<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=UTF-8" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
-------- Original Message --------<br>
Subject: Re: slow dhcp on 1G backbone.<br>
From: Blake Hudson <a class="moz-txt-link-rfc2396E" href="mailto:blake@ispn.net"><blake@ispn.net></a><br>
To: Users of ISC DHCP <a class="moz-txt-link-rfc2396E" href="mailto:dhcp-users@lists.isc.org"><dhcp-users@lists.isc.org></a><br>
Date: Friday, July 24, 2009 11:34:45 AM<br>
<blockquote cite="mid:4A69E2A5.60603@ispn.net" type="cite">
<meta content="text/html;charset=UTF-8" http-equiv="Content-Type">
-------- Original Message --------<br>
Subject: Re: slow dhcp on 1G backbone.<br>
From: Glen R. J. Neff <a moz-do-not-send="true"
class="moz-txt-link-rfc2396E" href="mailto:neff_glen@emc.com"><neff_glen@emc.com></a><br>
To: Users of ISC DHCP <a moz-do-not-send="true"
class="moz-txt-link-rfc2396E" href="mailto:dhcp-users@lists.isc.org"><dhcp-users@lists.isc.org></a><br>
Date: Friday, July 24, 2009 11:30:55 AM<br>
<blockquote
cite="mid:1248453055.610.11.camel@gneffpc.nas-v6.rtp.dg.com"
type="cite">
<pre wrap="">On Fri, 2009-07-24 at 08:51 -0700, Dean Montgomery wrote:
</pre>
<blockquote type="cite">
<pre wrap="">* HP Procurve 2810 core switch with 2610 switches in the classrooms.
</pre>
</blockquote>
<pre wrap=""><!---->
</pre>
<blockquote type="cite">
<pre wrap="">* No errors or overruns on ifconfig for both client and server.
* Any ideas why dhcp slows down when there is a 1G backbone?
</pre>
</blockquote>
<pre wrap=""><!---->
I'm not familiar with HP Procurve kit, but I really think you need to
look at individual switch port configuration.
I'm not saying it's broken, but often default settings will slow-down
how fast a port initializes, thus running afoul of DHCP.
-G
</pre>
</blockquote>
I agree, diskless clients tend to come up too quickly for default
spanning tree to initialize a port. Portfast on the access ports (ports
where a diskless client connects directly) will likely resolve the
issue.<br>
<br>
-Blake<br>
</blockquote>
Actually, upon review the op stated he is only changing the server
port... so client port initialization would not be affected by the
server moving from 100MB or 1Gbit... <br>
the issue still needs more investigation...<br>
</body>
</html>