Named locks are scoped for the server, so any other page on that
server with a lock using the same name will be affected. This is how
we originally ended up with <cflock
name="#application.applicationname#"> prior to the advent of the scope
attribute in the cflock tag. I've accidentally bound (and decreased)
the performance of multiple applications together this way using
cflock.
> You can use CFLock just use the name attribute instead of
> scope. Keep in
> mind that it will only lock that file within that
> particular CF Application.
> _____
> From: Jim Campbell [mailto:[EMAIL PROTECTED]
> Sent: Thursday, July 29, 2004 3:43 PM
> To: CF-Community
> Subject: CFTRY CFFILE CFWAYTOOTIRED
> I've been up for 18 hours... so I'm definitely sharp
> today. We've got
> an operation that needs to append to a single CSV. The
> request could
> come from any one of four load balanced servers to a
> single netapp filer
> system. The problem is the occasional race condition
> where the CSV will
> be in the process of being written to by one of the CF
> servers, and the
> other one needs to wait its turn like a good CF server, or
> risk being
> sent to bed without its dessert.
> I can't use CFLOCK with a file operation can I? It's easy
> enough to
> CFTRY CFCATCH the file operation, but when it fails, I
> want it to keep
> trying until it gets a chance (within reason, of course -
> maybe 20 tries
> before failing), since it will only be a few milliseconds
> before its
> chance comes up. I'd like to avoid the ignominious coding
> mess of
> stacking retries in CFCATCH blocks, and there must be a
> more graceful
> way of doing this, but I can't think of it. Any ideas?
> - Jim
> _____
>
[Todays Threads] [This Message] [Subscription] [Fast Unsubscribe] [User Settings] [Donations and Support]
