[73] in Hesiod
Re: How do I use hesiod efficiently?
daemon@ATHENA.MIT.EDU (Randall Winchester)
Mon Dec 23 15:40:40 1991
Date: Mon, 23 Dec 91 15:37:53 -0500
From: rsw@eng.umd.edu (Randall Winchester)
To: jh@efd.lth.se, hesiod@Athena.MIT.EDU
In-Reply-To: Mail from 'jh@efd.lth.se (Joergen Haegg)'
> From jh@baal.efd.lth.se Sun Dec 22 07:46:01 1991
> X-Mailer: Mail User's Shell (7.2.1 12/20/90)
> To: hesiod@Athena.MIT.EDU
> Subject: How do I use hesiod efficiently?
>
> I've unpacked and installed Hesiod version 1.3 (I think :-).
> That's what I found in hesiod.c,v.
I would suggest using the BIND 4.8.3 code along with the patches that
tytso@ATHENA.MIT.EDU put out for ftp on some server other than athena-dist.
I can not remember what its name was, but I still have them and can put
them up for ftp if people want/Ted does not remind us where he put them.
Then just use the libhes.a and hesinfo parts of the MIT Distribution.
>
> Anyway, are there more libraryfunctions available?
> If I would like to use hesiod instead of yp, then I must
> write a number of getpw and getgr functions myself.
> Only getpwnam is defined in the distribution.
Others have answered on this, we too have written our own librariy routines.
>
> And how do people normally use hesiod?
Hesiod is a tool, it is a flexible way for sysadmins to maintain information.
Its "normal use" will grow as your networks complexity grows or as you get
swamped with maintaining boring system files.
I look at it as a database for variable length text. We use it for system
information that we get tired of distributing on a machine by machine
or netgroup by netgroup (Think NIS) basis.
We use it for;
autoconfigurations on bootup (clusterinfo, hostinfo, other creative things)
autoconfigurations on login (MAILHOST, getpwent, etc)
note: these can make onetime queries that setenv or create info files.
this cuts down on redundent requests.
runtime site/cluster/group/host/user dependent requests.
automounts (We use amd, not automounter) and amd_clusters
We plan to use it with;
more system libraries
AFS maintainance.
> Is it common to replace all yp-maps?
The biggest trick here is the "netgroups" database and those things that use
it. The two most important to us when we implemented them where;
1) restricting NFS mounts by groups of machines
We then modified rpc.mountd to call a hes_resolve() routine instead
of the NIS call. This allowes "netgroups" to work in /etc/exports.
An example;
prompt% cat /etc/exports
/a/egypt/home/elves -access=ENG_unix
/export/os -access=DEV_mach
/export/software -access=DEV_mach
/export/glue1_1 -ro,access=ENG_unix
prompt% hesinfo bacchus.eng.umd.edu ngbyhost
EEserv_mach,EE_mach,ee_cluster,ENG_unix,GLUE_unix
2) restricting logins by groups of users.
Login.c (and others like xdm) were modified to use a /etc/restrict
file which contains a "NIS like" format for allowing or denying users
by groups.
An example;
prompt% cat /etc/restrict
-@Restrict
+@Elves
+@Special
+@Sys
+@Tools
+@EEdept
-
prompt% hesinfo try ngbyuser
eegrad4,EEgrads,EEdept
We used /usr/etc/yp/revnetgroup {-h,-u} | some_script to create a
netgroup_by_host, netgroup_by_user, and clustername.
> (I really would love to do that, our system has grown to much now :-)
A note here about hesiod_service/name_service. The size of the named
databases get pretty large when you start putting information for
every user's home, mailbox, and pwent. There are usually a lot more users
then machines, so some consideration for RAM on ther hesiod servers is
needed. I am assuming thousands of users here.
>
> --
> Joergen Haegg jh@efd.lth.se postmaster@efd.lth.se
> System manager @ efd 046-107492
> Lund Institute of Technology E-huset, DDG, Ole R|mers v. 3
> 221 00 LUND, Sweden
> The easiest way to figure the cost of living is to take your income and
> add ten percent.
Try it you will like it.
Randall Winchester
Systems Administrator
College of Engineering
University of Maryland
College Park, Maryland (USA)
