From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: linux-nfs-owner@vger.kernel.org Received: from nm8-vm0.bullet.mail.ird.yahoo.com ([77.238.189.203]:36514 "HELO nm8-vm0.bullet.mail.ird.yahoo.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with SMTP id S1751179Ab1KDPCf convert rfc822-to-8bit (ORCPT ); Fri, 4 Nov 2011 11:02:35 -0400 References: <1320353685.18396.119.camel@lade.trondhjem.org> <20111103211100.GA8393@umich.edu> <1320356241.80563.YahooMailNeo@web24706.mail.ird.yahoo.com> <92DF2E31-FABF-40A5-8F78-89B64363568B@oracle.com> <1320361764.48851.YahooMailNeo@web24708.mail.ird.yahoo.com> <39983D1A-70A8-49A1-A4E2-926637780F75@oracle.com> <1320399858.11675.YahooMailNeo@web24703.mail.ird.yahoo.com> <20111104132050.GB13788@umich.edu> <01668DEE-43F7-464B-9BCF-6E52DF0B5956@oracle.com> <20111104144617.GB911@umich.edu> Message-ID: <1320418951.17239.YahooMailNeo@web24716.mail.ird.yahoo.com> Date: Fri, 4 Nov 2011 15:02:31 +0000 (GMT) From: Lukas Razik Reply-To: Lukas Razik Subject: Re: [BUG?] Maybe NFS bug since 2.6.37 on SPARC64 To: Jim Rees , Chuck Lever Cc: Trond Myklebust , Linux NFS Mailing List In-Reply-To: <20111104144617.GB911@umich.edu> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Sender: linux-nfs-owner@vger.kernel.org List-ID: > Chuck Lever wrote: > >   On Nov 4, 2011, at 9:20 AM, Jim Rees wrote: >   >   > As for a fix...  we're trying to move away from udp transport > anyway.  Maybe >   > someone should figure out a way to get it to work with tcp?  I have zero >   > experience with nfs over udp, at least on linux.  Just for fun, have you >   > tried tcp transport (proto=tcp)? >   >   TCP is a real problem in this environment, because it deals poorly with >   NIC initialization timing issues.  UDP is still the best approach (as long >   as it is retransmitting appropriately).  To support TCP, ultimately what >   we need to do is to introduce serialization to make the kernel wait for >   the NIC to become ready before attempting network activity. > > Agreed, but we're grasping at straws here, and this guy just wants it to > work.  It's something to try. Yes, this could be a fast workaround. I'll test it in some minutes. Thanks for that and till later then...