From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from gabriel-2.zfn.uni-bremen.de ([134.102.20.37]:51692 "EHLO smtp.uni-bremen.de" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1751311Ab3FDVMH (ORCPT ); Tue, 4 Jun 2013 17:12:07 -0400 Date: Tue, 4 Jun 2013 22:58:02 +0200 From: Robin Kreis To: linux-btrfs@vger.kernel.org Subject: Recovering from btrfsck --init-csum-tree Message-ID: <20130604225802.0feb0367@quip> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Sender: linux-btrfs-owner@vger.kernel.org List-ID: Hello, how can I recover from running btrfsck --init-csum-tree on my 2TB btrfs? Every attempt to read a file results in no checksum being found, followed by a checksum mismatch which leads to the data block being zeroed out (see label zeroit in inode.c). My current fix is simply skipping the whole checksum validation, which enables me to read my files again. I still think at least one of the following should happen: 1. Add a warning that --init-csum-tree is pretty desctructive. 2. Make --init-csum-tree mark all inodes with INODE_NODATASUM. 3. Have some global flag or mount option to disable CRC checks. What do you think?