From mboxrd@z Thu Jan 1 00:00:00 1970 From: Trond Myklebust Subject: Re: PATCH: Support binding to a local IPv4 address when mounting a server. Date: Sun, 22 Feb 2009 14:01:54 -0800 Message-ID: <1235340114.7331.92.camel@heimdal.trondhjem.org> References: <4977C580.4040805@candelatech.com> <633CA802-DD5A-4082-B771-C524D367241F@oracle.com> <497805A7.4070205@candelatech.com> <4978AD75.9090900@candelatech.com> <65D69956-DB67-43A7-9101-9AFB7EC55A9F@oracle.com> <499FB08B.1070200@candelatech.com> <49A0E802.9000209@candelatech.com> <1235329791.7331.75.camel@heimdal.trondhjem.org> <7DBDEC87-CB31-404E-8C4D-2762358387DE@oracle.com> Mime-Version: 1.0 Content-Type: text/plain Cc: Linux NFS Mailing List , Patrick McHardy , Ben Greear To: Chuck Lever Return-path: Received: from mail-out1.uio.no ([129.240.10.57]:54858 "EHLO mail-out1.uio.no" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752523AbZBVWB7 (ORCPT ); Sun, 22 Feb 2009 17:01:59 -0500 In-Reply-To: <7DBDEC87-CB31-404E-8C4D-2762358387DE@oracle.com> Sender: linux-nfs-owner@vger.kernel.org List-ID: On Sun, 2009-02-22 at 15:29 -0500, Chuck Lever wrote: > On Feb 22, 2009, at Feb 22, 2009, 2:09 PM, Trond Myklebust wrote: > > I really dislike this idea of adding routing information at the sunrpc > > level. > > Well I think we have had problems in the past where RPC replies on > multi-homed systems go to the wrong interface and are dropped. Isn't > that why we now have the source address field in rpc_create_args? I fail to see the relevance to this case. We're not talking about RPC callbacks here. > I suspect we might have similar problems with NSM, for instance. See Tom Talpey's talk at Connectathon a few years ago. This is why we now have the 'nsm_use_hostnames' sysctl, and why the statd callback code saves FQDNs rather than ip addresses. Trond