From: Andreas Dilger Subject: Re: Lustre basic tests script Date: Sun, 21 Jun 2009 13:19:11 +0200 Message-ID: <20090621111911.GB11977@webber.adilger.int> References: <20090522213639.GC13578@webber.adilger.int> Mime-Version: 1.0 Content-Type: text/plain; CHARSET=US-ASCII Content-Transfer-Encoding: 7BIT Cc: Eric Sandeen , linux-ext4@vger.kernel.org, LTP To: Subrata Modak Return-path: Received: from sca-es-mail-2.Sun.COM ([192.18.43.133]:43282 "EHLO sca-es-mail-2.sun.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751000AbZFULTt (ORCPT ); Sun, 21 Jun 2009 07:19:49 -0400 Received: from fe-sfbay-09.sun.com ([192.18.43.129]) by sca-es-mail-2.sun.com (8.13.7+Sun/8.12.9) with ESMTP id n5LBJZcd028153 for ; Sun, 21 Jun 2009 04:19:50 -0700 (PDT) Content-disposition: inline Received: from conversion-daemon.fe-sfbay-09.sun.com by fe-sfbay-09.sun.com (Sun Java(tm) System Messaging Server 7u2-7.02 64bit (built Apr 16 2009)) id <0KLL0020062E9Y00@fe-sfbay-09.sun.com> for linux-ext4@vger.kernel.org; Sun, 21 Jun 2009 04:19:35 -0700 (PDT) In-reply-to: Sender: linux-ext4-owner@vger.kernel.org List-ID: On May 23, 2009 16:33 +0530, Subrata Modak wrote: > On Sat, May 23, 2009 at 3:06 AM, Andreas Dilger wrote: > > > Below are included some of the tests from the Lustre "sanity.sh" script, > > which tests filesystem behaviour in a variety of ways. Some of them are > > very simple and might concievably be removed, but are not necessarily bad. > > I've removed the tests that depend on Lustre-specific behaviour. > > > > It probably isn't useful to use this test script directly, since it > > depends on a number of helper functions, but it is definitely useful > > as a starting point for adding new tests to xfstests. > > > > Unfortunately, a bunch of tests had to be removed since they depend on > > Lustre's fault-injection code, but they would be useful for testing > > generic filesystem behaviour (e.g. ENOMEM or ENOSPC handling). There > > is also a helper program we use "multiop" which is a toolbox program > > that I could supply separately, if you want. > > > How about understanding these and find out if they can be made part of LTP > as well ? Any objection(s) ? Sorry for the late response. No objection to using these tests in LTP. I've separatel posted the "multiop" program to this list so that you can better understand what the tests are doing. Cheers, Andreas -- Andreas Dilger Sr. Staff Engineer, Lustre Group Sun Microsystems of Canada, Inc.