Traveller2003 wrote:
> Hi All ,
>
> Informix 9.4UC4
> Solaris 8
> Onbar
> Tivoli Storage Manager
>
> If I do a Full System Backup the night before and all the subsequent
> logical logs are then sent to the storage manager. I find the next day
> that one of the DB spaces is down and I have to restore from the
> backup.
> So I then recover the DB space.
>
> What next? Is IDS clever enough to roll forward just the transactions
> to do with that DB space while the rest of the system is on line?

During physical restore only pages from that dbspace will be extracted.

When it comes time to replay the logical logs, and you MUST do logical
recovery, the engine is clever enough to just rollforward the parts of
the transaction dealing with the dbspace being restored. This is
required for maintaining data consistency when doing a restore.



>
> My view is that is dependent on how the application is written for a
> start but I am not sure about getting the dbspace consistent up to the
> point in time when it failed. That is to say be able to roll forward
> the logical logs just for that DB space.


The onstat -d output will contain an "I" flag indicated the dbspace is
inconsistent. When the database restore process completes and all log
records have been replayed this inconsistent flag will be removed. At
that point all indexes and data are 100% consistent.


>
> Can I please get a view or feed back?
> Also if this is not possible can it be a feature request which other
> would no doubt benefit from.

This feature was introduced in version 6.0 of informix.

Hope this helps,
John


>
>
> Thanks
>
> Traveller