Remote File Sharing (RFS) - SVR3

Doug Gwyn gwyn at brl-smoke.ARPA
Thu Jan 8 02:09:38 AEST 1987


I think Guy's point was that the CONNECTION (e.g. "file") is
broken by the link going down, and even though RFS re-advertises
systems when they come back up, it is too late for the particular
connection, since the first I/O operation after the link outage
returned an I/O error rather than blocking until the link came
back up.

Which of the two ways of dealing with this is better depends on
what one is trying to do.  Personally I prefer the immediate I/O
error in most situations.  A broken link seems to me to be much
the same as a disk drive going off-line.



More information about the Comp.unix.wizards mailing list