On Fri, Aug 28, 2009 at 02:33:28PM -0700, Gary Mills wrote: > The problem I'm trying to solve is for services like telnet or rlogin which > don't > have a way to enable keepalive at either the application level or the TCP > level. > Yes, doing this may violate company policy, but that's a separate issue.
IMO TCP_KEEPALIVE should not be seen as violating idle session logout policies -- such policies should determine idle-ness using better measures than "are there packets moving for the connection's 5-tuple", such as "is there _data_ moving for the connection's 5-tuple". > > I would like to see a corresponding keepalive timer interval > > setting, though I don't think this project is strictly incomplete > > without one, it's certainly close to it. The default setting (two > > hours) is so much more than just a tad too long -- having to set it > > system-wide seems obnoxious to me. > > Yes, I agree, although that's not part of this RFE. It's actually > not too bad to set the timeout system-wide because it's still > only used if the service sets the corresponding socket option. Understood. Thanks so much for doing this! Nico -- _______________________________________________ networking-discuss mailing list [email protected]
