Date: Sat, 12 May 2012 03:47:12 +1000 (EST) From: Bruce Evans <brde@optusnet.com.au> To: Ulrich Spoerlein <uqs@freebsd.org> Cc: svn-src-head@freebsd.org, svn-src-all@freebsd.org, src-committers@freebsd.org Subject: Re: svn commit: r235265 - head/share/man/man7 Message-ID: <20120512031923.N2476@besplex.bde.org> In-Reply-To: <201205111013.q4BADY9g076888@svn.freebsd.org> References: <201205111013.q4BADY9g076888@svn.freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, 11 May 2012, Ulrich Spoerlein wrote: > Log: > Stop lying about default UFS blocksizes. > > This catches up with the year-old change to default blocksizes. Also > reduce the variants of spelling gigabyte from 3 down to 2 (GB and GiB). Why not reduce it to 1 (no GakaBytes, aka GiB)? > Modified: head/share/man/man7/tuning.7 > ============================================================================== > --- head/share/man/man7/tuning.7 Fri May 11 09:46:17 2012 (r235264) > +++ head/share/man/man7/tuning.7 Fri May 11 10:13:34 2012 (r235265) > @@ -148,15 +148,15 @@ If you do not use ports all that much an > system source > .Pq Pa /usr/src > on the machine, you can get away with > -a 1 gigabyte > +a 1 GB Since we are not disk manufacturers, GB here means 2**30. I like not spelling it GakaByte of course. Similarly in most places. > @@ -218,22 +218,22 @@ and > .Em cylinders/group . > .Pp > .Fx > -performs best when using 8K or 16K file system block sizes. > -The default file system block size is 16K, > +performs best when using 16K or 32K file system block sizes. > +The default file system block size is 32K, K here means 2**10. I like not spelling it KakaByte of course. > @@ -256,13 +256,13 @@ Do not use this option > unless you are actually storing large files on the partition, because if you > overcompensate you can wind up with a file system that has lots of free > space remaining but cannot accommodate any more files. > -Using 32768, 65536, or 262144 bytes/i-node is recommended. > +Using 65536, 131072, or 262144 bytes/i-node is recommended. Most places give block sizes in K. This is verbose. > You can go higher but > it will have only incremental effects on > .Xr fsck 8 > recovery times. > For example, > -.Dq Li "newfs -i 32768 ..." . > +.Dq Li "newfs -i 65536 ..." . newfs now uses expand_number() for most numeric args. expand_number() is horrible, starting with its name (it doesn't expand numbers, but parses string representations of numbers), but it supports sizes in K and its bugs don't include supporting sizes in GakaBytes, so the verboseness in the above is now unnecessary. > .Pp > .Xr tunefs 8 > may be used to further tune a file system. > @@ -525,8 +525,8 @@ sysctl governs VFS read-ahead and is exp > to pre-read if the heuristics algorithm decides that the reads are > issued sequentially. > It is used by the UFS, ext2fs and msdosfs file systems. > -With the default UFS block size of 16 KiB, a setting of 32 will allow > -speculatively reading up to 512 KiB. > +With the default UFS block size of 32 KiB, a setting of 64 will allow > +speculatively reading up to 2 MiB. This used and uses KakaBytes, but the above uses K (it also omits B). This uses MakaBytes for the first time in this file. Elsewhere, most places use M. The first of these are in the first paragraph, and near them, GB is spelled as G, so the reduction is from >= 4 to >=3, not from 3 to 2. Actually, I can't find any GakaB, so the reduction does seem to be to 2 (G and GB). For M, the spellings are at least M for fs and swap sizes, M and MB for memory sizes, MB for page sizes, and now MakaByte for a block size. > This setting may be increased to get around disk I/O latencies, especially > where these latencies are large such as in virtual machine emulated > environments. Bruce
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20120512031923.N2476>