From: "Talpey, Thomas" Subject: Re: nfs4 over rdma transport oops Date: Thu, 28 Feb 2008 03:13:58 -0500 Message-ID: References: <4a4634330802272049xe20ef80y21c2c59f55932c70@mail.gmail.com> <4a4634330802272049xe20ef80y21c2c59f55932c70@mail.gmail.com > Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Cc: linux-nfs@vger.kernel.org To: "Shirish Pargaonkar" Return-path: Received: from mx2.netapp.com ([216.240.18.37]:29381 "EHLO mx2.netapp.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754597AbYB1IPH (ORCPT ); Thu, 28 Feb 2008 03:15:07 -0500 In-Reply-To: <4a4634330802272049xe20ef80y21c2c59f55932c70-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org > References: <4a4634330802272049xe20ef80y21c2c59f55932c70-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org> Sender: linux-nfs-owner@vger.kernel.org List-ID: At 11:49 PM 2/27/2008, Shirish Pargaonkar wrote: >The code fix works. I did not apply the workaround you had mentioned earlier >and did not encounter oops for the same operation (cat) on the file on the >mounted filesystem. Great! I'll send it along to Trond then. If you have a chance after running for a while on an RDMA mount point, I'd be interested in the output of Chuck's "mountstats" python script, which can extract some of the rdma operation statistics from the rpc/rdma code. The "hardway" statistic is the one I'm especially interested in, in your NFSv4 case. You can get it at: http://oss.oracle.com/~cel/linux-2.6/2.6.24/mountstats There is also another script there which shows some very interesting timing and other performance data, which I recommend: http://oss.oracle.com/~cel/linux-2.6/2.6.24/nfs-iostat Thanks, Tom.