xfs
[Top] [All Lists]

Re: xfsdump stuck in io_schedule

To: Zlatko Calusic <zlatko.calusic@xxxxxxxx>
Subject: Re: xfsdump stuck in io_schedule
From: Chris Wedgwood <cw@xxxxxxxx>
Date: Fri, 15 Nov 2002 03:21:34 -0800
Cc: linux-xfs@xxxxxxxxxxx
In-reply-to: <dnfzu3yw8u.fsf@magla.zg.iskon.hr>
References: <dnfzu3yw8u.fsf@magla.zg.iskon.hr>
Sender: linux-xfs-bounce@xxxxxxxxxxx
User-agent: Mutt/1.4i
On Fri, Nov 15, 2002 at 11:15:29AM +0100, Zlatko Calusic wrote:

> xfsdump: page allocation failure. order:0, mode:0x0
> xfsdump: page allocation failure. order:0, mode:0x0
> ...

> % ps axo pid,command,state,wchan | grep xfsdump
>  5442 xfsdump D io_schedule

> The process is, of course, unkillable.  Dump was started to a file
> on a different disk.  Same problem happens on two different
> machines, where one is UP, second is SMP.  Kernel is (virgin) 2.5.47
> on both.  If you need any further info, please don't hesistate to
> ask.

I too had (have?) this problem.

I tried to reproduce it with CVS of late (last three days) and so far
haven't been able to.  I'm now running kdb kernels though in the hope
to get something useful when it does happen.

I'd love to know *why* I can't make this happen anymore (or why it's
harder to hit) as I didn't see anything relevant fixed as best as I
can tell.

Does it still happen for you with a kernel compile from CVS?



  --cw


<Prev in Thread] Current Thread [Next in Thread>