On Fri, 28 Jul 2000, Rajagopal Ananthanarayanan wrote:
> Thomas Graichen wrote:
> > i'am currently trying to use it in non critical but productive
> > situations (it gets used but it's no problem if it breaks :-)
> > so far and it seems to be useable but it looks like it still
> > has some seroius problems - for instance - i sometimes boot
> > with a etx2 rootfs (i ususally use xfs for /) and run the
> > repair_xfs and it finds problems in cleanly unmounted
> > filesystems ... so maybe xfs is not ready for rootfs
> > or repair_xfs is not finished so far ... so is this
> > problem expected - or should i mail the errors i
> > get - if - to someone ?
>
> There was a problem with inodes on disk getting corrupted.
> This was fixed a couple of days back:
> ...
> To quickly check if you have the fix, look in vmalloc.h. It
> should have the last argument to vmalloc_area_pages as
> "struct page ***pages", i.e with a triple indirection. The
> buggy code had only a double indirection.
>
> One other known and still open bug is:
>
> BUG 797419 - xfs_iget goes recursive and dies a horrible death
>
> but you would get a kernel oops when that happens, so
> I don't believe you are seeing it.
>
> Could you please ensure that you have the vmalloc fix
> and re-run your tests? If you still see problems, please
> send a report.
looks like it was the version with the bug - i have updated now
and will let you know if i find anything else interesting
thanks for the quick response
t
--
thomas.graichen@xxxxxxxxxxxxx
technical director innominate AG
clustering & security networking people
fon: +49.30.308806-13 fax: -77 web: http://innominate.de pgp: /pgp/tgr
|