[Unionfs] unionfs 2.0 in enterprise environment

[Jefferson Ogata]

On 07/05/07 15:13, Joe Landman wrote:
>   You should have some swap configured.  The crash says "Unable to
> handle kernel paging request at ..."  which usually means that you don't
> have swap configured, but need it.  Sometimes it also means your swap
> device went away, or is not responding.  You can try to turn swapping
> off in the kernel, though that might cause other problems.

I don't think a failed page fault handler necessarily means swap at all.
And unless the system has a tiny amount of physical memory, the idea
that the kernel would be needing swap during the boot process doesn't
seem plausible to me.

The stack trace shows EIP in nfs_dentry_iput. This implies the fault
occurred while reading the content of an NFS directory. Higher in the
stack trace we see reference to unionfs_mkdir and utimes. Looks like
something did utimes on an object in a unionfs branch.