Coda File System

Re: question + comment + doc patches

From: Peter J. Braam <>
Date: Fri, 11 Sep 1998 13:36:16 -0400 (EDT)
Maybe next week (when I return from 4 days in recluse writing NT Coda
kernel code) I can help more with clog: 

All you can do after a vanilla install is (NOT as root -- root cannot

clog admin (note: you are getting tokens for user admin). 
password: changeme

The messages you show may be debugging output after a first attempt to use
clog failed.  Clog admin ought to return (perhaps after 30secs) and then
ctokens should show the tokens. 

As to adding users: a user with his her group membership in principle
resides in the vice.pdb file; vice.pcf is an index into this file for fast
lookups.  You add users by recreating the vice.pdb vice.pcf files from
groups.coda and users.coda.  This is totally bad, and a better tool to
manage Coda groups and users is forthcoming - see this list, the Yale
Protection Server messages.

Passwords, like in Kerberos reside somewhere else.  While you can set up
an initial password file (and must do so to get an administrator) adding
other users and changing their passwords is done with the au program. A
member of System:Aministrator group can edit other peoples passwords.

- Peter -

On Fri, 11 Sep 1998, Jeff Breidenbach wrote:

> With Peter's good help, I seem to have a working 4.6.5 Coda server and
> client. I have high hopes to, someday soon, actually mount a
> volume. At least a df returns
> Filesystem         1024-blocks  Used Available Capacity Mounted on
> /dev/hda2             925095  605100   272202     69%   /
> /dev/hdc10            302463     182   286659      0%   /vice
> /dev/hdc11           3365574    4161  3187333      0%   /vicepa
> Coda                 9000000       0  9000000      0%   /coda
> Question:
>  It is not obvious to me how or where one adds a password for
>  clog to authenticate against. Chapter 11 of the administration
>  manual talks about adding users, but nowhere do we drop in
>  passwords. And authentication with clog seems to demand a password.
>  I think another person on the mailing list faces the same question.
> Comment:
>  After a few failed attempts with clog to authenticate, I too get
>  errors filling up the screen, with
>  [12:59:03]SocketListener:
>   "/usr/src/redhat/BUILD/coda-4.6.5/coda-src/rpc2/sl.c", line 421:
>   DecodePacket(RPC2_BUSY): bad bind seqno
>  It seems easy to reproduce, but I don't know exactly what triggers it.
> Documentation patches:
>  I've noted the sections in the manual that gave me the most trouble,
>  hopefully to make things easier for the next person. I hope this
>  feedback is helpful.
> (1) Chapter 7 of the adinistration guide
> -For Linux:
> -
> -/etc/rc.d/init.d/update.init start
> -/etc/rc.d/init.d/auth2.init start
> +For Linux:
> +
> +/etc/rc.d/init.d/update.init start
> +/etc/rc.d/init.d/auth2.init start
> +/etc/rc.d/init.d/codasrv.init start
> (2) Also in chapter 7
> -The last thing that must be done if an SCM server was being setup so
> -it will function is to create the root volume you specified in
> -vice-setup to be mounted by the client software:
> -
> -       createvol_rep coda:root E0000100 /vicepa
> +One final step is needed to make the SCM server function. Use
> +the root volume you specified in vice-setup in place of 
> +your-root-volume. This step creates the volume, which will 
> +be mountable by the client software.
> +
> +       createvol_rep your-root-volume E0000100 /vicepa
> (3) In chapter 11, there are two typos
> - Execute % pwd2pdb -p /vice/db/user.coda -g /vice/db/groups.coda > /vice/db/vice.pdb
> + Execute % pwd2pdb -u /vice/db/user.coda -g /vice/db/group.coda > /vice/db/vice.pdb
Received on 1998-09-11 13:41:04