[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: netatalk, ubuntu, and netboot

On 09/09/2010 09:05 PM, Ivan X wrote:

On 9/9/10 5:25 PM, in article A-CdneLBF68wzRTRnZ2dnUVZ_hudnZ2d@giganews.com,
"Steven Hirsch"<snhirsch@gmail.com>  wrote:


That's nice to hear.  Are you sure that urand authentication is supported out
of the box?  The Ubuntu/debian folks had that turned off due to perceived
licensing issues (IIRC).


Not exactly out of the box. What I have discovered is that if I use vanilla
Ubuntu Desktop 10.04, then I use Synaptics Package Manager to install these
six packages:
libdb-4.8-dev
libcrack2
libcrack2-dev
libpam-cracklib
cracklib-runtime
libssl-dev

And then download netatalk 2.0.5 source, and use:
./configure --enable-debian --enable-a2boot --with-cracklib
That all looks familiar. I think those were all mentioned in a readme that I posted along with my modified debian source package. 


and then make the appropriate tweaks to the config files and initscripts, I
can log in from an Apple IIe client (I haven't tried a IIgs yet), and
probably netboot (if I ever get my A2BOOT folder sorted out).

Interestingly, you don't seem to need --with-cracklib to log in with the Log
On v1.0 client on the Workstation Card disk. You do however, need those
libraries for the Log On v1.5 client that is embedded in the Apple //e Boot
Blocks file which is loaded during netboot, as well as the v1.3 client used
during netboot from an Apple IIe Card (for Macintosh LC). The v1.0 client
must permit clear text for login (you can omit uams_randnum.so from
afpd.conf, or have that file missing completely), while the newer ones seem
to require randnum number exchange (and you do need uams_randnum.so).
The IIgs with stock login support definitely needs randnum. At least it always has for me :-). 


I can't vouch one way or the other for the datestamp fix, yet, nor anything
having to do with IIgs.
Try placing a file in the shared volume on the server side and note the modify data. Then, copy it to the IIe in ProDOS and see if it's propagated. That was broken in the most recent netatalk I played with here. 


Once I get netboot working, hopefully before 2011, I'll try all this using
the latest netatalk (2.1.3) to see what holds up.
Really glad to see more folks using this. I hacked in the original AFP "short name" support and A2 authentication code over ten years ago and until recently was wondering if Marsha and I were the only folks who ever used it. 

Steve