public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: "Amit Gupta" <amit.gupta@amd.com>
To: "Luigi Genoni" <kernel@Expansa.sns.it>
Cc: "Alan Cox" <alan@lxorguk.ukuu.org.uk>,
	"Amit Gupta" <amit.gupta@amd.com>,
	linux-kernel@vger.kernel.org
Subject: Re: arpd not working in 2.4.17 or 2.5.1
Date: Wed, 16 Jan 2002 09:33:48 -0800	[thread overview]
Message-ID: <3C45B97C.9F7466B5@cmdmail.amd.com> (raw)
In-Reply-To: <Pine.LNX.4.44.0201161246490.31902-100000@Expansa.sns.it>


We do have more than 1024 systems dircetly connected using switches.. and I am
working on getting a "lsf linux server master", which requires to open connections
all the machines in almost every second or so. Everthing works fine but as soon as
the master daemon tries to connect to systems >1024.. arp kills it with

Jan 15 08:54:08 lx-master kernel: NET: 464 messages suppressed.
Jan 15 08:54:08 lx-master kernel: Neighbour table overflow.
Jan 15 08:54:08 lx-master last message repeated 9 times
Jan 15 08:54:54 lx-master kernel: NET: 442 messages suppressed.
Jan 15 08:54:54 lx-master kernel: Neighbour table overflow.
Jan 15 08:54:54 lx-master last message repeated 8 times
Jan 15 08:56:24 lx-master kernel: NET: 443 messages suppressed.
Jan 15 08:56:24 lx-master kernel: Neighbour table overflow.

-Amit

Luigi Genoni wrote:

> On Tue, 15 Jan 2002, Alan Cox wrote:
>
> > > But arp>1024 is Very Important, else linux will never be able to talk to more
> > > than 1024 clients !
> > >
> > > Linux is my favourite and I wonder if this limit will kill linux for the race
> > > with Solaris/M$ server market. So pls save me :) and help neighour.c/network
> > > layer in new kernel.
> >
> > ARP applies for local links only. So you need a network you are actively
> > talking to 1024 different hosts directly on. Furthermore all the
> > config items should now be soft anyway. Want more, enable more.
>
> To have this kind of network is not impossible at all, I received mail
> from people at intel who are dealing with around 5000 arp (they say).
> In this situation with arpd there was no sensible performance loss, right
> now there is a big slowdown.



  reply	other threads:[~2002-01-16 17:38 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2002-01-15  1:08 arpd not working in 2.4.17 or 2.5.1 Amit Gupta
2002-01-15 15:02 ` Luigi Genoni
2002-01-15 18:59   ` Amit Gupta
2002-01-15 19:58     ` Alan Cox
2002-01-16 11:48       ` Luigi Genoni
2002-01-16 17:33         ` Amit Gupta [this message]
2002-01-17 19:49           ` kuznet
  -- strict thread matches above, loose matches on Subject: below --
2002-01-15 19:34 Dana Lacoste
2002-01-16 11:50 ` Luigi Genoni
2002-01-16 14:06 Dana Lacoste
2002-01-16 21:15 ` Amit Gupta

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=3C45B97C.9F7466B5@cmdmail.amd.com \
    --to=amit.gupta@amd.com \
    --cc=alan@lxorguk.ukuu.org.uk \
    --cc=kernel@Expansa.sns.it \
    --cc=linux-kernel@vger.kernel.org \
    /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