[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference turris::digital_unix

Title:DIGITAL UNIX(FORMERLY KNOWN AS DEC OSF/1)
Notice:Welcome to the Digital UNIX Conference
Moderator:SMURF::DENHAM
Created:Thu Mar 16 1995
Last Modified:Fri Jun 06 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:10068
Total number of notes:35879

8566.0. "Why did TCP partner did?" by APACHE::CHAMBERS () Thu Jan 23 1997 16:58

T.RTitleUserPersonal
Name
DateLines
8566.1netrix.lkg.dec.com::thomasThe Code WarriorThu Jan 23 1997 17:201
8566.2VAXCPU::michaudJeff Michaud - ObjectBrokerThu Jan 23 1997 19:3516
8566.3Can't use TCPIP to ask why the TCPIP connection failed...WTFN::SCALESDespair is appropriate and inevitable.Fri Jan 24 1997 11:327
And, you'd still want the case of a physical failure (i.e., the machine on
the other end got power cycled or crashed, or a component of the intervening
network failed), and there's no driver or parent process which can tell you
what happened in that case...


					Webb
8566.4VAXCPU::michaudJeff Michaud - ObjectBrokerFri Jan 24 1997 12:1215
> And, you'd still want the case of a physical failure (i.e., the machine on
> the other end got power cycled or crashed, or a component of the intervening
> network failed), and there's no driver or parent process which can tell you
> what happened in that case...

	Not quite true.  In the case of the remote system crashing
	(or network failing), this can/will be detected by the local
	end of the connection.

	In theory (untested by me) you should be able to get that reason via
	getsockopt(SO_ERROR).  On NT, the condition is reported
	differently (getsockopt(SO_ERROR already returns 0 on NT),
	instead from experience the recv() will fail with SOCKET_ERROR
	and WSAGetLastError() will return a different error if the
	network failed.