more testing done

[Tom Maher <tardis@xxxxxxxxxx>]

Continuing to use arla for the client, I've been working on setting up a
test cell.  I've followed the steps outlined in

http://www.transarc.ibm.com/Library/documentation/afs/3.6/unix/en_US/HTML/QkBegin/auqbg005.htm

with the following problems:

1) kas, bos, and klog all completely screw up the terminal they're run
on when they prompt you for a password.  kas and klog seem to accept the
password and do the right thing with them, but bos doesn't.  I think
this problem comes from either lwp/waitkey.c or des/read_pssword.c, but
I'm not sure.  I should check that.

The workaround for bos is to specify the password on the command line.
Obviously, this isn't a viable solution for anything more than
development and debugging.

***

2) The big problem comes when I try a "vos create".  The command itself
spits back this...

root@dori:/usr/afs/bin# vos create dori /vicepa root.afs -noauth
Failed to create the volume root.afs 536870915 
: Input/output error
Error in vos create command.
: Input/output error

and the logs tell me...

Mon Apr  9 22:11:52 2001 Starting AFS Volserver 2.0 (/usr/afs/bin/volserver)
Mon Apr  9 22:15:58 2001 VCreateVolume:  Problem creating Volume information file associated with volume header /vicepa/V0536870915.vol
Mon Apr  9 22:15:58 2001 1 Volser: CreateVolume: Unable to create the volume; aborted, error code 103
Mon Apr  9 22:15:58 2001 : Unknown error: 103

According to volser/volser.p.h

#define VNOVOL          103     /* Volume not attached, doesn't exist, 
                                   not created or not online */

Which jives with the below diagnostic output

root@dori:/usr/afs/logs# vos listvol dori -noauth
Total number of volumes on server dori partition /vicepa: 1 
**** Could not attach volume 536870915 ****

Total volumes onLine 0 ; Total volumes offLine 1 ; Total busy 0

My butt says that this is due to volserver treating the partition like a
special transarc-style partition, but not having kernel support to do
so.  Then again, nothing has complained about invalid syscalls, not bos,
not volserver, not even the kernel, so maybe not.

The problem traces back to VCreateVolume, in vol/vutil.c, where there's
some ifdef AFS_NAMEI_ENV's floating around, but recompiling with that
either on or off didn't change anything.

***

In the next day or so, I'll probably fix the password-getting bug, as it
will drive me insane if I don't, and I think I can actually fix that
one.  I'm pretty sure I'll need the help of someone with great knowledge
of how volserver actually works to fix the volume creation problem.

Again, any feedback from those who, unlike myself, actually know what
they're doing would be greatly appreciated.

-- 
Tom Maher

To Unsubscribe: send mail to majordomo@xxxxxxxxxxx
with "unsubscribe freebsd-afs" in the body of the message