Not yet.

We just got another kernel to test this morning and so far in limited testing 
(not with WAS yet), it seems to be ok.  Interestingly, this kernel supposedly 
has this patch removed:

* Fri Mar 07 2014 [email protected]
- revert patch 
patches.fixes/mm-mmvmscan-only-evict-file-pages-when-we-have-plenty.patch
  bnc#859225

Puzzling...




-----Original Message-----
From: Linux on 390 Port [mailto:[email protected]] On Behalf Of Leland 
Lucius
Sent: Friday, March 07, 2014 12:15 PM
To: [email protected]
Subject: Re: [LINUX-390] SLES 11 SP3 problems

Hi Marcy,

Sorry to bother you about this, but did you ever get a resolution?
We've seen this as well and haven't yet gotten much relief from Novell, so I 
started poking around myself and tracked it down to (in the kernel package 
changelog):

* Thu May 23 2013 [email protected]
- Refresh
patches.fixes/mm-vmscan-Block-kswapd-if-it-is-encountering-pages-under-writeback-fix-2.patch
- commit ee1efa5

I just kept removing patches until I got a kernel that didn't exhibit the 
problem.  Removing that patch it works fine...put the patch back on and the 
problem comes back.

The problem is that the patch is on in upstream as well, but upstream kernels 
do not have the problem either.  I suspect it's because of this
commit:

http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/mm/vmstat.c?id=6e543d5780e36ff5ee56c44d7e2e30db3457a7ed

I'm going to try building a vanilla kernel without that and see if the problem 
returns.  Unfortunately, that specific patch looks to be a bugger to backport 
to our SP3 kernel.  :-(

Thanks,

Leland



On 11/1/2013 12:30 PM, Marcy Cortes wrote:
> We're experiencing a problem with kernel 3.0.93 (SP3).   It seems on servers 
> that are memory constrained can go into a loop doing tons of i/o - 8000+ / 
> second.
> It seems to happen with things like WAS deploys.    We may have also been 
> able to recreate use a simple perl array.
>
> Has anyone else seen this?
>
> Yes, we have a high priority SR opened with SUSE.
>
> Oddly we have seen something like this after a kernel update on SP2.  That 
> was solved by changing vm.swappiness to something > 0.    We have put it back 
> to 60 (the default) and still get the problem.
> Upping the memory solves it, but obviously, that's not something we want to 
> be doing, especially in our constrained dev environment.
>
>
> Marcy
>
> ----------------------------------------------------------------------
> For LINUX-390 subscribe / signoff / archive access instructions, send 
> email to [email protected] with the message: INFO LINUX-390 or 
> visit
> http://www.marist.edu/htbin/wlvindex?LINUX-390
> ----------------------------------------------------------------------
> For more information on Linux on System z, visit 
> http://wiki.linuxvm.org/
>


----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions, send email to 
[email protected] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390
----------------------------------------------------------------------
For more information on Linux on System z, visit http://wiki.linuxvm.org/

----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390
----------------------------------------------------------------------
For more information on Linux on System z, visit
http://wiki.linuxvm.org/

Reply via email to