Re: Transportable Tablespaces and Deferred segment creation "bug"

From: Norman Dunbar <>
Date: Tue, 27 Nov 2012 08:03:24 +0000
Message-ID: <>

Morning Pier Paolo,

On 26/11/12 23:11, pier paolo Bruno wrote:
> Pherhaps it seems stupid .

No, questions are never stupid!

> I imagine that at the end of the procedure to move your instance you have
> to make an import for loading all not segment objects...
Yes. After the TT import and some work setting default tablespaces and changing the transported tablespaces to read write, there is a norow import.

However, I like to have my problems resolved at one step before movving on to the next. When the problem occurred, I wasn't aware of what it was nor the cause. The docs are next to useless on the matter as they don't go in to any detail about the fact that deferred segments are not acceptable to SE databases.

> Can you let impdp with CONTENT=METADATA_ONLY recreate empty and missing
> table for you from a dump created with content=metadata_only (you can not
> use exp with rows=no for deferred_segment_creation )
Impdp, in this example, is unusable because doing a TT impdp requires a database link be created between the importing database and the exporting database. In our set up, production networks are air-gapped from test and/or QA networks, so the transfer of data must be by tape (and, as it turns out, secure courier).

Exp/imp is simple, works, and everything is there. Expdp/impdp is not!


Norman Dunbar
Dunbar IT Consultants Ltd

Registered address:
Thorpe House
61 Richardshaw Lane
West Yorkshire
United Kingdom
LS28 7EL

Company Number: 05132767
Received on Tue Nov 27 2012 - 09:03:24 CET

Original text of this message