From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: linux-nfs-owner@vger.kernel.org Received: from bombadil.infradead.org ([198.137.202.9]:53359 "EHLO bombadil.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754694Ab3KGP6h (ORCPT ); Thu, 7 Nov 2013 10:58:37 -0500 Date: Thu, 7 Nov 2013 07:58:37 -0800 From: Christoph Hellwig To: Jeff Layton Cc: steved@redhat.com, linux-nfs@vger.kernel.org Subject: Re: [PATCH] gssd: cut down log chatter in gssd_search_krb5_keytab Message-ID: <20131107155837.GA16658@infradead.org> References: <1383839253-8223-1-git-send-email-jlayton@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: <1383839253-8223-1-git-send-email-jlayton@redhat.com> Sender: linux-nfs-owner@vger.kernel.org List-ID: On Thu, Nov 07, 2013 at 10:47:33AM -0500, Jeff Layton wrote: > Because the kernel now looks for krb5 creds by default when mounting, we > need to always have clients run rpc.gssd to avoid the 15s hang on the > first mount attempt. Not sure if it's the same hangs, but I see about the same length of a hang on my Debian wheezy systems when accessing the first file for a data operation, with rpc.gssd running or not. Funnily enough this also happens when trying to read a non-existand file, but not for things like a simple readdir. Still looking into it, but being a non-interruptible sleep it's fairly annoying. That being said Debian wheezy also only starts rpc.gssd from the init scripts if it sees a krb option in /etc/fstab, which I don't have.