[Gluster-devel] [Feature request]: Regression to take more patches in single instance

[Niels de Vos]

On Wed, Jul 31, 2013 at 05:31:18PM +0100, Justin Clift wrote:
> On 31/07/2013, at 2:26 PM, Kaleb S. KEITHLEY wrote:
> > On 07/31/2013 09:17 AM, Kaleb S. KEITHLEY wrote:
> >> On 07/31/2013 08:44 AM, Anand Avati wrote:
> >>> 
> >>>        Multiple VMs will solve the problem, but I guess we need to
> >>>        figure out
> >>>        how to get a bigger box etc.
> >>> 
> >>> 
> >>>    Can the "worker" build machines be behind a firewall? I'm working on
> >>>    getting the old Sunnyvale lab machines on-line in our new lab. Can
> >>>    we use some of those?
> >>> 
> >>> 
> >>> That should work, I think!
> >> 
> >> I've escalated the ticket to get those machines on-line. We have some
> >> pretty beefy machines to throw at it that.
> > 
> > ... should be good for running lots of VMs.
> 
> Please be careful of this bug when you're getting the machines online:
> 
>   https://bugzilla.redhat.com/show_bug.cgi?id=981741
> 
> Running the regression test on an unpatched readdirplus enabled kernel
> will take out the box every time (kernel panic).
> 
> RHEL/CentOS 6.3 GA (no kernel updates applied) is known to work.  Niels
> has built some recent RHEL 6.4 kernels with a fixed readdirplus too,
> which should also work.  (get them from him?) :)

I've opened up that bug and added a link to download x86_64 test 
packages and patch. Anyone who hits this problem and has a valid 
entitlement to get support from Red Hat, is encouraged to open a support 
case through the Red Hat Customer Portal to get notified when the 
appropriate errata is released.

Cheers,
Niels