All of lore.kernel.org
 help / color / mirror / Atom feed
From: Trond Myklebust <trond.myklebust@primarydata.com>
To: Christoph Hellwig <hch@infradead.org>
Cc: linux-nfs@vger.kernel.org, Olga.Kornievskaia@netapp.com
Subject: Re: [PATCH 1/2] NFSv4.1: Ask for no delegation on OPEN if using O_DIRECT
Date: Wed, 4 Feb 2015 07:32:43 -0500	[thread overview]
Message-ID: <CAABAsM4Sy+252QAhwOk_6RM6rGq0g4AxtHA2Yp0dovKeN6At7A@mail.gmail.com> (raw)
In-Reply-To: <20150204081624.GA16543@infradead.org>

On Wed, Feb 4, 2015 at 3:16 AM, Christoph Hellwig <hch@infradead.org> wrote:
> On Tue, Feb 03, 2015 at 04:59:43PM -0500, Trond Myklebust wrote:
>> If we're using NFSv4.1, then we have the ability to let the server know
>> whether or not we believe that returning a delegation as part of our OPEN
>> request would be useful.
>> The feature needs to be used with care, since the client sending the request
>> doesn't necessarily know how other clients are using that file, and how
>> they may be affected by the delegation.
>> For this reason, our initial use of the feature will be to let the server
>> know when the client believes that handing out a delegation would not be
>> useful.
>> The first application for this function is when opening the file using
>> O_DIRECT.
>
> I can't see why O_DIRECT openers would not want to use a delegation, can you
> explain your rationale?
>

Delegations are all about allowing the NFS client to cache data
aggressively, and notifying when it is no longer safe to do so. That
is clearly not of interest to an application using O_DIRECT, since it
is by definition managing the data cache (if there is one) instead of
the NFS client. We don't share delegation state with userspace and
even if we did, there are no existing applications out there that are
capable (or even interested) of taking advantage of it.

You can argue that the client could still use the delegation to cache
metadata and open/lock state, but most of the users of O_DIRECT of
which I'm aware tend to be data intensive, and not very metadata/state
intensive. So why burden both the server and the with that extra state
management?

Cheers
  Trond

  reply	other threads:[~2015-02-04 12:32 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-02-03 21:59 [PATCH 1/2] NFSv4.1: Ask for no delegation on OPEN if using O_DIRECT Trond Myklebust
2015-02-03 21:59 ` [PATCH 2/2] NFSv4.1: Ask for no delegation on OPEN if already holding one Trond Myklebust
2015-02-03 22:47   ` Kornievskaia, Olga
2015-02-04  1:04     ` Trond Myklebust
2015-02-05 14:07   ` Christoph Hellwig
2015-02-05 14:49     ` Trond Myklebust
2015-02-04  8:16 ` [PATCH 1/2] NFSv4.1: Ask for no delegation on OPEN if using O_DIRECT Christoph Hellwig
2015-02-04 12:32   ` Trond Myklebust [this message]
2015-02-05 11:27     ` Christoph Hellwig

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=CAABAsM4Sy+252QAhwOk_6RM6rGq0g4AxtHA2Yp0dovKeN6At7A@mail.gmail.com \
    --to=trond.myklebust@primarydata.com \
    --cc=Olga.Kornievskaia@netapp.com \
    --cc=hch@infradead.org \
    --cc=linux-nfs@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.