Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 23 Oct 2008 08:52:08 -0700
From:      Jeremy Chadwick <koitsu@FreeBSD.org>
To:        Eduardo Meyer <dudu.meyer@gmail.com>
Cc:        stable@freebsd.org
Subject:   Re: gstat information on the CLI
Message-ID:  <20081023155208.GA90330@icarus.home.lan>
In-Reply-To: <d3ea75b30810230811h66ae957ej2e0afcb31fac8519@mail.gmail.com>
References:  <d3ea75b30810230720uef21cfft17b5c63507022482@mail.gmail.com> <20081023145054.GA88957@icarus.home.lan> <d3ea75b30810230811h66ae957ej2e0afcb31fac8519@mail.gmail.com>

next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, Oct 23, 2008 at 01:11:55PM -0200, Eduardo Meyer wrote:
> On Thu, Oct 23, 2008 at 12:50 PM, Jeremy Chadwick <koitsu@freebsd.org> wrote:
> > On Thu, Oct 23, 2008 at 12:20:45PM -0200, Eduardo Meyer wrote:
> >> Hello,
> >>
> >> Its me again bothering you with basic things I cant accomplish.
> >>
> >> I am planning on monitoring my disks, and the most important
> >> information for me is thorughput and lengh queue of operations. I can
> >> get the first information with iostat -w1 and sorta, which is perfect
> >> for scripting.
> >>
> >> However, I also need the L(q) information which FreeBSD gives me with
> >> gstat. However, this curses interface wont allow me to use grep+awk to
> >> get the information I need for the device (slices and disks, but not
> >> labels) I need.
> >
> > Can you tell me what the L(q) field actually represents in gstat?
> >
> > The BUGS section of the gstat man page should indirectly answer your
> > other question (re: non-curses).
> 
> Yes, I have read that. I am looking the source code for gstat. Its
> simple, small and clear. I guess can be asily modified to have what I
> want :)
> 
> >> So I ask, how can I get this information other than gstat? Or, can
> >> gstat work in non-interactive mode?
> >
> > iostat -x should provide what you're looking for.  And remember, the
> > first sample data shown in iostat should be generally discarded.
> 
> I need the queue lengh of pending disk operations. What L(q) shows is
> the lengh queue, the queued number of pending operations (I believe).

It isn't documented, so I really have no idea what it means, hence my
question.  I'm curious why you're interested in that number; why does it
matter?

iostat -x provides the same kind of value, and you won't have to modify
any code to get what you need:

           wait    transactions queue length

-- 
| Jeremy Chadwick                                jdc at parodius.com |
| Parodius Networking                       http://www.parodius.com/ |
| UNIX Systems Administrator                  Mountain View, CA, USA |
| Making life hard for others since 1977.              PGP: 4BD6C0CB |




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20081023155208.GA90330>