From: Ken Brown <kbrown@cornell.edu>
To: cygwin@cygwin.com
Subject: Re: cygrunsrv + sshd + rsync = 20 times too slow -- throttled?
Date: Thu, 26 Aug 2021 11:56:07 -0400 [thread overview]
Message-ID: <d0a8c57d-1ed1-6b4f-c6e7-cbe0e2ec8a1c@cornell.edu> (raw)
In-Reply-To: <20210826062934.54f2f2216021c095bb7ba13b@nifty.ne.jp>
On 8/25/2021 5:29 PM, Takashi Yano wrote:
> On Wed, 25 Aug 2021 13:52:19 -0400
> Ken Brown wrote:
>> On 8/25/2021 7:18 AM, Takashi Yano via Cygwin wrote:
>>> On Tue, 24 Aug 2021 12:49:52 -0700
>>> Chris Roehrig wrote:
>>>> I have a network of Windows, Linux and Mac machines and I use rsync to synchronize various directories between them.
>>>>
>>>> I'm trying to figure out why my rsync transfers are so slow (<4 MB/s) only when the remote endpoint is Cygwin rsync over sshd (with both a Linux or Cygwin rsync client). In all other scenarios, I get the full 100MB/s as expected from gigabit ethernet. This has been an ongoing problem for me for a couple of years over several Windows and Cygwin versions, and I'd like to try to fix it.
>>>>
>>>> If I run rsync --daemon --no-detach under mintty in the foreground on the remote Windows endpoint, I get the full 100 MB/s transfers, so it seems like it has something to do with rsync.exe running in the background under the cygrunsrv+sshd service (which was installed normally using ssh-host-config).
>>>>
>>>> If I do:
>>>> pv /dev/zero | ssh $WINHOST "cat > /dev/null"
>>>> or even
>>>> pv /dev/urandom | ssh $WINHOST md5sum
>>>> I also get the full 100 MB/s transfers, so it doesn't look like sshd itself is being throttled by bandwidth or CPU.
>>>>
>>>> The machines have less than 15% CPU utilization while transferring, with each of the 4 cores less than 30%, so it doesn't look to be CPU issue.
>>>> In Task Manager, sshd.exe and rsync.exe seem to be running normally using only few percent CPU, and show Power Throttling=Disabled, Priority=Normal. Setting their Priority to High doesn't seem to change things.
>>>>
>>>> Looking in Resource Monitor on the remote endpoint, the network usage is pretty much a flat horizontal line at about 18 Mbps (2.5 MB/s), so it sure looks to me as if rsync is somehow being bandwidth-throttled\b when run in the background under cygsshd.
>>>>
>>>> It's almost as if rsync has an implicit --bwlimit override when it is run from cygrunsrv+sshd (I've tried --bwlimit=0 on the client which makes no difference).
>>>>
>>>>
>>>> Any ideas? Not sure where to go from here.
>>>
>>> In cygwin, just scp is very slow.
>>>
>>> The transfer speed in my environment is as follows.
>>> The tests were done with 100MB of test.dat file.
>>>
>>> (1-1) From cygwin-PC,
>>> [yano@cygwin-PC ~]$ scp test.dat yano@linux-server:.
>>> yano@linux-server's password:
>>> test.dat 100% 100MB 4.0MB/s 00:24
>>> [yano@cygwin-PC ~]$ scp yano@linux-server:test.dat .
>>> yano@linux-server's password:
>>> test.dat 100% 100MB 8.0MB/s 00:12
>>>
>>> (1-2) From linux-server,
>>> yano@linux-server:~$ scp yano@cygwin-PC:test.dat .
>>> yano@cygwin-PC's password:
>>> test.dat 100% 100MB 4.0MB/s 00:24
>>> yano@linux-server:~$ scp test.dat yano@cygwin-PC:.
>>> yano@cygwin-PC's password:
>>> test.dat 100% 100MB 4.1MB/s 00:24
>>>
>>>
>>> I looked into this problem, and noticed that this is caused
>>> by cygwin pipe implementation. Pipe in cygwin is configured
>>> with FILE_FLAG_OVERLAPPED.
>>>
>>> If the pipe is configured without FILE_FLAG_OVERLAPPED,
>>> the transfer speed is much improved as follows.
>>>
>>>
>>> (2-1) From cygwin-PC,
>>> [yano@cygwin-PC ~]$ scp test.dat yano@linux-server:.
>>> yano@linux-server's password:
>>> test.dat 100% 100MB 85.5MB/s 00:01
>>> [yano@cygwin-PC ~]$ scp yano@linux-server:test.dat .
>>> yano@linux-server's password:
>>> test.dat 100% 100MB 69.7MB/s 00:01
>>>
>>> (2-2) From linux-server,
>>> yano@linux-server:~$ scp yano@cygwin-PC:test.dat .
>>> yano@cygwin-PC's password:
>>> test.dat 100% 100MB 80.1MB/s 00:01
>>> yano@linux-server:~$ scp test.dat yano@cygwin-PC:.
>>> yano@cygwin-PC's password:
>>> test.dat 100% 100MB 57.7MB/s 00:01
>>>
>>> I am not sure why this happens and how to fix this.
>>
>> A couple years ago I had an idea for changing the pipe implementation to avoid
>> overlapped I/O:
>>
>> https://cygwin.com/pipermail/cygwin-patches/2019q2/009393.html
>> https://cygwin.com/pipermail/cygwin-patches/2019q2/009423.html
>>
>> I never followed up on it. But if you think it might help with this problem, I
>> could dust it off and try to finish it.
>
> Interesting.
>
> It will be also helpfull for:
> https://cygwin.com/pipermail/cygwin/2021-March/247987.html
> which seems to be the same issue with
> https://stackoverflow.com/questions/10385424/good-alternatives-to-cygwin-cygwin-doesnt-support-natively-support-win32-app
>
> However, I wonder why scp dislikes overlapped I/O.
I agree that it would be good to understand this. When I first proposed the
change, I was thinking in terms of code simplification. If it also improves
performance (which we don't know yet), it becomes a higher priority, but in that
case it would be nice to understand why it improves performance.
Ken
next prev parent reply other threads:[~2021-08-26 15:56 UTC|newest]
Thread overview: 86+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-08-24 19:49 Chris Roehrig
2021-08-24 23:43 ` Mark Geisert
2021-08-25 0:35 ` Chris Roehrig
2021-08-25 2:12 ` Chris Roehrig
2021-08-25 2:02 ` NightStrike
2021-08-25 3:20 ` Mark Geisert
2021-08-25 17:24 ` Chris Roehrig
2021-08-25 11:18 ` Takashi Yano
2021-08-25 17:31 ` Chris Roehrig
2021-08-25 17:52 ` Ken Brown
2021-08-25 18:18 ` Chris Roehrig
2021-08-25 18:51 ` Chris Roehrig
2021-08-26 15:44 ` Ken Brown
2021-08-25 20:33 ` Mario Emmenlauer
2021-08-26 15:47 ` Ken Brown
2021-08-25 21:29 ` Takashi Yano
2021-08-26 15:56 ` Ken Brown [this message]
2021-08-26 22:18 ` Ken Brown
2021-08-27 11:24 ` Takashi Yano
2021-08-27 16:00 ` Ken Brown
2021-08-27 17:21 ` Takashi Yano
2021-08-28 2:00 ` Takashi Yano
2021-08-28 3:03 ` Takashi Yano
2021-08-28 8:43 ` Corinna Vinschen
2021-08-28 9:41 ` Takashi Yano
2021-08-28 11:58 ` Corinna Vinschen
2021-08-28 15:43 ` Takashi Yano
2021-08-28 20:55 ` Ken Brown
2021-08-29 8:41 ` Takashi Yano
2021-08-29 22:42 ` Ken Brown
2021-08-29 10:27 ` Corinna Vinschen
2021-08-29 9:07 ` Takashi Yano
2021-08-29 15:57 ` Ken Brown
2021-08-29 19:24 ` Chris Roehrig
2021-08-29 22:24 ` Ken Brown
2021-08-30 23:58 ` Chris Roehrig
2021-08-31 19:05 ` Ken Brown
2021-08-31 19:53 ` Chris Roehrig
2021-08-31 20:23 ` Chris Roehrig
2021-08-31 21:29 ` Brian Inglis
2021-09-01 21:11 ` Chris Roehrig
2021-09-02 15:25 ` Ken Brown
2021-09-02 19:03 ` Chris Roehrig
2021-09-03 17:26 ` Ken Brown
2021-09-03 19:55 ` Corinna Vinschen
2021-09-03 20:59 ` Chris Roehrig
2021-09-04 8:32 ` Achim Gratz
2021-09-04 16:45 ` Brian Inglis
2021-09-05 8:18 ` Achim Gratz
2021-09-05 15:11 ` Brian Inglis
2021-09-06 10:42 ` Achim Gratz
2021-09-04 22:37 ` mmap failure [was: cygrunsrv + sshd + rsync = 20 times too slow -- throttled?] Ken Brown
2021-09-04 22:54 ` Ken Brown
2021-09-04 22:58 ` Ken Brown
2021-09-05 0:04 ` Ken Brown
2021-09-05 13:24 ` Ken Brown
2021-09-06 15:32 ` Corinna Vinschen
2021-09-06 17:12 ` Ken Brown
2021-09-06 17:38 ` Ken Brown
2021-09-06 17:43 ` Eliot Moss
2021-09-06 17:59 ` Corinna Vinschen
2021-09-06 18:07 ` Corinna Vinschen
2021-09-06 18:40 ` Ken Brown
2021-09-06 20:52 ` Peter Dons Tychsen
2021-09-06 20:54 ` Peter Dons Tychsen
2021-09-06 21:24 ` Ken Brown
2021-09-06 21:31 ` Ken Brown
2021-09-07 3:34 ` Brian Inglis
2021-09-07 16:28 ` Ken Brown
2021-09-07 21:52 ` Lavrentiev, Anton (NIH/NLM/NCBI) [C]
2021-09-07 22:44 ` Ken Brown
2021-09-08 6:14 ` Sam Edge
2021-09-08 8:18 ` Corinna Vinschen
2021-09-07 21:41 ` Peter Dons Tychsen
2021-09-04 23:18 ` Brian Inglis
2021-08-29 19:37 ` cygrunsrv + sshd + rsync = 20 times too slow -- throttled? Takashi Yano
2021-08-29 21:09 ` Ken Brown
2021-08-29 21:04 ` Ken Brown
2021-08-30 0:13 ` Takashi Yano
2021-08-30 0:22 ` Takashi Yano
2021-08-30 2:15 ` Ken Brown
2021-08-30 8:02 ` Takashi Yano
2021-08-28 15:17 ` Ken Brown
2021-09-16 22:00 ` Keith Christian
2021-09-16 22:48 ` Ken Brown
2021-09-16 22:58 ` Keith Christian
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=d0a8c57d-1ed1-6b4f-c6e7-cbe0e2ec8a1c@cornell.edu \
--to=kbrown@cornell.edu \
--cc=cygwin@cygwin.com \
/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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).