Kris Buelens <[email protected]> wrote :-

> You could use Q FILEPOOL CONFLICT to see wait conditions

> SFS guarantees read integrity, that is, if someone has a file in read, 
SFS keeps that image 
> of the file until the last reader closes it (this image is kept in 
shadow tables).  
> And -just tested- you can indeed ERASE the file while someone has it 
open.  
> The user reading it can just continue.

> Maybe you can to stop using the lock files by using SFS CREATE LOCK 
commands

Thanks, Kris, for your very helpful answer. 

Yes, the Q FILEPOOL CONFLICT is a good idea if I can catch it again. The 
last time it happened was at the weekend and I did not have any chance to 
investigate.

Your information on the ERASE is great - that gives me a strategy (limited 
to one simple sub EXEC on the upload server (A) ) to overcome any 
potential conflicts. 

Yes, using the SFS standard locking mechanism would have been a good idea 
if I were starting from scratch but the code in Server B would need to be 
modified to deal with finding a lock in place. As it is, it will not start 
scanning the file if the code defined lock file is in place. I want to 
avoid any changes that are too widespread due to the sensitivity of the 
application (there are daily production logs). 

Thanks anyway. I now have a way to proceed.


Colin Allinson
VM Systems Support
Amadeus Data Processing GmbH

Reply via email to