This is the mail archive of the
systemtap@sourceware.org
mailing list for the systemtap project.
Re: New example script: nfstop
- From: David Smith <dsmith at redhat dot com>
- To: "Bryn M. Reeves" <bmr at redhat dot com>
- Cc: Systemtap List <systemtap at sourceware dot org>
- Date: Thu, 05 Apr 2012 14:39:11 -0500
- Subject: Re: New example script: nfstop
- References: <4F6C8F31.3070608@redhat.com> <4F6C932A.1050206@redhat.com>
On 03/23/2012 10:13 AM, Bryn M. Reeves wrote:
> Funnily enough I hacked something similar together last year. It's
> been on my TODO since then to clean it up a bit and use it for a Red
> Hat article.
>
> I started working on a version using aggregates but that got ugly fast
> (as it needed to handle empty and non-empty aggregates explicitly at
> the time).
>
> I think I talked to Frank at the time about ways to make this cleaner
> (iirc all the @ops were wrapped in if(have_xxx_stats) constructs) but
> I haven't been back to look again since.
>
> Here's my original version. It just tracks lookups, reads and writes
> and maintains global and per-client byte counters. It's been
> successfully used in production at a few sites (iirc on RHEL5) for
> tracking down badly behaved NFS clients and observing general server
> performance.
>
> I should probably dust off the improved version and see if I can
> de-ugly it a bit.
I went ahead and checked in this version. I then went through and
converted it to use aggregates, which simplified things a bit.
--
David Smith
dsmith@redhat.com
Red Hat
http://www.redhat.com
256.217.0141 (direct)
256.837.0057 (fax)