Date: Sat, 16 Apr 2005 21:46:51 +0200 From: "Poul-Henning Kamp" <phk@phk.freebsd.dk> To: Eric Anderson <anderson@centtech.com> Cc: FreeBSD Current <freebsd-current@freebsd.org> Subject: Re: gstat shows > 100% busy Message-ID: <3703.1113680811@critter.freebsd.dk> In-Reply-To: Your message of "Sat, 16 Apr 2005 14:37:25 CDT." <42616975.9060303@centtech.com>
next in thread | previous in thread | raw e-mail | index | archive | help
In message <42616975.9060303@centtech.com>, Eric Anderson writes: >Is gstat supposed to show > 100% sometimes? What does that mean, >or is it a bug? > >dT: 0.501 flag_I 500000us sizeof 240 i -1 > L(q) ops/s r/s kBps ms/r w/s kBps ms/w %busy Name > 2 260 146 14912 10.7 114 14565 2.8 148.1| ad0 > 0 0 0 0 0.0 0 0 0.0 0.0| ad0s1 The reason gstat shows >100% busy is that there are some outstanding requests. (the 2 in the left hand column). I tried to make the statistics collection as cheap as possible, and as a side effect some of the columns can be somewhat misleading. The length of the queue "L(q)" can be plain wrong due to a race in updating the counters and %busy can go over 100% while there are outstanding requests. The sysctl kern.geom.collectstats can be used to tune some aspects of the statistics collection, but the %busy issue is just something you have to live with. The reason why I don't want to spend cpu time on the %busy field is that it is useless as a performance indication for all modern disks and most ancient ones as well. The "ms/r" and "ms/w" give you the time it takes to send a transaction through (in milliseconds, for read and write respectively) and those are the numbers you should monitor. -- Poul-Henning Kamp | UNIX since Zilog Zeus 3.20 phk@FreeBSD.ORG | TCP/IP since RFC 956 FreeBSD committer | BSD since 4.3-tahoe Never attribute to malice what can adequately be explained by incompetence.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3703.1113680811>