Re: [9fans] keyfs warnings

[erik quanstrom <quanstro@coraid.com>]

> 
> - overwriting my nvram with bla

i think what you have should work, but i think you'd rather zero
out the whole file:

	dd -if /dev/zero -of /n/9fat/plan9.nvr -count 1

> - truncating the /adm/keys file with echo '' > /adm/keys
> - rebooting, filling in password and secstore phrase
> - killing keyfs
> - starting auth/keyfs on the server console

i don't understand this step.  if keyfs is running on the console
of your auth server, restarting it shouldn't make any difference.

(you should always be dealing with keyfs and changeuser from the
console of the auth server.)

> - creating new keys with auth/changeuser for bootes and another user
> - rebooting
> 
> After that I am able to log in as bootes but not the other user (the
> user is present on the fossil filesystem)

where are you "logging in" from.  i assume, since you mentioned this
is a combined fs/cpu/auth server, that you are drawterming in from
another machine?

> 
> When I now kill/restart keyfs on the server keyfs prints the
> following warnings:
> % auth/keyfs
> keyfs: warning: bad status in key file
> keyfs: warning: bad status in key file
> keyfs: warning: bad status in key file
> 4 keys read
> 
> 1. Shouldn't there be only 2 keys as I have added 2 users?
> 2. When trying to auth/debug the user login still fails
> 3. lc /mnt/keys displays garbage

what is in /mnt/keys on the console after you have just booted the machine?

- erik