CentOS - Jun 2006 - [CSL #301925] centosplus kernel xfs patches and CVE-2006-1855

Hi All, we've run into a problem with the xfs component in the centosplus 
kernel (2.6.9-34.106.unsupportedsmp) where synchronous writes sometimes fail 
with EAGAIN.  It's causing headaches for DB2 installs and for our database 
research group.  There is a fix reported at:

http://marc.theaimsgroup.com/?l=linux-xfs&m=110808122904352&w=2

...that's supposed to take care of things.  If there is any way these
patches
could be included in a future centosplus kernel, we would greatly appreciate 
it.  We are enjoying the performance lift from xfs on many systems, and hate 
to throw it away to get a database to work.

Also, are there plans to produce a centosplus kernel that corresponds to the 
newest RHEL errata kernel (2.6.9-34.0.1.EL)?

Cheers and thanks,

Dave Thompson
UW-Madison

On Thu, 2006-06-01 at 16:24 -0500, David Thompson wrote:
> Hi All, we've run into a problem with the xfs component in the
centosplus
> kernel (2.6.9-34.106.unsupportedsmp) where synchronous writes sometimes
fail
> with EAGAIN.  It's causing headaches for DB2 installs and for our
database
> research group.  There is a fix reported at:
> 
> http://marc.theaimsgroup.com/?l=linux-xfs&m=110808122904352&w=2
> 
> ...that's supposed to take care of things.  If there is any way these
patches
> could be included in a future centosplus kernel, we would greatly
appreciate
> it.  We are enjoying the performance lift from xfs on many systems, and
hate
> to throw it away to get a database to work.
> 
> Also, are there plans to produce a centosplus kernel that corresponds to
the
> newest RHEL errata kernel (2.6.9-34.0.1.EL)?
> 
> Cheers and thanks,
> 
> Dave Thompson
> UW-Madison
David,

There is a centosplus kernel that has that errata ... it is version:
 kernel-2.6.9-34.107.plus.c4

However ... if the only thing you need from the centosplus kernel is the
xfs ... you should use the normal (non-centosplus) kernel and the
kernel-xfs-module that matches it from here:

http://dev.centos.org/centos/4/testing/i386/RPMS/

or

http://dev.centos.org/centos/4/testing/x86_64/RPMS/

(That code is from SGI and has all the latest patches.)

Thanks,
Johnny Hughes
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
URL:
<http://lists.centos.org/pipermail/centos/attachments/20060601/a2b35f4c/attachment-0003.sig>

Johnny Hughes wrote:
>
>There is a centosplus kernel that has that errata ... it is version:
> kernel-2.6.9-34.107.plus.c4
Excellent.  Is there a general way to map from upstream errata kernels to 
centosplus kernels?  The (apparently??) previous convention of suffixing 
.106.unsupported to the upstream release number seemed to work well, and I 
don't want to either guess or take up mailing list bandwidth asking every
time
upstream revs.
>However ... if the only thing you need from the centosplus kernel is the
>xfs ... you should use the normal (non-centosplus) kernel and the
>kernel-xfs-module that matches it from here:
>
>http://dev.centos.org/centos/4/testing/i386/RPMS/
>
>or
>
>http://dev.centos.org/centos/4/testing/x86_64/RPMS/
Ooooh.  That would be cool.  I'll look into it.

Thanks again.

Dave Thompson
UW-Madison

Johnny Hughes wrote:
>
>There is a centosplus kernel that has that errata ... it is version:
> kernel-2.6.9-34.107.plus.c4
>
>However ... if the only thing you need from the centosplus kernel is the
>xfs ... you should use the normal (non-centosplus) kernel and the
>kernel-xfs-module that matches it from here:
>
>http://dev.centos.org/centos/4/testing/i386/RPMS/
>
>or
>
>http://dev.centos.org/centos/4/testing/x86_64/RPMS/
>
>(That code is from SGI and has all the latest patches.)
Do I understand from this that (for example) kernel-module-xfs-2.6.9-34.0.1.EL-
0.1-2.i686.rpm contains xfs errata this is not present in (the centosplus) 
kernel-smp-devel-2.6.9-34.107.plus.c4.i686.rpm?

Dave