Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-2.5 required=3.0 tests=HEADER_FROM_DIFFERENT_DOMAINS, MAILING_LIST_MULTI,SPF_PASS,USER_AGENT_MUTT autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 575FCC10F13 for ; Tue, 16 Apr 2019 20:46:06 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 206072075B for ; Tue, 16 Apr 2019 20:46:06 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1728000AbfDPUqF (ORCPT ); Tue, 16 Apr 2019 16:46:05 -0400 Received: from fieldses.org ([173.255.197.46]:39552 "EHLO fieldses.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1727136AbfDPUqF (ORCPT ); Tue, 16 Apr 2019 16:46:05 -0400 Received: by fieldses.org (Postfix, from userid 2815) id 218493EB; Tue, 16 Apr 2019 16:46:05 -0400 (EDT) Date: Tue, 16 Apr 2019 16:46:05 -0400 To: Bruno Santos Cc: Linux NFS Mailing List Subject: Re: Fwd: nfs v4.2 leaking file descriptors Message-ID: <20190416204605.GB6662@fieldses.org> References: MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.5.21 (2010-09-15) From: bfields@fieldses.org (J. Bruce Fields) Sender: linux-nfs-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-nfs@vger.kernel.org On Mon, Apr 15, 2019 at 06:00:56PM +0100, Bruno Santos wrote: > We have a debian stretch HPC cluster(#1 SMP Debian 4.9.130-2 > (2018-10-27)). One of the machines mounts a couple of drives from a > Dell compellent system and shares it across a 10GB network to 4 > different machines. > > We had the nfs server crashing a few weeks ago because the file-max > limit had been reached. At the time we increased the number of file > handles it can handle and been monitoring since. We have noticed that > the number of entries on that machine keeps increasing though and > despite our best efforts we have been unable identify the cause. > Anything I can find related to this is from a well known bug in 2011 > and nothing afterwards. We are assuming this is caused but a leak of > file handles on the nfs side but not sure. > > Does anyone has anyway of figuring out what is causing this? Output > from the file-ne, lsof, etc is below. Off the top of my head, the only idea I have is to try watching grep nfsd4 /proc/slabinfo and see if any of those objects are also leaking. --b. > > Thank you very much for any help you can provide. > > Best regards, > Bruno Santos > > :~# while :;do echo "$(date): $(cat /proc/sys/fs/file-nr)";sleep > 30;done > Mon 15 Apr 17:23:11 BST 2019: 2466176 0 4927726 > Mon 15 Apr 17:23:41 BST 2019: 2466176 0 4927726 > Mon 15 Apr 17:24:11 BST 2019: 2466336 0 4927726 > Mon 15 Apr 17:24:41 BST 2019: 2466240 0 4927726 > Mon 15 Apr 17:25:11 BST 2019: 2466560 0 4927726 > Mon 15 Apr 17:25:41 BST 2019: 2466336 0 4927726 > Mon 15 Apr 17:26:11 BST 2019: 2466400 0 4927726 > Mon 15 Apr 17:26:41 BST 2019: 2466432 0 4927726 > Mon 15 Apr 17:27:11 BST 2019: 2466688 0 4927726 > Mon 15 Apr 17:27:41 BST 2019: 2466624 0 4927726 > Mon 15 Apr 17:28:11 BST 2019: 2466784 0 4927726 > Mon 15 Apr 17:28:41 BST 2019: 2466688 0 4927726 > Mon 15 Apr 17:29:11 BST 2019: 2466816 0 4927726 > Mon 15 Apr 17:29:42 BST 2019: 2466752 0 4927726 > Mon 15 Apr 17:30:12 BST 2019: 2467072 0 4927726 > Mon 15 Apr 17:30:42 BST 2019: 2466880 0 4927726 > > ~# lsof|wc -l > 3428