From mboxrd@z Thu Jan 1 00:00:00 1970 From: Arnd Hannemann Subject: Re: 3.4.0-rc6: WARNING: at fs/btrfs/super.c:219 __btrfs_abort_transaction+0xae/0xc0 [btrfs]() Date: Mon, 14 May 2012 18:16:33 +0200 Message-ID: <4FB12FE1.2070909@arndnet.de> References: <4FB10636.2010003@arndnet.de> <20120514150509.GJ25501@twin.jikos.cz> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Cc: linux-btrfs@vger.kernel.org, linux-kernel@vger.kernel.org To: dave@jikos.cz Return-path: In-Reply-To: <20120514150509.GJ25501@twin.jikos.cz> List-ID: Hello David, Am 14.05.2012 17:05, schrieb David Sterba: > On Mon, May 14, 2012 at 03:18:46PM +0200, Arnd Hannemann wrote: >> I just got the following warning on a compressed btrfs filesystem, while writing on it >> until no remaining space was available. >> Looks like a bit verbose "disk full" message, is it expected behavior? > > Shouldn't be the case for regular ENOSPC (ie. when the space is reserved > in advance and with the possibility to fail early). If the transaction > is aborted, it means it's too late and there was some logic error. > > Can you please describe more the fs and how did you fill it? Like single > process or multiple, dd or whatever. the particular filesystem is used for chroots (/opt/chroots). It got full while installing texlive-extra on a debian squeeze (i386) chroot. It was probably a single process, but it involed a lot of small files. Best regards Arnd