Apple needs to investigate this issue promptly. These kinds of problems could be hard to diagnose without the help of Apple and a good analysis of the network traffic. However, part of the cause could be related to an inappropriate network architecture.

the awesome writing/editing continues. yay.

I traced my problem of this nature to the use of Safari 3.2.1 by network accounts. Once I switched everybody to Firefox and disallowed the use of that version of Safari the problem hasn't resurfaced. I have six Xserves serving home folders and running 10.5.5 Server and 450 clients on 10.4.11 and 32 of those on 10.5.6.

Actually it's "Apple Filing Protocol" not Apple File Protocol. Ain't trvia fun?

...to finally implement rw FTP in the Finder? They even embedded Panic's FTP in iWeb. This is eternally puzzling.

What can we really expect forma company that still has a finder that hangs just because network store isn't responding? Or a CD that wont eject for mysterious reasons? They should also be able to tell the user EXACTLY what program is halting media ejections, as the generic messages rarely help when they say "start closing programs" WHAT?. They should be able to cancel a network store mounting WHEN I CLICK CANCEL and not halt the finder. Most beginning users know nothing of the terminal and how to recover from this when the relaunch finder command fails. There are serious issues that need to be fixed and if Microsoft was a pompous as apple was, they would make commercials saying how crappy certain fundamental parts of OS X STILL work nowhere near properly. As mac lovers we need to get on Apple to get this tuff fixed! 99% of everything else of course works like a charm which is why Apple rocks!

We've been using Apple servers for eons now, AFP problems are nothing new. We had issues with stability (caused by AFP) as far back as 10.2.

Apple's line is a consistent "try this patch", and we've never got one that works. So far, what does seem to work is to reduce the number of users on each server (50 seems to be the maximum for a quad-core and 100-or-so for an 8-core), and set the servers to restart every night when no-ones using them.

Personally, I think it's a memory leak somewhere in the AFP implementation, simply because the issue is cumulative. Starting with a fresh reboot, everything works well, then as time progresses, it gets worst and worst, until the AFP thread takes over all the processing power and brings the servers to a halt.

OsX does have its of problems, but it has its advantages as well. Don't bash a system just because of one "glitch" (persistent as it may be).

M. (ACSA)