Received-SPF: pass (google.com: best guess record for domain of linux-nfs-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
References: <CANpxKHHAngJWaW0uiOfUnBCi2Tv36h33dVg8pAuAmMcbnYzzJQ@mail.gmail.com>
 <1589957450.8701645.1572273893609.JavaMail.zimbra@desy.de>
In-Reply-To: <1589957450.8701645.1572273893609.JavaMail.zimbra@desy.de>
From:   Naruto Nguyen <narutonguyen2018@gmail.com>
Date:   Tue, 29 Oct 2019 14:19:28 +0700
Message-ID: <CANpxKHFXgM2nuZ9WEME_XA08U0xWDV5SQuS1Vy3cx4=_vS6oeQ@mail.gmail.com>
Subject: Re: NFS latency and throughput minimum requirement
To:     "Mkrtchyan, Tigran" <tigran.mkrtchyan@desy.de>
Cc:     linux-nfs <linux-nfs@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-nfs-owner@vger.kernel.org
Precedence: bulk

Hi Tigran,

Thanks a lot for your advice. I will try to find out the bottlenecks first.

Brs,
Naruto

On Mon, 28 Oct 2019 at 21:44, Mkrtchyan, Tigran
<tigran.mkrtchyan@desy.de> wrote:
>
> Hi Naruto,
>
> NFS performance is a function with multiple variables:
>
>   - CPU
>   - kernel
>   - TCP
>   - Networks
>   - Disk subsystem
>   - File system
>
> and every component has it's own set of knobs to turn...
>
> I will inter the question: What is expected latency and throughput
> of your applications? Can you identify the bottlenecks? At the best
> try to reproduce the load with application itself or fio and then
> go one-by-one and tweak the performance.
>
> Regards,
>    Tigran.
>
> ----- Original Message -----
> > From: "Naruto Nguyen" <narutonguyen2018@gmail.com>
> > To: "linux-nfs" <linux-nfs@vger.kernel.org>
> > Sent: Monday, October 28, 2019 6:25:24 AM
> > Subject: NFS latency and throughput minimum requirement
>
> > Hi everyone,
> >
> > Could you please advice me how to identify the minimum NFS requirement
> > of latency and throughput on a specific system? I have some test cases
> > to test performance of NFS and tune the system to have better NFS
> > performance but I do not know how good or bad with my
> > measurement/result.
> >
> > Thanks,
> > Brs,
> > Naruto