xfs
[Top] [All Lists]

Re: filesystem shrinks after using xfs_repair

To: Michael Monnerie <michael.monnerie@xxxxxxxxxxxxxxxxxxx>
Subject: Re: filesystem shrinks after using xfs_repair
From: Christoph Hellwig <hch@xxxxxxxxxxxxx>
Date: Fri, 30 Jul 2010 06:29:43 -0400
Cc: xfs@xxxxxxxxxxx, Christoph Hellwig <hch@xxxxxxxxxxxxx>
In-reply-to: <201007301223.12134@xxxxxx>
References: <20100726034545.GE655@dastard> <20100730091554.2c9ca1ec@xxxxxxxxxxxxxx> <20100730075746.GA20444@xxxxxxxxxxxxx> <201007301223.12134@xxxxxx>
User-agent: Mutt/1.5.20 (2009-08-17)
On Fri, Jul 30, 2010 at 12:23:08PM +0200, Michael Monnerie wrote:
> On Freitag, 30. Juli 2010 Christoph Hellwig wrote:
> > Recent enough kernel work fine with filesystems that inode64 was used
> > on even if it's not specified anymore.
>  
> Really? Since when exactly? That would be a nice feature. If we can 
> define it clearly, I could put that on the FAQ.

Linux 2.6.35 will be the first kernel with the bugfixes for this to
work.

> But how does it truncate the numbers >int32 and avoid collisions?

It doesn't.  Existing inodes won't nessecarily fit into 32-bits, but
no new inodes above it will be allocated.

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