From: "Chuck Lever" Subject: Re: [NFS] NFS performance debugging Date: Fri, 27 Jun 2008 09:44:38 -0400 Message-ID: <76bd70e30806270644j3e67c83and7b1f7fd6cc2f5f5@mail.gmail.com> References: <48624F34.1070108@RedHat.com> Reply-To: chucklever@gmail.com Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Cc: nfs@lists.sourceforge.net, Steve Dickson To: "Krishna Kumar2" Return-path: Received: from neil.brown.name ([220.233.11.133]:58000 "EHLO neil.brown.name" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754917AbYF0Now (ORCPT ); Fri, 27 Jun 2008 09:44:52 -0400 Received: from brown by neil.brown.name with local (Exim 4.63) (envelope-from ) id 1KCEFt-0002Fc-7W for linux-nfs@vger.kernel.org; Fri, 27 Jun 2008 23:44:49 +1000 In-Reply-To: Sender: linux-nfs-owner@vger.kernel.org List-ID: On Fri, Jun 27, 2008 at 3:30 AM, Krishna Kumar2 wrote: > linux-nfs-owner@vger.kernel.org wrote on 06/25/2008 07:29:16 PM: > >> But please do not let this caveat stop you from trying this script and >> mountstats script. We definitely need feedback on how well they do or >> don't work (and as always... patches are welcomed! ;-) ) Once things > harden >> up (via any feedback), I'll added the install code to make these scripts > be >> installed (probably in /usr/sbin) which means they will be ready for >> prime time... > > Hi Steve, > > When I ran this util, it worked the first time: > > [root@localhost ~]# /tmp/nfs-iostat > > localhost:/local mounted on /nfs: > > op/s rpc bklog > 1.27 0.01 > read: ops/s Kb/s Kb/op retrans > avg RTT (ms) avg exe (ms) > 0.619 39.772 64.273 0 (0.0%) > 16.889 16.927 > write: ops/s Kb/s Kb/op retrans > avg RTT (ms) avg exe (ms) > 0.626 39.112 62.523 0 (0.0%) > 26.357 4039.241 > > 8.124.111.71:/mnt mounted on /mnt: > > op/s rpc bklog > 3.00 0.00 > read: ops/s Kb/s Kb/op retrans > avg RTT (ms) avg exe (ms) > 0.000 0.000 0.000 0 (0.0%) > 0.000 0.000 > write: ops/s Kb/s Kb/op retrans > avg RTT (ms) avg exe (ms) > 0.000 0.000 0.000 0 (0.0%) > 0.000 0.000 > > (KK: for some reason, the real NFS mount gives all zero. Also, for > loopback, the read/write > BW is very poor: less than 40 Kb/s). As an aside, Jim Rees pointed out to me that that should be KB/s, not Kb/s. I will post a patch to address that. > But from the next time I ran, and all subsequent runs (and even after a > reboot), I get: > > [root@localhost init.d]# /tmp/nfs-iostat > localhost:/local mounted on /nfs: > > op/s rpc bklog > 0.04 0.00 > read: ops/s Kb/s Kb/op retrans > avg RTT (ms) avg exe (ms) > 0.000 0.000 0.000 0 (0.0%) > 0.000 0.000 > write: ops/s Kb/s Kb/op retrans > avg RTT (ms) avg exe (ms) > 0.000 0.000 0.000 0 (0.0%) > 0.000 0.000 > > 8.124.111.71:/mnt mounted on /mnt: > > op/s rpc bklog > 6.00 0.00 > read: ops/s Kb/s Kb/op retrans > avg RTT (ms) avg exe (ms) > 0.000 0.000 0.000 0 (0.0%) > 0.000 0.000 > write: ops/s Kb/s Kb/op retrans > avg RTT (ms) avg exe (ms) > 0.000 0.000 0.000 0 (0.0%) > 0.000 0.000 > > (KK: every filesystem gives zero for all parameters). > > What could be wrong? At a guess, it's because the client is now caching your file data? Can you post a copy of your mount command line, and your /proc/self/mountstats file after a test run (but before you unmount)? -- Chuck Lever ------------------------------------------------------------------------- Check out the new SourceForge.net Marketplace. It's the best place to buy or sell services for just about anything Open Source. http://sourceforge.net/services/buy/index.php _______________________________________________ NFS maillist - NFS@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/nfs _______________________________________________ Please note that nfs@lists.sourceforge.net is being discontinued. Please subscribe to linux-nfs@vger.kernel.org instead. http://vger.kernel.org/vger-lists.html#linux-nfs