From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from userp2120.oracle.com ([156.151.31.85]:42840 "EHLO userp2120.oracle.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752676AbeBZIZW (ORCPT ); Mon, 26 Feb 2018 03:25:22 -0500 Subject: Re: Help with leaf parent key incorrect To: Paul Jones , "linux-btrfs@vger.kernel.org" References: From: Anand Jain Message-ID: <2fd186fe-34b3-c2dd-ad36-a043368104eb@oracle.com> Date: Mon, 26 Feb 2018 16:26:43 +0800 MIME-Version: 1.0 In-Reply-To: Content-Type: text/plain; charset=utf-8; format=flowed Sender: linux-btrfs-owner@vger.kernel.org List-ID: > There is one io error in the log below, Apparently, that's not a real EIO. We need to fix it. But can't be the root cause we are looking for here. ---- > Feb 24 22:41:59 home kernel: BTRFS: error (device dm-6) in btrfs_run_delayed_refs:3076: errno=-5 IO failure > Feb 24 22:41:59 home kernel: BTRFS info (device dm-6): forced readonly static int run_delayed_extent_op(struct btrfs_trans_handle *trans, struct btrfs_fs_info *fs_info, struct btrfs_delayed_ref_head *head, struct btrfs_delayed_extent_op *extent_op) { :: } else { err = -EIO; goto out; } ---- > but other than that I have never had io errors before, or any other troubles. Hm. btrfs dev stat shows real disk IO errors. As this FS isn't mountable .. pls try btrfs dev stat > file search for 'device stats', there will be one for each disk. Or it reports in the syslog when it happens not necessarily during dedupe. > One of my other filesystems share the same two discs and it is still fine, so I think the hardware is probably ok. Right. I guess that too. A confirmation will be better. > I've copied the beginning of the errors below. At my end finding the root cause of 'parent transid verify failed' during/after dedupe is is kind of fading as disk seems to be had no issues. which I had in mind. Also, there wasn't abrupt power-recycle here? I presume. It's better to save the output disk1-log and disk2-log as below before further efforts to recovery. Just in case if something pops out. btrfs in dump-super -fa disk1 > disk1-log btrfs in dump-tree --degraded disk1 >> disk1-log [1] btrfs in dump-super -fa disk2 > disk2-log btrfs in dump-tree --degraded disk2 >> disk2-log [1] [1] --degraded option is in the ML. [PATCH] btrfs-progs: dump-tree: add degraded option Thanks, Anand