[Cluster-devel] Re: [PATCH 10/24] GFS2: Don't warn when delete inode fails on ro filesystem

[Steven Whitehouse]

Hi,

On Wed, 2009-06-10 at 05:49 -0400, Christoph Hellwig wrote:
> On Wed, Jun 10, 2009 at 09:30:51AM +0100, Steven Whitehouse wrote:
> > If the filesystem is read-only, then we expect that delete inode
> > will fail, so there is no need to warn about it.
> 
> Umm, ->delete_inode should never be called on a read-only filesystem
> by the VFS.
> 

On a single node, thats true. With multiple nodes, some mounted ro and
some mounted rw, sometimes the ro node gets the baton passed to it, if
it happens to have a process holding open an inode thats been unlinked
on the rw node. In that case it has to ignore it and one of the rw nodes
will deal with the deallocation later,

Steve.