xfs
[Top] [All Lists]

Re: kernel OOPS for XFS in xfs_iget_core (using NFS+SMP+MD)

To: Jakob Oestergaard <jakob@xxxxxxxxxxxxx>
Subject: Re: kernel OOPS for XFS in xfs_iget_core (using NFS+SMP+MD)
From: Joshua Baker-LePain <jlb17@xxxxxxxx>
Date: Wed, 18 May 2005 16:00:08 -0400 (EDT)
Cc: Chris Wedgwood <cw@xxxxxxxx>, Gregory Brauer <greg@xxxxxxxxxxxxx>, linux-kernel@xxxxxxxxxxxxxxx, linux-xfs@xxxxxxxxxxx
In-reply-to: <20050518195251.GY422@unthought.net>
References: <428511F8.6020303@wildbrain.com> <20050514184711.GA27565@taniwha.stupidest.org> <428B7D7F.9000107@wildbrain.com> <20050518175925.GA22738@taniwha.stupidest.org> <20050518195251.GY422@unthought.net>
Sender: linux-xfs-bounce@xxxxxxxxxxx
On Wed, 18 May 2005 at 9:52pm, Jakob Oestergaard wrote

> On Wed, May 18, 2005 at 10:59:25AM -0700, Chris Wedgwood wrote:
> > On Wed, May 18, 2005 at 10:38:07AM -0700, Gregory Brauer wrote:
> > 
> > > May 18 02:59:47 violet kernel: xfs_iget_core: ambiguous vns: 
> > > vp/0xf53f8ac8, invp/0xe49ccc4c
> > 
> > I'm pretty sure it's NFS that aggravates this --- can anyone recall
> > why?
> 
> Not why no - but there where *major* problems with SMP+NFS+XFS up until
> 2.6.11.
> 
> I run 2.6.11(.8/9) on both SMP (dual athlon) and NUMA (64 bit kernel on
> dual opteron) with NFS and XFS and haven't yet seen any problems (knock
> the wood).
> 
> Seriously, any 2.6 earlier than .11 is *unusable* for file serving over
> NFS (at least with XFS which at the moment is the only FS with
> journalled quota so at least for me that's the only option).

Do you have a test case that would show this up?  I've been testing a 
centos-4 based server with the RH-derived 2.6.9-based kernel tweaked to 
disable 4K stacks and enable XFS and haven't run into any issues yet.  
This includes running the parallel IOR benchmark from 10 clients (and 
getting 200MiB/s throughput on reads).

-- 
Joshua Baker-LePain
Department of Biomedical Engineering
Duke University


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