Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 11 May 2005 12:08:36 +0530
From:      Subhro <subhro.kar@gmail.com>
To:        Rob <spamrefuse@yahoo.com>
Cc:        FreeBSD Stable <freebsd-stable@freebsd.org>
Subject:   Re: xl(4) & polling
Message-ID:  <4281A86C.5010102@gmail.com>
In-Reply-To: <20050511054441.69092.qmail@web54008.mail.yahoo.com>
References:  <20050511054441.69092.qmail@web54008.mail.yahoo.com>

next in thread | previous in thread | raw e-mail | index | archive | help
This is a multi-part message in MIME format.
--------------070203080108000103080909
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

On 5/11/2005 11:14, Rob wrote:

>>When I have to put up a new FreeBSD box, I start
>>from 100 and start beefing up the number until I
>>find a good balance.
>>    
>>
>
>Hmmm, how do you "find a good balance" ?
>Network access speed vs. lost connections.....?
>
>  
>
Yes. The access times during the top load period and the average load 
period helps me to decide the "ideal value". Also its worthwhile to 
adjust the values:

sysctl -a | grep nmbcluster

They determine the number of buffers available for storing network 
connections.

>
>Interestingly: HZ=1000 is apparently a problem with
>the xl devices (3Com 3c905B-TX), but not with the
>rl devices (RealTek 8139).
>What could cause that difference? Could a difference
>in buffer size on the LAN card cause this?
>  
>
Highly possible. But it also depends on the sysctl values I mentioned 
above.

Regards
S.

--------------070203080108000103080909--



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4281A86C.5010102>