OK, although in the last several kernel versions I've had installed
there have been two coda modules -- one in unsupported that is included
in the redhat kernel and one that we install in the regular tree -- with
this latest update modprobe decided that the unsupported coda module was
the one it wanted to load.  So I'm resorting to insmod to load the
proper module and 6.0.12 (which I have once again installed) seems to be
working fine.

	Sorry for the false alarm.


On Tue, 2005-10-04 at 11:47 -0600, Patrick Walsh wrote:
> 	Although I though I checked for this, I think modprobe was loading the
> wrong coda.o module and grabbing the unsupported redhat version instead
> of my custom version.  I'm testing things and working on the details...
> 
> 
> On Tue, 2005-10-04 at 11:08 -0600, Patrick Walsh wrote:
> > 	I just attempted an upgrade from coda client 6.0.11 to 6.0.12.  This is
> > with the coda kernel module 6.1 as 6.2 may be broken.  Unfortunately,
> > venus fails on startup with this message:
> > 
> > 10:46:31 WorkerInit: Version Skew with kernel! Get a newer kernel!
> > 10:46:31 WorkerInit: Kernel version is 2
> > 
> > 
> > 	We are using a Redhat enterprise kernel, 2.4.21.  We didn't have this
> > issue in 6.0.11, but in case nothing changed in coda, we did just
> > upgrade our kernel to the latest set of patches, which were minor bug
> > fixes and security updates.  From 2.4.21 to 2.4.21.
> > 
> > 	The check appears to be in worker.cc in the WorkerInit() function.  It
> > wants kernel_version == 3, but apparently my kernel_version is 2.  
> > 
> > 	coda.o loads smoothly.
> > 
> > 	It looks like this check hasn't changed from 6.0.11, but I'm not sure
> > exactly how kernel_version is determined or where that number comes
> > from.  If it comes from coda.o, I'm a bit confused.  I rebuild it from
> > the 6.1 sources because I'm forced to rebuild it with each kernel
> > upgrade.  
> > 
> > 	I'm going to attempt to move back to 6.0.11 to see if the problem goes
> > away.  I'd appreciate any pointers though as to what may have changed in
> > 6.0.12 that could possibly have caused this to happen or if you think it
> > may be the kernel update that's causing the problem.
> > 
> > Thanks,
> > 
-- 
Patrick Walsh
eSoft Incorporated
303.444.1600 x3350
http://www.esoft.com/