From: "Fredrik Lindgren" Subject: Understanding iostat-ms output Date: Fri, 18 Jan 2008 10:37:55 +0100 Message-ID: <2a29d0a3c209d9488b0a3e8263c53e38@swip.net> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" To: "linux-nfs@vger.kernel.org" Return-path: Received: from mailfe11.swip.net ([212.247.155.65]:40013 "EHLO swip.net" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1751017AbYARKiA convert rfc822-to-8bit (ORCPT ); Fri, 18 Jan 2008 05:38:00 -0500 Received: from [192.71.219.1] (account fli-FpffG6+3qsA@public.gmane.org) by mailbe03.swip.net (CommuniGate Pro IMAP 5.1.13) with XMIT id 90184203 for linux-nfs@vger.kernel.org; Fri, 18 Jan 2008 10:37:56 +0100 Sender: linux-nfs-owner@vger.kernel.org List-ID: Hello, I have been messing around with the excellent "iostat-ms" tool in an effort to understand what is going on with our machines that are showing lots of IOwait. Below are two 15 second samples, the first one is taken during peak and the second one during the night. The obvious difference is in the "exe" field. Please excuse my ignorance, but what does this field mean? Where is this additional time spent? On the filer or in the NFS stack on the host? stbnas1:/foo mounted on /misc/foo: op/s rpc bklog 2364.60 0.96 read: ops/s Kb/s Kb/op retrans avg RTT (ms) avg exe (ms) 43.267 659.624 15.246 0 (0.0%) 17.066 293.880 write: ops/s Kb/s Kb/op retrans avg RTT (ms) avg exe (ms) 16.133 60.314 3.738 0 (0.0%) 3.868 255.074 stbnas1:/foo mounted on /misc/foo: op/s rpc bklog 1446.07 0.26 read: ops/s Kb/s Kb/op retrans avg RTT (ms) avg exe (ms) 105.200 189.155 1.798 0 (0.0%) 8.264 8.702 write: ops/s Kb/s Kb/op retrans avg RTT (ms) avg exe (ms) 105.067 297.774 2.834 0 (0.0%) 1.048 1.310 Regards, Fredrik Lindgren