[OpenAFS] recursive address error

Todd T. Fries todd@fries.net
Fri, 09 Dec 2005 09:54:34 -0600


Since nobody had any thoughts on this, I'll follow up for future list 
archive searches...

I did end up finding the solution, though I'm not entirely pleased to 
have found it, and I'm still not quite sure what to make of it.

After having wiped my laptop clean of any openafs server bits, and set 
it up several times to no avail, I finally figured out that there is one 
thing I can change which makes it work just fine, and if I put that 
change back, it does not work until I remove that change.

This change I am talking about is adding an IP onto a gif tunnel.

If I have an ethernet IP of 10.0.0.93 and setup the server to work, it 
works fine.

If I add a gif tunnel IP of 10.9.9.2, and do a 'bos shutdown localhost 
-localauth' and then a 'bos startup localhost -localauth' it falls into 
the below pattern.

If I remove the gif tunnel IP and then redo the 'bos shutdown..' 'bos 
startup...' it works again.

FYI.

Todd T. Fries wrote:
> I setup an openafs cell on one system on the inside of my network, 
> running on OpenBSD.
> 
> Things worked fine, I've moved several gigabytes of data onto it.
> 
> I decided I wanted to try making a second server on my laptop, so I 
> could have RO volumes available when elsewhere to cut down on bandwidth 
> and perhaps even be available when offline.
> 
> So I succeeded in setting up my laptop to be an openafs server 
> internally, and since have roamed to several clients and back, finding 
> that the openafs servers were not so happy on my laptop anymore.
> 
> Through several trials and errors, I've finally realized I have a 
> recursive address error situation.
> 
> The fileserver says this:
> Thu Dec  8 11:31:51 2005 File server starting
> Thu Dec  8 11:31:52 2005 afs_krb_get_lrealm failed, using fries.net.
> Thu Dec  8 11:31:52 2005 VL_RegisterAddrs rpc failed; The IP address 
> exists on a different server; repair it
> Thu Dec  8 11:31:52 2005 VL_RegisterAddrs rpc failed; See VLLog for details
> Thu Dec  8 11:31:52 2005 Fatal error in library initialization, exiting!!
> 
> Then the bosserver starts a slavager:
> 12/08/2005 11:31:52 SALVAGING FILE SYSTEM PARTITION /vicepa (device=vicepa)
> 12/08/2005 11:32:08 ***Forced salvage of all volumes on this partition***
> 12/08/2005 11:32:08 3 nVolumesInInodeFile 84
> 12/08/2005 11:32:08 CHECKING CLONED VOLUME 536870925.
> 12/08/2005 11:32:08 openbsd.ports.readonly (536870925) updated 
> 12/07/2005 17:45
> 12/08/2005 11:32:09 totalInodes 74386
> 12/08/2005 11:32:09 SALVAGING VOLUME 536870961.
> 12/08/2005 11:32:09 blue.test (536870961) updated 11/29/2005 08:32
> 12/08/2005 11:32:11 totalInodes 44468
> 12/08/2005 11:32:16 Salvaged blue.test (536870961): 44464 files, 486125 
> blocks
> 12/08/2005 11:32:16 SALVAGING OF PARTITION /vicepa COMPLETED
> 
> Which completes ok, but then the fileserver is unhappy with the address 
> on the laptop again, and so it reiterates the above, ad nauseum.
> 
> How do I make the the fileserver happy, and how do I debug this to know 
> what it is trying to do so I can make it happy?
> 
> Thanks,

-- 
Todd Fries .. todd@fries.net

  _____________________________________________
|                                             \  1.636.410.0632 (voice)
| Free Daemon Consulting, LLC                 \  1.405.227.9094 (voice)
| http://FreeDaemonConsulting.com             \  1.866.792.3418 (FAX)
| "..in support of free software solutions."  \          250797 (FWD)
|                                             \
  \\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\

               37E7 D3EB 74D0 8D66 A68D  B866 0326 204E 3F42 004A
                         http://todd.fries.net/pgp.txt