From mboxrd@z Thu Jan 1 00:00:00 1970 From: Marian Csontos Date: Wed, 12 Feb 2014 10:06:34 +0100 Subject: dmeventd generates segfault when creating snapshots In-Reply-To: <52F9FC6D.6090708@redhat.com> References: <20140210203415.7529.DDAE3F33@valinux.co.jp> <52F9FC6D.6090708@redhat.com> Message-ID: <52FB399A.9090406@redhat.com> List-Id: To: lvm-devel@redhat.com MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit On 02/11/2014 11:33 AM, Zdenek Kabelac wrote: > Dne 10.2.2014 12:34, Chiyohiko Morita napsal(a): >> >> (I'm not on this list; please Cc me) >> >> With a high load, the dmeventd daemon sometimes generates >> a segmentation violation when createing a lvm-snapshot. >> >> In _register_for_event(), it may take 10 or more seconds between >> calling _register_for_timeout() and _create_thread(). >> (Please refer to the following _register_for_timeout().) >> >> If the _timeout_thread thread calls pthread_kill() before >> _create_thread() is called, dmeventd will generate segfault. >> This is because the value of thread->thread is still NULL. >> (Please refer to the following _timeout_thread().) >> >> A related source code and coredump are shown below. >> > > Could you please open regular lvm2 bugzilla > (at http://bugzilla.redhat.com/ - use Fedora Rawhide in case you don't > have match for your distro) > > Attach all your information provided here in your email. > > Add info from 'vgs, lvs -a, pvs' and your in-use lvm.conf > version of your lvm2 package, > how many volumes are active. IMO this is closely related to Bug 889651 - segfault in dmeventd in low memory situations. https://bugzilla.redhat.com/show_bug.cgi?id=889651 -- Marian > > Thank you > > Zdenek > > -- > lvm-devel mailing list > lvm-devel at redhat.com > https://www.redhat.com/mailman/listinfo/lvm-devel >