SUMMARY: Hi Collision rate using Solaris 2.5.1 & hme
Wed, 25 Feb 1998 09:52:58 +0000

Content-type: text/plain; charset=US-ASCII
Content-description: cc:Mail note part
Content-transfer-encoding: 7bit

Hi All,

Thanks go out to the following for there suggestions.

The solution to my problem was to install the latest patch 104212-09
(hme driver patch) from sunsite, collision rate is down to acceptable
level now.

Apparently the hme driver assumes MAUI is a full duplex device and
runs it at full duplex mode. This will cause excessive collisions and
late collisions in any packet transfer test.

Original Question below:

______________________________ Reply Separator
Subject: Slow ultra30 NFS connection
Author: at Internet-Gateway
Date: 10/02/97 16:50

Hi All,

Looking for suggestions to solve the following problem,

I have a Ultra 30 running under 2.5.1 (Hardware 4/97) which is
connected to our thin ethernet network via the Ultra's MII port, a Sun
MII to AUI converter (X467A) and a Allied Telesyn MX10 10base2

The Network is served by a Sparc2 running SunOs 4.1.1, I have mounted
users home directories and also software directories from it onto the

My problem is that the ultra30 is very slow at reading/writing any
information on the mounted directories, infact it is slower than any
of our older stations on the same network segment.

Also when I initially run any software contained on the mounted
directories (Binary Compatibility Mode), they take an age to start if
even, but on the next try or subsequent tries they start up without
any problem.

I have noticed that there is a large amount of collisions occuring
when the Ultra 30 is accessing the mounted directories. Using netstat
-i on the ultra produced the following

Name Mtu Net/Dest Address Ipkts Ierrs Opkts Oerrs Collis Queue
lo0 8232 loopback localhost 29859 0 29859 0 0 0
hme0 1500 sol-etherfielding 555806 6922 378071 0 17290 0

My first thought is that it is a problem with NFS3 and NFS2, but I
have mounted the directory from the sparc2 using nfs ver=2 and also
proto = udp, also I have noticed that the ultra30 is very slow when
ftping to it.

My other concern is that possibly the arrangement of using the MII
connection and the adapter card and transceiver maybe causing the

Has anybody come across similar problems, if so any fixes.

Thanks for any help in advance


Unix System Manager

Content-type: text/plain; name="RFC822 message headers"; charset=US-ASCII
Content-description: cc:Mail note part
Content-disposition: inline; filename="RFC822 message headers"
Content-transfer-encoding: 7bit

Received: from ( by with
(IMA Internet Exchange 2.12 Enterprise) id 00003613; Mon, 9 Feb 98 20:42:20
Received: from ( []) by (8.6.10/8.6.10) with ESMTP id UAA01743 for
<>; Mon, 9 Feb 1998 20:43:22 GMT
Received: from localhost (daemon@localhost) by (8.8.3/8.8.3) with
SMTP id MAA29479; Mon, 9 Feb 1998 12:58:06 -0600 (CST)
Received: by (bulk_mailer v1.5); Mon, 9 Feb 1998 12:56:45 -0600
Received: (from daemon@localhost) by (8.8.3/8.8.3) id LAA29021
for sun-managers-outbound; Mon, 9 Feb 1998 11:29:03 -0600 (CST)
Received: (from listserv@localhost) by (8.8.3/8.8.3) id LAA29014
for smmsgs-out; Mon, 9 Feb 1998 11:28:59 -0600 (CST)
Received: from ( []) by (8.8.3/8.8.3) with SMTP id LAA29007 for
<>; Mon, 9 Feb 1998 11:27:58 -0600 (CST)
Received: from ccMail by
(IMA Internet Exchange 2.12 Enterprise) id 00003371; Mon, 9 Feb 1998 17:12:30
Mime-Version: 1.0
Date: Mon, 10 Feb 1997 16:50:16 +0000
Message-ID: <>
Subject: Slow ultra30 NFS connection
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Content-Description: cc:Mail note part
Precedence: bulk