Export/Import is a bad solution to use for this.  Instead you should use
the Copy StgPool feature, plus to help you in your automation you should
consider obtaining the DRM (Disaster Recovery Manager) feature of the
product.

Paul Bradshaw
----------
> From: Tom Denier <tom AT STAFF.UDC.UPENN DOT EDU>
> From: Tom Denier <tom AT STAFF.UDC.UPENN DOT EDU>
> To: ADSM-L AT VM.MARIST DOT EDU
> Subject: IMPORT node questions
> Date: Thursday, January 30, 1997 11:01 AM
>
> My site is currently running ADSM Version 1 under AIX. We are using the
> EXPORT NODE command with the FILEDATA option to produce offsite backup
> tapes for disaster recovery. We are preparing for a disaster recovery
> test. Since the EXPORT currently takes almost 24 hours and the whole
> test period is only 24 hours, the performance of the corresponding IMPORT
> operation is a matter of some concern.
>
> How does the running time for an IMPORT from tapes with file data
> compare with the running time for the corresponding EXPORT?
>
> The documentation for the IMPORT command indicates that it is possible
> to import a subset of the data on a tape produced by EXPORT NODE,
> selecting by node, filespace, or both. Is this in fact true? If so,
> how does the performance of a selective IMPORT compare with that of
> a full IMPORT? For example, does importing half of the data take 50%
> of the elapsed time of the full import, or almost the same time, or
> something in between?
>
> Are the files on an EXPORT tape grouped by node? For example, would
> the command
>
>   export node alpha.umis.upenn.edu,beta.umis.upenn.edu filedata=allactive
>
> produce a tape with all the active files from alpha followed by all the
> active files from beta? Similarly, will the files from a given filespace
> be grouped together on an EXPORT tape?
>
> Once an IMPORT command places some files in the storage pools on the
> disaster recovery server, will it be possible for the clients owning
> those files to begin restoring them while the IMPORT command is still
> running?