--
<zifnab_at_islandsofmyth.remove.org> wrote in message
news:Xns93D45777E326Dzifnabislandsofmytho_at_206.127.4.21...
> "Burton Peltier" <burttemp1REMOVE_THIS_at_bellsouth.net> wrote in
> news:tVFZa.14238$vK4.11346_at_fe05.atl2.webusenet.com:
>
>
> I think you might slightly be missing Quarkman's point here, maybe a
> timeline will help (assuming I understand it correctly as well).
>
> time A ) Cold backup taken including redo logs
> time B ) Backup complete
> time C ) Database opens, users start doing transactions. Recorded in Redo
> not archived yet.
The "not archived yet" is key to any use of the REDO logs you are referring
to. I don't think anyone was advocating using the "current" REDO logs after
restoring from a cold backup. I suppose that would work if you ensured the
"not archived yet".
I say use the REDO logs backed up at the same time as the cold backup and
not have to worry about a "clean shutdown" .
Quarkman (and others) say don't use any REDO logs (correct me if I say this
incorrectly) , make sure you always have a "clean shutdown", and do an
"open reset logs" after recovering from a cold backup (which re-creates
empty REDO logs) and never backup REDO logs.
> time D ) For whatever reasons the cold backup needs to be applied from
> time A.
> time E ) Time A's cold backup is restored including redo logs.
> time F ) database reopens
>
> At time F you have now lost any transactions from C that were not
> archived yet because you restored the full cold backup + redo logs.
I thought Quarkman was advocating a reset logs after recovering the "clean
shutdown" datafiles of the cold backup!
This means you lose the same transactions in C .
>
> I think that is the point that is trying to be made.
>
>
>
Received on Mon Aug 11 2003 - 20:47:37 CDT
