xfs
[Top] [All Lists]

Re: Questions about pagebuf code

To: Craig Tierney <ctierney@xxxxxxxx>
Subject: Re: Questions about pagebuf code
From: Russell Cattelan <cattelan@xxxxxxx>
Date: Tue, 04 May 2004 15:29:09 -0500
Cc: linux-xfs@xxxxxxxxxxx
In-reply-to: <1083698993.2376.4.camel@xxxxxxxxxxxxxxxxxx>
References: <1083435856.2302.3.camel@xxxxxxxxxxxxxxxxxxxxx> <20040501194709.A23768@xxxxxxxxxxxxx> <1083446482.2302.22.camel@xxxxxxxxxxxxxxxxxxxxx> <1083614597.24397.16.camel@xxxxxxxxxxxxxxxxxxxxxx> <1083698993.2376.4.camel@xxxxxxxxxxxxxxxxxx>
Sender: linux-xfs-bounce@xxxxxxxxxxx
On Tue, 2004-05-04 at 13:29 -0600, Craig Tierney wrote:
> On Mon, 2004-05-03 at 14:03, Russell Cattelan wrote:
> > Look at this bug and see if matches you're situation.
> > http://oss.sgi.com/bugzilla/show_bug.cgi?id=198
> 
> 
> I ran doublewrite on my server.  I bumped up the threads
> to 20 and was easy able to generate a problem.  
> 
> I wrote a script that submitted 1 doublewrite process
> to each of 20 nodes, to more mimic what I am doing.  I ran
> these for 6 hours and did not have a problem.  It could
> be that my jobs are putting more of a load on the system,
> or the problem isn't the same.
The test was designed to mimic the out of order writes that 
nfs often generates.

So 20 nfs clients banging on the same file on a server would 
probably end up corrupting it.


> However, since the doublewrite test passes on a uni-processor
> kernel, I am rebuilding my kernel and will rerun my test
> to see if it works or not.  
> 
> The codes might not be the same, but if it is a race condition
> there still might be something similar there.
The problem still exists I just don't have a clue at to where
the race might be.

> 
> Craig
> 
> 
-- 
Russell Cattelan <cattelan@xxxxxxx>

Attachment: signature.asc
Description: This is a digitally signed message part

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