Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 22 Jul 2004 11:07:54 -0600
From:      Scott Long <scottl@freebsd.org>
To:        Nate Lawson <nate@root.org>
Cc:        cvs-src@freebsd.org
Subject:   Re: cvs commit: src/sys/kern kern_shutdown.c
Message-ID:  <40FFF46A.2080703@freebsd.org>
In-Reply-To: <40FFEB86.2050209@root.org>
References:  <200407212045.i6LKjHvX090599@palm.tree.com> <40FEE569.2010209@elischer.org> <40FEE6CA.3090005@samsco.org> <20040722092441.GH3001@cirb503493.alcatel.com.au> <40FFEB86.2050209@root.org>

next in thread | previous in thread | raw e-mail | index | archive | help
Nate Lawson wrote:

> Peter Jeremy wrote:
> 
>> On Wed, 2004-Jul-21 15:57:30 -0600, Scott Long wrote:
>>
>>> Implementing a journalling filesystem would be a much more beneficial
>>> use of time here.
>>
>>
>> You still wind up with unwritten data in RAM, just less of it.
>>
>> How much effort would be required to add journalling to UFS or UFS2?
>> How big a gain does journalling give you over soft-updates?
> 
> 
> Kirk pointed out something to me the other day which many people don't 
> think about.  None of the journaling systems has had its recovery mode 
> fully tested, especially on very large systems (dozen TB).  It turns out 
> that memory pressure from per-allocation unit state is a big problem 
> when you are trying to recover a huge volume.
> 
> Just because it says "journaling" doesn't make it good.
> 

You are very correct that there are issues like this, and that's why I 
said that it would take a while to chase out the bugs and make it 
production quality.  However, given the enterprise nature of Sun, I'd
say it's a bit of a stretch to think that they haven't tested their
f/s on multi-terabyte arrays.  Even Apple advertises multi-terabyte
storage with their XServe, so I'd be surprised if they hadn't done at
least some testing there.

Scott



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?40FFF46A.2080703>