DMARC-Filter: OpenDMARC Filter v1.4.1 sourceware.org 4D61D3857412
Date: Thu, 10 Jun 2021 12:23:03 +1000
From: Duncan Roe <duncan_roe AT optusnet DOT com DOT au>
To: cygwin AT cygwin DOT com
Subject: Re: odd prob for cygwin 'dd' from a character device on network disk
Message-ID: <20210610022303.GA14223@dimstar.local.net>
Mail-Followup-To: cygwin AT cygwin DOT com
References: <60C168AF DOT 4070400 AT tlinx DOT org>
MIME-Version: 1.0
Content-Disposition: inline
In-Reply-To: <60C168AF.4070400@tlinx.org>
User-Agent: Mutt/1.10.1 (2018-07-13)
Precedence: list
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
Errors-To: cygwin-bounces+archive-cygwin=delorie DOT com AT cygwin DOT com
Sender: "Cygwin" <cygwin-bounces+archive-cygwin=delorie DOT com AT cygwin DOT com>

On Wed, Jun 09, 2021 at 06:19:43PM -0700, L A Walsh wrote:
> I've been using a character device on linux in my
> home directory named 'zero' that is a copy of the
> 'zero' device in /dev:
> crwxrwxrw- 1 1, 5 Jun 15  2015 zero
>
> to do read benchmarks using 'dd' (and write benchmarks
> using a file named 'null' thats a copy of /dev/null).
>
> I run it "occasionally", but not on a regular basis, since
> it stays a bit boring.  Nevertheless, ballpark numbers
> consistent with historical norms verify correct network
> function (separate from read/write files).  Read/write speeds
> at last check a few weeks ago were typical with
> reads at 700MB/s and writes at about 300MB/s.
>
> A few hours ago I tried it again due to some strange network
> probs where I seemed to be getting file xfer speeds as low
> as 200K/s.
>
> I tried the bench script, and its half broken now -- because
> I can no longer get anything back from the zero device
> on my server via the network.  Locally, I can do 'dd' from
> /dev/zero (or the zero file) and it takes a few seconds and
> gave about 800MB/s.  So seemed to have been working fine,
> but remotely -- still nada... reads 0 bytes from /dev/zero
> and about 300MB/s write speed.
>
> Can anyone think of anything that might have changed in
> cygwin such that it would know the remote device is a
> "/dev/zero".  Wondered if the 'dd' prog might have changed,
> but just realize it did same with 'cat' as well.
>
> Anyway -- I'm a bit stumped as to possible causes...
> The writing to a remote /dev/null part is still working,
> so not sure why one would change and not the other.
>
> If anyone can think of anything, please let me know -- I'm
> also gonna ping the samba list and maybe my distro list
>
> Thanks!
> -linda
>
nfs / nodev?

-- 
Problem reports:      https://cygwin.com/problems.html
FAQ:                  https://cygwin.com/faq/
Documentation:        https://cygwin.com/docs.html
Unsubscribe info:     https://cygwin.com/ml/#unsubscribe-simple