Return-Path: linux-nfs-owner@vger.kernel.org Received: from mx2.netapp.com ([216.240.18.37]:37996 "EHLO mx2.netapp.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751076Ab2IJVD7 convert rfc822-to-8bit (ORCPT ); Mon, 10 Sep 2012 17:03:59 -0400 From: "Adamson, Andy" To: "Myklebust, Trond" CC: Jim Rees , Jeff Layton , "Adamson, Andy" , "linux-nfs@vger.kernel.org" Subject: Re: [PATCH 0/4] RFC Avoid expired credential keys for buffered writes Date: Mon, 10 Sep 2012 21:03:51 +0000 Message-ID: <7851A663D3BF5E41B9E166C2C7A42145102B312E@SACEXCMBX02-PRD.hq.netapp.com> References: <1346961251-2554-1-git-send-email-andros@netapp.com> <20120910145732.4d800f5b@corrin.poochiereds.net> <20120910195650.GA3095@umich.edu> <4FA345DA4F4AE44899BD2B03EEEC2FA908F97C34@SACEXCMBX04-PRD.hq.netapp.com> In-Reply-To: <4FA345DA4F4AE44899BD2B03EEEC2FA908F97C34@SACEXCMBX04-PRD.hq.netapp.com> Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Sender: linux-nfs-owner@vger.kernel.org List-ID: Jeff, Jim, Trond - Thanks for all the comments. I'm looking again at the code and will incorporate your suggestions and do some more testing. Should have results and a modified patch set tomorrow. -->Andy On Sep 10, 2012, at 4:14 PM, Myklebust, Trond wrote: > On Mon, 2012-09-10 at 15:56 -0400, Jim Rees wrote: >> Jeff Layton wrote: >> >> My only concern is this behavior described in the last patch: >> >> "If the buffered WRITE is using a credential key that will expire >> within low watermark seconds, fail the WRITE in nfs_write_begin >> _before_ the WRITE is buffered and return -EACCES to the application." >> >> Shouldn't rejecting the write attempt be the purview of the server? >> >> It seems to me that we'd be best off just switching to synchronous >> writes when we start approaching credential expiration, and letting the >> server handle the case where the credential expires. >> >> It seems to me that the client should just be in the business of >> recognizing when the credential might be about to expire and attempt to >> ensure that we don't end up with a bunch of buffered data in that case. >> >> I don't like it. It's easy to understand why writes might fail if the creds >> are getting stale. It's not so easy to understand why client behavior >> changes in a subtle way, with performance implications. > > We don't know that the creds are stale. We just know that the GSS > session is about to expire. > > -- > Trond Myklebust > Linux NFS client maintainer > > NetApp > Trond.Myklebust@netapp.com > www.netapp.com >