Nicolas Williams wrote:
On Tue, Feb 21, 2006 at 08:42:00AM -0800, Bart Smaalders wrote:
If we don't do automatic reconfigs of network state, is there
a way to insure that network operations depending on a currently
unavailable network "fail fast"? If my wireless link has dropped,
waiting for YP/DNS/LDAP to timeout is going to be rather tiresome;
I'd much rather it failed right away with a "no route to host"
kind of message.
This is hard to address in the path of the hang -- the functions that
are hanging are hanging because that's the only thing they're expected
to do in these situations, and changing this behaviour isn't likely to
be feasible.
The best thing to do in the short-term, IMO, on a GUI console login is
to pop-up a little ballon saying something to the effect that "you've
lost your network, things will hang, click here for some options."
Fast failures is a two edged sword.
On a Windows XP machine, link breakages are reported with a little
balloon popup and followed by all of your TCP connections breaking.
This is bothersome if you've got a putty window open and the cable
is either knocked out or you need to unplug-replug to sort out a birds
nest or you need to power cycle your own AP or....
For TCP connections, one path to becoming aware of the other end
going away is to use keep alive messages. Maybe we need to have
a think about how keep alives are enabled, set, tuned and used in
Solaris as part of addressing this issue.
Darren
_______________________________________________
networking-discuss mailing list
[email protected]
