From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from fieldses.org ([174.143.236.118]:57698 "EHLO fieldses.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753948Ab1AXSFZ (ORCPT ); Mon, 24 Jan 2011 13:05:25 -0500 Date: Mon, 24 Jan 2011 13:05:24 -0500 To: Max Veprinsky Cc: linux-nfs@vger.kernel.org Subject: Re: releasing space after rapid grwoth Message-ID: <20110124180524.GA15056@fieldses.org> References: Content-Type: text/plain; charset=us-ascii In-Reply-To: From: "J. Bruce Fields" Sender: linux-nfs-owner@vger.kernel.org List-ID: MIME-Version: 1.0 On Fri, Jan 21, 2011 at 05:06:45PM -0600, Max Veprinsky wrote: > Hello all, > > I'm trying to reclaim disk space on stable storage which was consumed > by a crashed apache process at a rate of ~100GB/hour. This issue has > happened a few times before and normally resolved by discovering a > .nfsXXXXX file, stopping the process on the the application server > that was holding the file. Sometimes the used filespace is recovered > quickly without manual intervention, other times I had to delete the > .nfsXXXX file manually. > > Last time this happened 3 TB of disk space was consumed and no > .nfsXXXX file of that size was discovered, nor. I dismounted all of > the volumes from the clients 1/2 of the space was recovered. > > showmount shows clients connected but there are no network nfs > connections open to any client. And all the clients *have* unmounted, right? > I have not restarted nfs server > service fearing that the used space would not be recovered. > > Any suggestion greatly appreciated More details about the setup? Are clients and server all Linux, and what kernel versions? --b.