From mboxrd@z Thu Jan 1 00:00:00 1970 From: Chris Mason Subject: Re: How to remount btrfs without compression? Date: Wed, 9 Nov 2011 08:01:34 -0500 Message-ID: <20111109130134.GL4149@shiny> References: <4EB87E01.1040704@parallels.com> <4EB880A0.1030304@gmail.com> <4EB88BC9.7020509@gmail.com> <4EB88D4A.5050908@parallels.com> <20111108150151.GA4954@shiny> <20111108151208.GB4954@shiny> <4EB9D0D9.1000009@parallels.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Lubos Kolouch , linux-btrfs@vger.kernel.org To: "Fajar A. Nugraha" Return-path: In-Reply-To: List-ID: On Wed, Nov 09, 2011 at 03:04:01PM +0700, Fajar A. Nugraha wrote: > On Wed, Nov 9, 2011 at 2:48 PM, Lubos Kolouch wrote: > > Sorry for possibly OT question - when I have historical btrfs system > > mounted with zlib compression, > > > > can I remount it with lzo ? > > yes > > > What will happen? Will the COW be broken > > and the files taking duplicate space? Or will the Universe explode and > > be replaced with something even more bizzare? > > New written block/extents will use lzo compression (if it's > compressible, or if it's mounted with compress-force). Old, unmodified > block/extents will remain unchanged, using zlib or uncompressed. This is correct, your old stuff will be zlib and your new stuff lzo. The compression type is recorded on a per-extent basis and the global switch just decides what type of compression to use for new writes. -chris