Scott L Riggen wrote:
Bob,
Yeah, I read the how-to and manual. It's still foreign to me. It would help
if I knew at what point things are executed.
This is covered in the "Technical Details for the Interminably Curious"
section of the How-To.
It's OK to mix sparc and x86 in a single FOG, but think of your
application environment. Is it truly transparent? Or do users have to do
things somewhat differently on sparc and x86? Do users have their own
binaries (built for one specific platform)? These are the sorts of
considerations that will drive your decision.
What sort of "token maintenance" do you need to do? None is required,
unless you utilize the registration information somehow.
-Bob
I thought I had it working when I had it pointed to user. It seemed silly to
me that it would not switch the user until they had typed in their login name.
I think Token switching is where I really want to be. I just don't see how the
amgh script ties it all in. If I wanted to write my own script what do I
return so that it can switch users to a particular FOG based on token install ?
In my configuration I currently have 3 (and might eventually have 4) types of
machine groups (But only 1 FOG). In the current SR3.1 install we are switching
users to their correct group using a custom Xsetup and utswitch. We switch
users around because the machine groups have a particular use such as regular
users, special users, qa, linux.
In the new 4.2 install I'd like to use amgh and I'm not sure if I want 1 big
failover group like I currently have or 4 smaller groups. I like having 1 big
FOG since it makes token maintenance a lot easier. However, I might be mixing
in some non-sparc machines and I don't think I can have a mix in the same FOG.
I appreciate the response.
Regards,
Scott
Have you read my blog How-To?
This is all described in there:
http://blogs.sun.com/bobd
Look on the right frame for the link to the AMGH HOW-TO Guide.
NFS shared directory or rdist seem like good choices for keeping the
scripts synchronized.
-Bob
Scott L Riggen wrote:
First off I wanted to thank Willam Yang for all the help.
I adjusted amgh to use
/opt/SUNWutref/amgh/lib/libutamghref_token.so ->> libutamghref_token.so.1
I also adjusted my file with the insert_token stuff and it seems to be ok.
My question about the amgh stuff is.
If I wanted to create my own script and change amgh to use it at what point
does that get called. Can I somehow grab the token on insert and lookup the
desired host / FOG from something like a nis map ? What does my script need to
return in order to do some sort of token_insert and move the users to the
correct FOG ?
On tokens my question is more about having multiple masters (1 for each FOG)
How are you keeping them all in sync (using scripts, etc)
Thanks for all the help so far. Sunrays and software can be a pain but they
also have alot of use with users moving all over the place :)
Regards,
Scott
Yep. And use insert_token= instead of token= since you are using registered
tokens. Also you might need to use a different library or script; it sounds
like from the filename of your library that you can only use usernames with
that particular script.
_______________________________________________
SunRay-Users mailing list
[email protected]
http://www.filibeto.org/mailman/listinfo/sunray-users
_______________________________________________
SunRay-Users mailing list
[email protected]
http://www.filibeto.org/mailman/listinfo/sunray-users
