|
Peter, I don't understand your problem. Surely
your application shouldn't care what node your FTSV job
runs on ?
If it does then why ?
Please remember, FTSV always used to actually queue
to SYS$BATCH, which as we all know would use RDGE00_BATCH OR
RDGE28_BATCH.
As for the reason for changing the queue, thats
another question !!!
But for the record I asked for it to be changed, my
reason was simple. If System management insist that RDGE28
network database does not have ALL nodes defined then whats
the point of of using that node to copy files around ? It
only proves frustrating if you are on U9, where you can see
the file you want to copy, so you set off and FTSV copy to
go and get it, only to find that the job ran on RDGE28
which didn't know about the node :-(
Iain..
|
|
RE .2
> Peter, I don't understand your problem. Surely
> your application shouldn't care what node your FTSV job
> runs on ?
> If it does then why ?
Simple answer is that it uses a mailbox to communicate with FTSV.
That mailbox must be on the same node as the FTSV process that wants to
write to it otherwise, unsurprisingly it is "not found" and then the
mailbox process sits there for ever waiting to be told what happened.
> Please remember, FTSV always used to actually queue
> to SYS$BATCH, which as we all know would use RDGE00_BATCH OR
> RDGE28_BATCH.
If you look back at my note you will see that we reassign the
logical in our process table which causes FTSV to use OUR assigned queue.
The value in the process table is set up automatically in a LOGIN.COM by
looking at the nodename.
> But for the record I asked for it to be changed, my
> reason was simple. If System management insist that RDGE28
> network database does not have ALL nodes defined then whats
> the point of of using that node to copy files around ? It
> only proves frustrating if you are on U9, where you can see
> the file you want to copy, so you set off and FTSV copy to
> go and get it, only to find that the job ran on RDGE28
> which didn't know about the node :-(
I couldn't agree more, however wouldn't it have been more sensible
and no more difficult to have simply changed the value of the logical
FTSV$QUEUE ? This would have been totally transparent to everyone.
Peter
|