Re: RMS read- (write) gateway

Well.. I don't think it's going to happen from the Oracle gateway perspective. We just really like to have transactions that ARE transactions and locking (well concurrency) that can be a little more trsutworthy. I guess the best thing to do is to ask your account rep to submit an enhancement request. There will be other (programmatic) solutions at some point (where you can kinda cheat and do what you want in a C routine that PL/SQL can call).

In article <43sh32$pkj_at_news.dknet.dk>, cfalk_at_pip.dknet.dk wrote:

> > kbradley_at_us.oracle.com (Kirk Bradley) writes:
> > In article <43egqn$3ld_at_news.dknet.dk>, cfalk_at_pip.dknet.dk wrote:
> >
> > > The RMS gateway from Oracle is only a read gateway.
> > >
> > > Has anyone developed a write gateway to RMS, or know of any easy way of
 doing it ?
> > > I would like all tables Oracle and RMS to appear in the dictionary.
> > > The easy way is to use a RMS library from C and code an easy interface.
> > >
> > > Now that Oracle has bought RDB from Digital will they support RMS
 as well ?
> > >
> > > cheers
> > >
> > > Claus Falk
> >
> > what would you want done with 2-PC? Or, record level locking to
> > prevent 2 users from updating the same row? or, what if a write
> > fails for some reason... how would it get backed out? There are
> > a lot of problems with write access to FILE systems.
> >
> >>>>
> There will not be any possibilities for roll-back in a gateway like
 this. (RMS-jounaling ???)
>
> There might also be some constraints regarding locking.
>
> RMS for VMS is just like C-ISAM for Unix, and sometimes it makes sense
 to store
> information in index-sequential files, eg. for speed or exchange with
 legacy systems.
>
> Both reasons are valid for my situation, and the migration to Oracle
would be speeded up by a 2-way interface. Received on Fri Sep 22 1995 - 00:00:00 CEST