Return-Path: Received: from daytona.panasas.com ([67.152.220.89]:24090 "EHLO daytona.panasas.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757835Ab1EWWbx (ORCPT ); Mon, 23 May 2011 18:31:53 -0400 Message-ID: <4DDAE041.2080707@panasas.com> Date: Tue, 24 May 2011 01:31:29 +0300 From: Boaz Harrosh To: Benny Halevy CC: Trond Myklebust , linux-nfs@vger.kernel.org Subject: Re: [PATCH] SQUASHME: Bugs in new global-device-cache code References: <4DD99F9B.2040406@panasas.com> <1306108720-28762-1-git-send-email-bhalevy@panasas.com> <4DD9E805.2020106@panasas.com> <4DDA64C6.2080909@panasas.com> <4DDAC93A.2020209@panasas.com> <4DDAD8AC.8030206@panasas.com> In-Reply-To: <4DDAD8AC.8030206@panasas.com> Content-Type: text/plain; charset=UTF-8 Sender: linux-nfs-owner@vger.kernel.org List-ID: MIME-Version: 1.0 On 05/24/2011 12:59 AM, Boaz Harrosh wrote: > 2. Something funny I'm still investigating. When I do > a small Io couple of requests the devices get freed > at the end on the release of the layout. (Which is not > layout_returned ever) > But when I do very large IO and lots of concurrent requests > the devices do not get to be released at all, they stay in > cache. But am still investigating OK That was a stupid print problem. The buffer was full and the prints did not show. It's fine, I guess I need to sleep ;-) Boaz