Date: Fri, 28 Mar 2008 16:26:16 +0100 From: Ivan Voras <ivoras@freebsd.org> To: freebsd-stable@freebsd.org Subject: Re: Question about file system checks Message-ID: <fsj2mo$dgc$1@ger.gmane.org> In-Reply-To: <20080328142843.GD28690@dan.emsphone.com> References: <47EBA3AB.40307@infracaninophile.co.uk> <f9ae3129fa235b31251ec97bc12c1e78@localhost> <200803280029.08136.danny@ricin.com> <fshdv1$jbt$1@ger.gmane.org> <20080328142843.GD28690@dan.emsphone.com>
next in thread | previous in thread | raw e-mail | index | archive | help
This is an OpenPGP/MIME signed message (RFC 2440 and 3156)
--------------enig7910809D3E036755CF293A1D
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
Dan Nelson wrote:
> In the last episode (Mar 28), Ivan Voras said:
> Note that you can tweak the SU caching time by adjusting the sysctls
> kern.{meta,dir,file}delay. Take them down to 10 seconds instead of 30
> and you'll lose less files (at the cost of more disk I/O of course).
Yes, but the other file systems were also tested with their respective
defaults.
>> 5. ZFS on FreeBSD was the fastest, in the sense of creating the most f=
iles=20
>> during this benchmark (though speed was not the target for this benchm=
ark=20
>> so this is a low-quality observation), closely followed by JFS and XFS=
=2E
>=20
> ZFS's transaction commit interval is only 5 seconds (see txg_time in
> uts/common/fs/zfs/txg.c); how many more files/second did it create vs
> the others to be able to lose the most files in that window? :)
All were tested within the same time: 50 seconds. Details: the machine
being tested was connected to a "reporter" machine via plain crossover
cable, the reporter had a TCP server and the tested machine had a TCP
client that run a tight loop of IO operations, single threaded, randomly
choosing between creating files and directories, appending to them and
changing (a random amount of data in a random position) them, then
sending to the server a description (log) of each IO operation after it
has been done. These were several Python scripts I wrote.
After 50 seconds I'd pull the plug off the server (meaning it's not
*exactly* 50 seconds but 50 seconds+time of my action, but it doesn't
matter as everything is timestamped), plug it back in, run fsck. The
reporting server would replay the log of IO operations from the first to
the last and generate a report of what should have been on the tested
machine which would then be compared with what was found on the tested
machine (by scripts/automated).
--------------enig7910809D3E036755CF293A1D
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: OpenPGP digital signature
Content-Disposition: attachment; filename="signature.asc"
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
iD8DBQFH7Q4YldnAQVacBcgRAo3TAKCtKLUE5VeftLm1dP5zxD50+ro8/ACgrMIS
03T1I5fsIVcVnfnpL/d5bxo=
=FfHf
-----END PGP SIGNATURE-----
--------------enig7910809D3E036755CF293A1D--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?fsj2mo$dgc$1>