On Sun, 2004-05-02 at 22:38, Nathan Scott wrote:
> On Sat, May 01, 2004 at 03:21:23PM -0600, Craig Tierney wrote:
> > On Sat, 2004-05-01 at 12:47, Christoph Hellwig wrote:
> > >
> > > Everything dealing with xfs_buf_t (= mostly metadata, + O_DIRECT data
> > > I/O in 2.4) is handled by xfs_buf.
> > >
> > I am trying to debug a problem with file corruption writing to my xfs
> > filesystem (over nfs) when the server is under heavy load (16+ clients
> > writing simultaneously to different directories). It does happen with
> > different versions under the 2.4 kernel. It doesn't happen wen ext2 or
> > jfs is used as the underlying filesystem. It does happen more often on
> > when faster servers are used, and the corruption is always page aligned
> > (starts at ADDR%4096==0, ends at ADDR%4096=4095).
>
> I'd guess your pagesize and fs blocksize are both 4K?
Yes. ia32 page size is 4k, and I didn't change the blocksize
(not that XFS would let me).
>
> If you read between the lines of Christophs mail, pagebuf isn't
> the right place to be looking for regular file data corruption,
> and probably the mem alloc interfaces are unrelated here also.
> You'll want to focus on the get_block and writepage operations
> in xfs_aops.c, I suspect.
I will look around there. That makes sense.
Craig
|