[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: [Linux-cachefs] Problems testing Lustre filesystem with fscache / cachefiles



David,

Thanks for the help.  It turns out (though I'm waiting to receive it) there
may be a lustre patch to get me into a modern kernel.  That begs the
question of how to build a 2.6.2x kernel with a current fscache &
cachefiles.

Since I've been taking the lazy approach - using a kernel that already
contained it (albeit an old unstable version) - I need to find instructions
as to how to build late model kernels with the fscache components.  I've
looked through your ftp space and found tarballs and patches, but not
specific info about exactly which kernels the tarballs and patches go with.
Is there a page or document I could read in order to get this straight?

I'll add that I'm not sure yet which kernel.org versions I'll be able to use
yet.  I was told in a vague way that 2.6.25 was probably a possibility, but
as yet that's unconfirmed.

Any advice appreciated,
John


On Wed, Dec 17, 2008 at 6:36 PM, David Howells <dhowells redhat com> wrote:

> John Groves <grovesaustin gmail com> wrote:
>
> > Dec 17 12:50:14 violin cachefilesd[5323]: readdir returned unknown type:
> > errno 4 (Interrupted system call)
>
> That would be:
>
>        if (dirent.d_type == DT_UNKNOWN)
>                oserror("readdir returned unknown type");
>
> Which suggests the backing filesystem didn't indicate file type of a dir
> entry
> through readdir().  Is this from Ext3, or is it from XFS?
>
> > I might consider trying the back port... Do you have a recommendation as
> to
> > what patch to start with?
>
> Start with the latest patches.  Trim away anything that requires weird
> security stuff.
>
> > Also, might it be easier just to back port cachefiles?  I suppose that
> will
> > be sensitive to the readv/aio_read change, etc.  If I could back port a
> > patch that was better, but not complete modern, it might be easier...or
> does
> > that sound like a mess to you?
>
> You can't just backport cachefiles, unfortunately.  The newer FS-Cache
> stuff
> has a load of asynchronicity stuff in there that affects cachefiles too.
>
> One problem you've got with RHEL-5 is that you need to add an extra address
> operation, but that would violate the kABI:-/
>
> David
>
> --
> Linux-cachefs mailing list
> Linux-cachefs redhat com
> https://www.redhat.com/mailman/listinfo/linux-cachefs
>
> _______________________________________________________________
> Pre-Boxing Day Domain Sales: Hosting + Domain = US$4.95/year
> Offer Ends: Dec 31, 2008.        http://www.doteasypromo.com
>

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]