SUMMARY: Problem with Helios Ethershare 2.2

From: Stacy Lindberg (
Date: Fri May 29 1998 - 12:37:25 CDT

Thank you to Blayne Puklich<> for pointing me in the
right direction and in particular, the following web page

The problem it seems was not with Ethershare itself, but with the
Appleshare client on the Macintoshes. MacOS 8 and previous versions have
buggy appleshare clients, versions 3.7.1 and earlier. You should get
version 3.7.2 or greater and have version 1.1.2 or greater of Apple's Open
Transport. You can get the new Appleshare client from
in the following directory:

Thanks also to:
Andrew M Townsend <>
Bismark Espinoza <bismark@alta.Jpl.Nasa.Gov>

I've included all responses below and the orignal message.


Stacy Lindberg
Graphics Network Administrator
The Calgary Sun
Date: Thu, 28 May 1998 14:16:25 -0400
From: Blayne Puklich <>
To: Stacy Lindberg <>
Subject: Re: Problem with Helios Ethershare 2.2

I spent the last three years dealing with Helios, Xinet and IPT, so I might be
able to help you out. Let me ramble on with a bunch of hints, observations
and questions for second here.

What are the MacOS versions? The AppleShare client version? Open Transport
version? MacOS 8 shipped with a terribly crappy version of the client.
Certain versions of OT work pretty poorly. There are some good combinations
of things that you can actually find in the Tech Support area of Xinet's site
( Some Finder versions also suck some bad wind
especially when copying between volumes.

Are the Macintosh 100BaseTX drivers up to date? If they're Asante, some
versions have well-known problems. If they're off-brand, sometimes they just
don't work well ever.

Macs should _always_ be forced full duplex. Heck, until you completely
stabilize a "fass" AppleShare environment, the whole thing should be forced
half, switch ports and NICs.

Do the "latest patches" for Helios include ASIP? Are you using AppleShare IP?

Could there be an AppleTalk routing problem? Sometimes this will look like
poor performance. Other times some Macs will see some Chooser devices and
others a different set of Chooser devices.

What are the transfer rates, reading and writing, from something like
Photoshop, off each server? Other measurable transfers? How about a transfer
between the UNIX machines using something such as rcp or ftp? How about
trying to use Fetch on a Mac between two AppleShare-mounted volumes?

Is it possible that something somewhere could have a crappy cable? How about
a NIC going bad?

How about spanning tree on your switches? How's that set up?

What's the rough network topology? You say servers are on a Linkswitch, but
the Macs?

That's a lot of stuff, but lots can go wrong! Hopefully some of this will at
least point you off in a good direction!

Let me know if I can help more.

>From ATOWNSEND@DOLETA.GOV Thu May 28 13:22:43 1998
Date: Thu, 28 May 1998 15:09:32 -0400
From: Andrew M Townsend <ATOWNSEND@DOLETA.GOV>
Subject: Re: Problem with Helios Ethershare 2.2

Hi - I used Helios years ago and found it very reliable. Have you tried to sniff the traffic to look for anything unusual? Try running snoop on an affected Sparc 10.

Could be something is setting off a broadcast storm, or each packet sent is receiving 100 acks, that sort of thing.

It might help you further identify the problem.


>From bismark@alta.Jpl.Nasa.Gov Thu May 28 15:28:46 1998
Date: Thu, 28 May 98 14:30:52 PDT
From: Bismark Espinoza <bismark@alta.Jpl.Nasa.Gov>
Cc: bismark@alta.Jpl.Nasa.Gov
Subject: Re: Problem with Helios Ethershare 2.2

Make sure the computer running SUNOS and the Macs
do only version 2 nfs.

Check half-duplex/ full-duplex settings and fragmentation flags/MTUs.

Orignal Message
> Hello Managers, I am in need of your assistance.
> We have 4 Sun machines:
> 1 Sparc 10 running SunOS 4.1.3_U1
> - Ethershare 2.2 has only update u0008 installed.
> 2 Sparc 10's running Solaris 2.6
> - All recommended and security bug patches (as of 05/13/98)
> - Ethershare 2.2 with all the latest updates
> 1 E3000 running Solaris 2.5.1
> - All recommended and security bug patches (as of 05/13/98)
> - Ethershare 2.2 with all the lastest updates
> The problem is that when doing a file transfer on a Mac from an Ethershare
> volume (on the E3000) to another Ethershare volume (on one of the Sparc
> 10's), the transfer takes like an hour for 90mb instead of a few minutes.
> I had a chance to reboot the servers and then everything was back to
> normal for a few days but then the problem returned. This time a
> scheduled reboot took place, but the problem remained.
> This only happens when transferring from the ethershare volumes mounted on
> the Mac from different servers. You can copy files to or from the Mac
> from the ethershare volume and you have no problems or to two different
> volumes on the same server.
> I should also mention that the servers have 100mbps connections to a
> linkswitch and that I installed all the latest ethershare updates to fix a
> bug in the afpsrv program that didn't handle volume size reporting to the
> Mac's properly when serving large disks (>4GB). Previous to that, we
> didn't have any problems other than the latter.
> Any help or suggestions would be greatly appreciated!
> Thanks in advance
> Stacy Lindberg
> Graphics Network Administrator
> The Calgary Sun

Blayne Puklich                           
Collective Technologies, a Pencom Company
Minneapolis, MN                              
Pager: 800/759-8888 pin 303-7304 * Home: 612/474-3173
Currently at: University of Minnesota Geometry Center

This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:12:41 CDT