Date: Fri, 11 Dec 2009 09:59:40 +0100 From: Jeremie Le Hen <jeremie@le-hen.org> To: Andriy Gapon <avg@icyb.net.ua> Cc: freebsd-current@FreeBSD.org, Jeremie Le Hen <jeremie@le-hen.org> Subject: Re: Panic while doing zfs rename Message-ID: <20091211085940.GA7114@felucia.tataz.chchile.org> In-Reply-To: <4B20EAA3.5010009@icyb.net.ua> References: <20091210072147.GA4963@felucia.tataz.chchile.org> <4B20EAA3.5010009@icyb.net.ua>
next in thread | previous in thread | raw e-mail | index | archive | help
Hi Andriy, On Thu, Dec 10, 2009 at 02:33:39PM +0200, Andriy Gapon wrote: > on 10/12/2009 09:21 Jeremie Le Hen said the following: > ... > > % zfs_statfs(866b2a10,866b2a70,1d8,cb41c844,865a3a10,...) at zfs_statfs+0x39 > > % __vfs_statfs(866b2a10,cb41c844,0,0,0,...) at __vfs_statfs+0x1f > > % nullfs_statfs(865a3a10,865a3a70,806bd68b,865a3a70,865a3a10,...) at nullfs_statfs+0x46 > > % __vfs_statfs(865a3a10,865a3a70,1d8,a5889340,cb41cb78,...) at __vfs_statfs+0x1f > > And you also seem to have nullfs in the picture. > Did you by a chance renamed a zfs filesystem under nullfs? Unfortunately I can't tell you for sure. I'd say yes according to the backtrace... ISTR I did some tests to null-mount the _removed_ filesystem (former data/repos). I think the following rename from data/crepos to data/repos triggered something in the nullfs layer that led to this panic. Now I'm thinking about it, I remember doing a forced-destroy (-f) of the original data/repos. Would this mean the nullfs filesystem will panic if the underlying filesystem disappear? Unfortunately I can't afford crashing this server once more, to verify this. Regards, -- Jeremie Le Hen < jeremie at le-hen dot org >< ttz at chchile dot org >
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20091211085940.GA7114>