Date: Mon, 1 Jun 1998 23:18:27 +0000 (GMT) From: Terry Lambert <tlambert@primenet.com> To: kpielorz@tdx.co.uk (Karl Pielorz) Cc: mi@video-collage.com, current@FreeBSD.ORG, toasty@home.dragondata.com Subject: Re: NFS discovery Message-ID: <199806012318.QAA29727@usr05.primenet.com> In-Reply-To: <35730F59.510D55FB@tdx.co.uk> from "Karl Pielorz" at Jun 1, 98 09:30:17 pm
next in thread | previous in thread | raw e-mail | index | archive | help
> > NFS hung ups are a strange topic, in my experience. People agree > > that they are "bad", but one is not supposed to complain about > > them... > > I remember having a long conversation with a friend a few years back (can I > get any more vague?) - Where he was praising NFS's ability to crash - as it > assures that say your running a program on a remote system, it will either > run to completion - or hang if the server dies... > This I presume works on the assumption that it helps somehow to have a > client that's 'hung' in mid-air (i.e. at least you know if failed) rather > than risking any corruption that might have been caused by the server > disappearing for a while... The program is supposed to hang only until the server comes back up. The problem is the use of TCP, and that the RPC call is not retried after a request ack before a request completion. This is exactly analogous to the FIN_WAIT_2 problem: --> request <-- response 1 <-- response 2 (this one never comes) Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199806012318.QAA29727>