Subject: Re: Major Netatalk confussion happening!
From: Luke McNeilage (lmn@d2p.com.au)
Date: Thu Aug 03 2000 - 22:31:23 EDT
Steve
I have been suggesting for many months that we pack the whole thing up and
release netatalk V2. As the umich admins seem to have no interest in
furthering the project, and with Mr Sun gone to cobalt ( and if you don't
think that he's there to produce a competing commercial product to Helios
Ethertalk.....) there needs to be a concerted effort by those of us who can
really make this shit work to pull it together.
With regards to your fervor for the umich list: I have had numerous problems
with this list, I get little response to my questions and propositions, and
it is populated mostly by well meaning, but lazy newbies (there is more to
life than RPM). More to the point, I have only just been able to
re-subscribe, because it had just plain disappeared for about two months.
With regards to sourceforge: there is nothing happening, and it is pathetic
that the site does not support IE on the Macintosh, requires 5.01/128bit on
WIN. As much as I hate Microsoft, Netscape really sucks. No forum messages,
not patches, no bugs, no tech manager, NO DOCUMENTATION. What exactly are we
trying to achieve here?
The recent discovery (by some of you) of Webmin is a good case in point. The
first thing I do to any installation is remove linuxconf, and install
Webmin, but I put out a request 6 months ago for any input into a webmin
module, and got nothing back. If you (collectively) manage to write one,
where are going to post it; webmin, umich, thehampton, sourceforge, or
cobalt?
If you are going to look into this, are you going to look at the 1.3, 1.4b,
asun2.1......., asun2.1.4p39b subsection 47 paragraph 9 .... etc, (you get
my point). The difference in the configuration files for asun's new work is
considerable, and while I admire where he is going, the module needs to be
retro-fitable to the installed base.
I would be willing to entertain giving webspace, support, list server, and
even development grants to anybody who wishes to participate in establishing
a real development site.
To begin with the integration of the afp & atalk into a conf that makes
sense, and this is all launched, by a SysV std init. Now I know Mr. Asun's
version 2.1.4 is supposed to be heading down this path, but the rc.d//atalk
& netatalk.conf files in his pre- versions are completely undocumented. And
I think we are all looking for something that runs with the current
'distributions'.
Due to the complete lack of documentation, and the very poor implementation
of routing compatibility, there is a big gap in Netatalk's ability to fit
into networks that are run by stupid NT administrators that insist on
ticking the little box under "services for Macintosh" that enables routing
that separates the accountants from the graphic designers.
Let's try and get over this....
As the documentation for netatalk is founded in V1.2, it falls very short on
clarity.
Has anybody got an example of how they have configured their system to
support "zones", especially just a configuration to fit into an existing
zone, not the "*" default.
The standard SysV init script that comes with RH.rpm, does not give a
methodology for passing zone information. I have inserted a "ZONE_NAME" in
the /etc/atalk/config file and modified the initd atalk script so that it
now picks up and passes the zone to the nbp-reg
daemon atalkd
nbprgstr -p 4 "${ATALK_NAME}:Workstation@${ZONE_NAME}"
nbprgstr -p 4 "${ATALK_NAME}:netatalk@${ZONE_NAME}"
This works except for the AFPServer component, and the atalk fails.
The problem is now; if afpd reads from its own /etc/atalk/afpd.conf file,
what is all this about?
if [ ${AFPD_RUN} = yes ] ; then
daemon afpd ${AFPD_NOCLR} ${AFPD_NOAFS} ${AFPD_NOKRB} \
${AFPD_NOGUEST} -g ${AFPD_GUEST} \
-c ${AFPD_MAX_CLIENTS} -n ${ATALK_NAME}
The suggestion from the documentation is that I should be able to launch
afpd -n @whatever_zone. This should imply that I could also insert the
${ZONE_NAME} after the -n call; guess what; not..
This is my afpd.conf
Box -tcp -ddp -nocleartxt -address 10.1.0.254
But as there appears to be no indication or documentation that this file
ever gets called, what has it really to do with the afp. Reading from man,
there should be a mechanism to insert "Box:AFPServer@all" into this line,
but it doesn't seem to work.
Neither does the atalk.conf
eth0 -addr 1.233 -zone "all"
And lastly, if anybody has ever run Appletalk services with the
back-grounding turned off,
do you get this line, and what does it mean?
Starting AppleTalk services: ./atalk: [: =: unary operator expected
I won't even begin on papd..
Thanks in advance.....
============================================================================
Luke McNeilage
Technical Director
D2P (Australasia) Pty Ltd
Email: lmn@d2p.com.au
Web: www.d2p.com.au
Australia's leading file delivery system for publishing & graphic arts.
Tel: +61 (0)3 9429 3233
Fax: +61 (0)3 9427 0929
Mobile: 0419 512 868
Suite 3/243 Bridge Road
Richmond VIC 3121
Australia
----------
This archive was generated by hypermail 2b28 : Wed Jan 17 2001 - 14:31:50 EST