From: "Shankar Anand" Subject: Re: PATCH [1/2] Capture nfsv4 server operations statistics Date: Sat, 20 May 2006 02:07:34 -0600 Message-ID: <446F1B9F.AD69.00CC.0@novell.com> References: <4469C82B.AD69.00CC.0@novell.com> <20060517213638.GE11237@fieldses.org> <446E3040.AD69.00CC.0@novell.com> <20060519162747.GI3721@fieldses.org> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Cc: , Return-path: Received: from [10.3.1.94] (helo=sc8-sf-list2-new.sourceforge.net) by sc8-sf-list2.sourceforge.net with esmtp (Exim 4.30) id 1FhMTR-0001Ol-0L for nfs@lists.sourceforge.net; Sat, 20 May 2006 01:06:09 -0700 Received: from sc8-sf-mx2-b.sourceforge.net ([10.3.1.92] helo=mail.sourceforge.net) by sc8-sf-list2-new.sourceforge.net with esmtp (Exim 4.43) id 1FhMTQ-0002kQ-SW for nfs@lists.sourceforge.net; Sat, 20 May 2006 01:06:08 -0700 Received: from lucius.provo.novell.com ([137.65.81.172]) by mail.sourceforge.net with esmtp (Exim 4.44) id 1FhMTO-0007Dm-Lu for nfs@lists.sourceforge.net; Sat, 20 May 2006 01:06:08 -0700 To: "J. Fields" In-Reply-To: <20060519162747.GI3721@fieldses.org> Sender: nfs-admin@lists.sourceforge.net Errors-To: nfs-admin@lists.sourceforge.net List-Unsubscribe: , List-Id: Discussion of NFS under Linux development, interoperability, and testing. List-Post: List-Help: List-Subscribe: , List-Archive: Hi, > On 5/19/2006 at 9:57:47 pm, > "J. Bruce Fields" wrote: > On Fri, May 19, 2006 at 09:23:20AM -0600, Shankar Anand wrote: >> Verified compilation when nfsv4 is configured out. And also=20 >> made changes so that the nfsv4 operation statistics are captured >> only when nfsv4 is configured. >=20 > Thanks! Looks fine to me. >=20 > One last thing I wonder about--the current /proc/net/rpc/nfsd ends with > lines like: >=20 > proc2 18 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > proc3 22 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > proc4 2 4 172 >=20 > where the two operations counted on the last line are NULL and COMPOUND. > You're appending one more line, to make: >=20 > proc2 18 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > proc3 22 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > proc4 2 4 172 > proc4ops 37 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 >=20 > I like the fact that that doesn't bother the preexisting proc4 line. On > the other hand, as far as we know nobody's ever used the data on that > line, and it's hard to see how they would in the future (the number of > NULL calls to the server isn't very important), so may it would be > simplest just to replace the proc4 line entirely, as was done on the > client side. >=20 > Any strong opinions from anyone else? There are some small issues in doing that. The proc2, proc3 and proc4 lines get written into the /proc file by=20 rpc_proc_show() function in sunrpc library unlike the proc4ops line = that=20 I have written now, which gets written into /proc file at nfs level in=20 nfs_proc_show(). All the procedures that are registered as rpc procedures, get their = statistics=20 recorded in the /proc file in the sunrpc rpc_proc_show() function. If = we=20 have to avoid the original proc4 line, we will have to introduce a special = case=20 to specifically refrain from writing v4 rpc statistics into the /proc = file,=20 which may not be appropriate. In the client, all the procedures are registered as rpc procedures in = nfs4xdr.c. Hence, the proc4 line for the client contains all the v4 procedures. Cheers, Shankar ------------------------------------------------------- Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642 _______________________________________________ NFS maillist - NFS@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/nfs