Re: Help with troubleshooting performance issue

Its a shame everyone is dropping suppport for pptp.  Nothing else is easier to setup on Windows clients.


Sent from my Verizon Wireless 4G LTE Smartphone



-------- Original message --------
From: James Cameron <[email protected]>
Date: 04/02/2013 1:04 AM (GMT-06:00)
To: Jafaruddin Lie <[email protected]>
Cc: [email protected]
Subject: Re: [Poptop-server] Help with troubleshooting performance issue


On Tue, Apr 02, 2013 at 04:37:06PM +1100, Jafaruddin Lie wrote:
> And if I get this correct, this could be because of the contention
> between pptp daemon and other user-space processes?

Yes, that is possible, but you would need to instrument the daemon to
prove it.

> I did top command while the transfer was going.  On both machines,
> the load average didn't go above 0.08 for both machines, and CPU %us
> on both machines didn't go above 9%.

Sounds like the only contention will be between SCP, FTP and PPTP.
You might try operating SCP and FTP on another host instead of the
host running PPTP.  That is how most PPTP deployments would work.

(In other words, using SCP or FTP on the same host may be invalidating
your results, since PPTP is rarely used in that way.)

> Sorry to be banging on about this again, but I do want to re-confirm.
> Do you mean you won't be suprised if ppptp might not sustaion a
> bandwidth of more than a tenth of native link bandwidth?
> From what you wrote, I gather that pptp *should* be able to sustain
> traffic more than a tenth of native link bandwidth.

Actually, I would not be surprised either way.  CentOS 6 is reasonably
recent, and the changes in kernel since my previous tests would
invalidate my tests and my experience.  I doubt there has been any
rigorous testing of PPTP with recent releases of CentOS or Ubuntu,
because nobody should care much about PPTP anymore, as it is so
insecure.

So I shall re-estimate and restate; I would not be surprised if the
bandwidth via tunnel is between 0.05 and 0.95 of the native bandwidth.

I would be surprised if it was greater than 0.95.  I would be very
critical of any result that shows a value greater than 1.00, or less
than 0.05.

Greg's results fit nicely into my not-surprised zone.

> I have also tested this with a Ubuntu server on different machine
> (same deal), with Windows client, and OS X client.  They all give me
> almost the same result, some slightly faster than the other.

That's great.  Have you tried a Windows server or an OS X server?

--
James Cameron
http://quozl.linux.org.au/

------------------------------------------------------------------------------
Own the Future-Intel(R) Level Up Game Demo Contest 2013
Rise to greatness in Intel's independent game demo contest. Compete
for recognition, cash, and the chance to get your game on Steam.
$5K grand prize plus 10 genre and skill prizes. Submit your demo
by 6/6/13. http://altfarm.mediaplex.com/ad/ck/12124-176961-30367-2
_______________________________________________
Poptop-server mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/poptop-server

------------------------------------------------------------------------------
Own the Future-Intel(R) Level Up Game Demo Contest 2013
Rise to greatness in Intel's independent game demo contest. Compete 
for recognition, cash, and the chance to get your game on Steam. 
$5K grand prize plus 10 genre and skill prizes. Submit your demo 
by 6/6/13. http://altfarm.mediaplex.com/ad/ck/12124-176961-30367-2

_______________________________________________
Poptop-server mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/poptop-server