Date: Sun, 11 Oct 2009 12:00:36 -0600 From: Josh Carter <josh@multipart-mixed.com> To: freebsd-fs@freebsd.org Subject: Re: ZFS vs. df -h completely different size of filesystem [solved] Message-ID: <DADFFD2B-469D-47B4-A57D-A89F6CB59B14@multipart-mixed.com> In-Reply-To: <4AD20B41.3070405@quip.cz> References: <4AD1616C.8060504@quip.cz> <4AD20B41.3070405@quip.cz>
next in thread | previous in thread | raw e-mail | index | archive | help
Miroslav, > But space remains occupied by data of snapshots! Thats why df showed > just 24G size and not 360G. Df knows nothing about snapshots. Some additional factors to keep in mind when looking at ZFS and used/ available disk space: - Snapshots (as you discovered). - Compression: when compression is turned on (as you have), you can't know exactly how much more data will fit into the filesystem because it depends on how well the data compresses. - Sparse files: "ls -h" will show you how large a file says it is; "du -h" and "zfs list" should show how much space is actually used. These will disagree on sparse files. - Space reserved for copy-on-write: "zpool list" and "zfs list" will differ on available space because ZFS reserves some amount of slop space; truly running out of blocks is disastrous in a COW system. In general, learn to trust "zfs list" because the traditional tools don't know the full story. Best regards, Josh
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?DADFFD2B-469D-47B4-A57D-A89F6CB59B14>