Date: Wed, 4 Apr 2018 12:46:33 +0500 From: "Eugene M. Zheganin" <eugene@zhegan.in> To: "Patrick M. Hausen" <hausen@punkt.de> Cc: freebsd-stable@freebsd.org Subject: Re: another question about zfs compression numbers Message-ID: <60635355-c0f3-9c3e-2375-7637bb073d87@zhegan.in> In-Reply-To: <8A5C86CA-C959-4DFF-9168-DD94CF46AC91@punkt.de> References: <52b1c557-bdb5-3b9f-1ce1-32f698ae982c@zhegan.in> <8A5C86CA-C959-4DFF-9168-DD94CF46AC91@punkt.de>
next in thread | previous in thread | raw e-mail | index | archive | help
Hi, On 04.04.2018 12:35, Patrick M. Hausen wrote: > Hi all, > >> Am 04.04.2018 um 09:21 schrieb Eugene M. Zheganin <eugene@zhegan.in>: >> I'm just trying to understand these numbers: >> >> file size is 232G, it's actual size on the lz4-compressed dataset is 18G, so then why is the compressratio only 1.86x ? And why logicalused is 34.2G ? On one hand, 34.2G exactlyfits to the 1.86x compresstaio, but still I don't get it. dataset is on raidz, 3 spans across 5 disk vdevs, with total of 15 disks if it matters: > A sparse file, possibly? The ZFS numbers refer to blocks. "Skipping" zeroes at the > VFS layer is not taken into account as fas as I know. Seriously, how should it? > If I'm not mistaken, ZFS will never get to see these empty blocks. > Looks so, thanks. Although it's a mysql tablespace file. But yeah, in hex viewer looks like it's filled with zeroes in many places. Eugene.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?60635355-c0f3-9c3e-2375-7637bb073d87>