No... don't go there. Standby databases are supposed to be run on
identical, or nearly identical, machines. The disk layouts can be different
(different directories etc), but not completely different O/Ses. You might
get away with transfering logs to a different O/S, but you also have to get
away with transferring the original data files onto the different O/S. You
might get away with it... not having a VMS box ready to hand in my study, I
can't tell you for sure! I *think* someone once posted here that a Linux to
Solaris transfer once worked. But I can't be sure of that, and even if it
did, it was a sort-of Unixish-to-Unix transfer. Last time I called VMS a
sort of Unix, though, one of my work colleagues practically decked me! (He
happens to be quite keen on VMS!!) I know Linux to Windows won't work, and
I think that's the kind of gulf you're talking about here.
So, this idea is really a non-starter, and I shouldn't have mentioned it and
got your hopes up. Sorry.
Resources for Oracle: http://www.hjrdba.com
> Thanks, thats great info. So to get a standby database kicked off on
> another server do I have to make a mirror copy of all database files
> and then start manually recovering?
> Are archive log files binary compatible across platforms - could I use
> a VMS box as the standby instead of another linux box?
> > You're describing a standby database, and that was possible in 7.3. But
> > *managed* recovery mode was not available until 8i (where the machines
> > automatically exchange archives, and they are applied automatically at
> > standby machine). So if you're going to do it, you'll need to sit there
> > manually transferring archives, and manually applying them, unless you
> > dab hand with a shell script and cron.
> > I seriously doubt that 60Gb can be upgraded to 8i in 3 hours, but it's
> > impossible to tell with the information provided.
> > If the only reason you are contemplating the standby database is that it
> > gives you a 'clone' which can be upgraded leaving the source unaffected
> > with no downtime, that's a fairly convoluted way to go about it.
> > databases can be invalidated by a single 'nologging' (or
> > DML operation, and assorted bits of DDL will play havoc, too.
> > The upgrade process does not require you to use export and import: you
> > do an in-place upgrade, which wouldn't require very much extra disk
> > all, and would be much quicker than a full export. So downtime would be
> > required, but not as much as with export. But whether 3 hours is
> > possible... as I say, its impossible to say for sure, but I doubt it.
> > Regards
> > HJR
> > --
> > ----------------------------------------------
> > Resources for Oracle: http://www.hjrdba.com
> > ===============================
> > > Hi,
> > > I have an intel box running RH 5.2 with Oracle 7.3.2 (archive log
> > > mode).
> > > I have another intel box running RH 6.2 and Oracle 7.3.2.
> > > Can I copy archived redo logs from RH 5.2 to RH 6.2 and have that
> > > database in a constant recovery mode, thus synchronizing the data?
> > > I do not want to stop the database and do exports as it would take too
> > > long (60GB db, 10 Mb ethernet) and I have limited disk space left!
> > > I then want to upgrade RH 6.2 from Oracle 7.3.2 to Oracle 8.1.7 - all
> > > with a minimum (1-3 hrs?) of downtime - is it possible?
> > > I guess this is where a Oracle DBA really shows her colours but it is
> > > hard to justify the cost for this low funded project.
> > > thanks!
> > > Tim