Return-Path: Received: from fieldses.org ([173.255.197.46]:39066 "EHLO fieldses.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753138AbcDNMrb (ORCPT ); Thu, 14 Apr 2016 08:47:31 -0400 Date: Thu, 14 Apr 2016 08:47:28 -0400 From: "J. Bruce Fields" To: YuGiOhJCJ Mailing-List Cc: linux-nfs@vger.kernel.org Subject: Re: Sometimes all the applications using sqlite do not work correctly Message-ID: <20160414124728.GA12667@fieldses.org> References: <20160407010927.bf07f12fa3f62b1586dd11ad@laposte.net> <20160407181422.GB30676@fieldses.org> <20160408004839.247958f8c43329122b48d75f@laposte.net> <20160411210939.447af7a06255f9ec8b491cc2@laposte.net> <20160413174439.GA2808@fieldses.org> <20160414061417.1cbbe93f9f070d57d1133f25@laposte.net> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 In-Reply-To: <20160414061417.1cbbe93f9f070d57d1133f25@laposte.net> Sender: linux-nfs-owner@vger.kernel.org List-ID: On Thu, Apr 14, 2016 at 06:14:17AM +0200, YuGiOhJCJ Mailing-List wrote: > On Wed, 13 Apr 2016 13:44:39 -0400 > bfields@fieldses.org (J. Bruce Fields) wrote: > > > On Mon, Apr 11, 2016 at 09:09:39PM +0200, YuGiOhJCJ Mailing-List wrote: > > > On Fri, 8 Apr 2016 00:48:39 +0200 > > > YuGiOhJCJ Mailing-List wrote: > > > > > > > > So when the problem happens, restarting the server fixes the problem? > > > > > > > > Indeed, that's it. > > > > > > > > > It's probably the kernel version that's most important. So it might be > > > > > interesting to know what your kernel version was before and what it is > > > > > now. > > > > > > > > My Linux kernel was 3.18.20 before and is still 3.18.20 now. > > > > I have not changed of kernel since the problem happens. > > > > > > > > > Also: what's your NFS server? Was it upgraded at the same time, or was > > > > > only your client upgraded? > > > > > > > > I have not upgraded the server and not upgraded the client. > > > > They are still the same: nfs-utils-1.2.8. > > > > > > It happened again today. > > > > > > Output on mumble: > > > $ mumble > > > CELT bitstream 8000000b from libcelt0.so.0.7.0 > > > CELT bitstream 80000010 from libcelt0.so.0.11.0 > > > Locale is "en_US" (System: "en_US") > > > TextToSpeech: Compiled without support for speech-dispatcher > > > SQL Query failed "CREATE TABLE IF NOT EXISTS `servers` (`id` INTEGER PRIMARY KEY AUTOINCREMENT, `name` TEXT, `hostname` TEXT, `port` INTEGER DEFAULT 64738, `username` TEXT, `password` TEXT)" > > ... > > > ALSAAudioInput: Releasing ALSA Mic. > > > ALSAAudioOutput: Destroyed > > > SQL Query failed "PRAGMA journal_mode = DELETE" > > > -1 "database is locked Unable to execute statement" > > > SQL Query failed "VACUUM" > > > -1 "database is locked Unable to fetch row" > > > > > > Output on xmoto: > > > $ xmoto > > > fatal exception : xmDb: database is locked > > > > > > No Output on firefox but I am unable to show bookmarks which means that there is a problem to access the sqlite database. > > > > > > I have just restarted the NFS server like that: > > > $ bash /etc/rc.d/rc.nfsd restart > > > > > > And the client works now perfectly again with applications using sqlite. > > > > > > So I guess that there is a problem with the NFS server (not the client). > > > Is there something I could check like processes running on the server when the problem happens? > > > What information could be useful to understand what happens when I get the problem? > > > > Are there any warnings in the client or server's logs? > > > > --b. > > I am not sure where all these logs are located. The kernel logs are the most interesting. You can get them from "dmesg", or they usually get dumped some place like /var/log/messages depending on distro. --b. > I found the "/var/log/nfsd/" directory on the server but is is empty. > The script running the nfs server is here: > http://ftp.slackware.com/pub/slackware/slackware-14.1/source/n/nfs-utils/rc.nfsd > > I see no options used for logs. > > Reading the man of nfsd, I see these two options: > $ man nfsd > [...] > -d or --debug > enable logging of debugging messages > [...] > -s or --syslog > By default, rpc.nfsd logs error messages (and debug messages, if > enabled) to stderr. This option makes rpc.nfsd log these mes‐ > sages to syslog instead. Note that errors encountered during > option processing will still be logged to stderr regardless of > this option. > [...] > > I guess I should replace in the script this line: > /usr/sbin/rpc.nfsd 8 > by this line: > /usr/sbin/rpc.nfsd -ds 8 > in order to see logs from the server. > > Is it correct? > > About the client logs, I have no idea how to activate them. > > Any idea? > > I saw that my server root partition was full (usage: 100%). > I removed some useless files in the /tmp directory and now I am not full anymore (usage: 98%). > Does NFS need a write access on the server in order to work correctly? > In this case, that could be a reason of my problem.