Re: Help: Standby database

"Casey Dyke" <cdyke_at_excitehome.com.au> wrote in message news:3AF257E6.E700BE13_at_excitehome.com.au...
> >
> > select file_name
> > from v$archived_log
> > where seq#=(
> > select max(SEQ#)
> > from v$archived_log
> > where archived='TRUE'
> > );
> >

>
> We do the same thing, but archived (unless I missed something) is a
> varchar(3) column - hence we use "YES".   Same def in 8.0.5,8.0.6.8.1.7.
>
> I might add that although the 8i improvements are pretty nice, you'd still
> need to consider checking that sql*net has indeed successfully transferred
> your logs.  With our scripts (run as daemons), network errors (and many
> others) are detected upon next loop and propagated out via our alerting
> system to a 24*7 noc.  Going up to 8i and relying on the new features
> could, theoretically, expose us.  Our scripts work like a charm and form a
> hugely important part of our backup and recovery strategy ...  Not
> poo-pooing the new features - they'll have their place at many sites,
> particularly those where scripting skills are lacking.
>
> Cheers,
>
> Casey ...
>
>

Just FYI: the standby mechanism can be configured to retry and to consider locations as optional or mandatory.
So you don't necessarily need to check whether sqlnet has copied your logs. Oracle is robust enough to do that and deals with those errors when configured correctly.
I wouldn't like to make an estimate how much developing your scripts would cost, and they will be still (as you describe above) less dependable. If I have to rely on NT's replication features, I would change to standby immediately.

Regards,

Sybrand Bakker, Oracle DBA Received on Fri May 04 2001 - 05:12:45 CDT