Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!rutgers!apple!voder!pyramid!ncc!alberta!ccu!rpjday
From: rpjday@ccu.UManitoba.CA (rpjday)
Newsgroups: comp.sys.apollo
Subject: sockets bug in 10.1
Message-ID: <231@ccu.UManitoba.CA>
Date: 19 Jul 89 14:54:23 GMT
Reply-To: rpjday@ccu.UManitoba.CA ()
Organization: University of Manitoba, Winnipeg, Manitoba, Canada
Lines: 13


  Being a bit of a rookie, both in terms of the Apollo and this newsgroup,
chances are this issue has come up before.  There appears to be a rather serious
bug in our DN3000 regarding sockets in BSD.  A set of application programs
communicating via sockets run merrily along for awhile, then just lock up.
A trace facility shows that a pair of processes has deadlocked, with
the process A send-blocked on process B, and B receive-blocked on process
A.  I have since learned that we are not the only Apollo site to have
experienced this bit of nastiness.
  Is this a recognized bug?  Is there a fix?  Enquiring minds want to know.

Rob P. J. Day
rpjday@ccu.umanitoba.ca