Conference movies::decdtm-vms

    Anyone got any ideas regarding the base note?

Don,

Sorry, I don't have any idea as to what you're getting here.  Have you asked
the Rdb folk what that error really means?  Without that information, there's
no way I could even hazard a guess...

regards,
//alan

    Alan,
    
    I work in Oracle Rdb/DBI technical support.  DEC contracts to Oracle
    for support of the DBI product suite.  This is until June 24 of this
    year.  This is why I still have access to DEC systems even though I was
    sold to Oracle in the DBI sale last March.
    
    The person resporting the problem is from Rdb support in Denmark.  I
    have forwarded your reply.
    
    I have also mentioned to this person that I have seen this error before
    but in rare cases.  Where I have seen this is with DBI and field test 
    versions of Rdb.  In my instances this is what happened...DBI always
    queries to see what the highest version of Rdb/Dispatch is on a system.
    It does this for capabilities purposes.  I've seen cases where DBI will
    make a call to Rdb/Dispatch, and a field test Dispatch doesn't know
    about the call.  When I mentioned this to my friend he said that he'd
    check on this but he was pretty sure that there was only one version of
    Rdb involved and that it wasn't field test.  I have yet to hear back.
    
    We were just wondering if this symptom/error has ever been reported to
    DECdtm or if anyone in DECdtm was familiar with the symptoms described.
    
    Thanks for the note.  I'll write back when I hear more from the other
    side.
    
    Don

Alan,

Some more info on this.

The RCI error is indicating that an DSRI (RCI) call was incorrectly 
constructed. The most likely cause in an rdb_attach_database call is 
that someone has set the instantation number to non-zero for a 
remote attach.

One posibility is that for some reason DTM is trying a remote attach 
as part of its handling of this distributed multi-database transaction.

Another is that DTM in some way corrupts calls or Transaction IDs.

This is not an attack on DTM. We are investigating if VMS, ACMS or Rdb 
are doing this too. But we need to follow up all the paths, especially 
if there are any known issues?  We have already thoroughly checked the 
other 3 so now it is time to ask DTM.

Thanks,
Don

Don,

Thanks for your replies.  All I can say is that this is entirely foreign to
me - I have heard nothing similar before.  Equally, I should point out that
I no longer work on DECdtm.  The maintainer for the product is Carey Donat.

It might be worth dropping a mail to her - STAR::CDONAT to see if she has
heard of anything similar.

It would be invaluable if you could find out whether this error is being 
generated as the result of a specific call.  For example, I think I'm right
in saying that an attach in Rdb parlance corresponds to the DECdtm system
service SYS$ADD_BRANCH.  A specific error code back from that service would
make this problem much easier to solve.

I'm not going to be around for the next week or so, so I would recommend
pursuing this with Carey.

Good luck,

regards,
//alan

    Thanks Alan.
    
    Don