Re: kpasswd and retransmits

[Assar Westerlund <assar@sics.se>]

Derrick J Brashear <shadow@dementia.org> writes:
> It turns out what's happening is the daemon is delayed just long enough
> that the client sends a second request, and caches the new ctime/cusec in
> the auth context, then gets a reply with the old one. 
>
> I'm not sure what the "right" answer to this is... not retransmit?
> remember all the ctime/cusec pairs?

I took the easy way out of increasing the initial timeout (which
should make this not happen) and not sending a new request if we
actually got an old packet back, hoping that we will get an reply to
the newer packet too.  How does that sound?

/assar