking header / lines tables, first free no. is again changed due to the fact somebody created a new order.

seq. dump is good if u r not having too much of data.

I do not your oracle version, but in older oracle version u have export/import utility. ... and in newer version they recommend using Oracle Data pump instead of oracle export/import

We use to take seq. dump. and it use to take close to 14-16 hrs, but with data pump it now takes close to 3 hours. You can set parallelism while using data pump ( like kicking multiple CPU simultaneously )

The only disadvantage that i see of data pump is that since it produces oracle propriety format ( only dpimp can read the files ), u do not have a provision to change the co. no., in case if u plan to see this dump to refresh ur test server

-Learner