steve@umiacs.umd.edu (10/06/89)
[ Gee, after answering this and a previous question, I'm beginning to think
that some of the network-related traffic should up and move to the
sun-nets@umiacs.umd.edu list... If you're not on, and want to be, contact
sun-nets-request@umiacs.umd.edu or {...}!uunet!mimsy!umiacs!sun-nets-request.]
The rsh hang problem comes about when the destination machine is alive and
OK enough to accept TCP connections, but not alive enough to make any
progress on them. The connection should stay open -- after all, the
machine may spontaneously come alive again, say once a hung NFS mount of
/usr becomes functional again. (Connections that time out spontaneously
are a Bad Thing, especially in the packet radio world, where TCP transfers
over packet radio have apparently been known to hang for days while
someone fixes a transmitter, then pick up again.) It's somewhat
inconvenient for scripts such as this.
We cheated: in most of our scripts of this sort, we don't say "rsh machine
bla bla bla", but instead say, "timer -t 1800 rsh machine bla bla bla".
The timer program runs the rsh, and kills it off if it goes on too long.
The timer program is short; I can send it around or put it out for
anonymous FTP, if there is interest.
-Steve
Spoken: Steve Miller Domain: steve@umiacs.umd.edu UUCP: uunet!mimsy!steve
Phone: +1-301-454-1808 USPS: UMIACS, Univ. of Maryland, College Park, MD 20742