Mac OS X Leopard servers mis-using the Apple File Protocol

Jan 15, 2009 09:35 GMT  ·  By

Macnn points out to a mile-long Apple Discussions forum thread debating (what else?) a big problem. Apparently, there have been severe problems with Mac OS X servers causing corporate networks to crash constantly. The main bug is believed to be Apple's own AFP.

According to the report, complaints from network administrators indicate that the problem is most likely caused by Apple's File Protocol used by Mac OS X Leopard Server.  Administrators have noticed that quad- and eight-core Xserves have suddenly become non-functional due to AFP activity.

“My problem is that my os x 10.5.4 server with about 30 networked homeddrive [sic] users have an issue with the afp server,” one complaint reads. “The afp server process uses all 8 cores of this newest intel xserve with 14 gigs of ram installed. When this happens, all users get a spinning wheel. The incoming network traffic is reduced to some kb´s [sic],” the user explains.

“[...] All users shut down there [sic] clients – restart server and about 30 minutes later i have the same problem,” the post continues. “I have dumped the network traffic with Wireshark, and there i see some tcp retransmissions. Now i need someone who can help me analyse [sic] the Wireshark protocol, because i cant´s [sic] handle that.”

It is still unknown why AFP overloads are being experienced. The only resolution administrators could find was to kill unnecessary processes and tasks and/or restart the server. Others managed to reduce server load by adjusting the afp_wan_quantum and afp_wan_threshold variables on a server, coupled with killing unnecessary processes or tasks, the same forum thread reveals.

Apple has been known to respond to widespread issues posted on its official forum by regular users, even calling up customers at their homes. Apple hasn't yet responded to this issue in any way, as far as we know.