On Fri, Jun 18, 2021 at 1:09 AM Darrick J. Wong <djwong@xxxxxxxxxx> wrote:
>
> On Fri, Jun 18, 2021 at 12:06:40AM +0800, Haiwei Li wrote:
> > Hi,
> >
> > Sorry to bother. I get a xfs error on kernel 3.10.x. And i don't know how to
> > debug. I got nothing useful from search engines. So I sent an e-mail
> > here. If there
> > are other more suitable ways to discuss the problem, please let me know, thanks!
> >
> > I have gotten a message on the console.
> >
> > '-bash: /data/.my_history: Input/output error'
> >
> > I tried:
> >
> > # ls -l / | grep data
> > ls: cannot access /data: Input/output error
> > d????????? ? ? ? ? ? data
> >
> > The mount point info is:
> >
> > '/dev/vdb on /data type xfs (rw,noatime,attr2,inode64,prjquota)'
> >
> > System log messages as below:
> >
> > ffff882b86a34000: 31 38 38 32 30 31 36 0a 00 00 00 00 00 00 00 00
> > 1882016.........
> > ffff882b86a34010: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
> > ................
> > ffff882b86a34020: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
> > ................
> > ffff882b86a34030: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
> > ................
> > XFS (vdb): Internal error xfs_inode_buf_verify at line 410 of file
> > XXXX/fs/x[2021-06-17 18:28:51]fs/xfs_inode.c. Caller
> > 0xffffffffa04d410e
> >
> > CPU: 0 PID: 7715 Comm: kworker/0:1H Tainted: G U O
> > 3.10.107-1-tlinux2_kvm_guest-0051 #1
> > Hardware name: Smdbmds KVM, BIOS seabios-1.9.1-qemu-project.org 04/01/2014
> > Workqueue: xfslogd xfs_buf_iodone_work [xfs]
> > ffff882ec52d9000 000000001a1ab7e6 ffff882efa97fd50 ffffffff819f1d23
> > ffff882efa97fd68 ffffffffa047da9b ffffffffa04d410e ffff882efa97fda0
> > ffffffffa047daf5 0000019a00000001 0000000000000001 ffff882b86a34000
> > Call Trace:
> > [<ffffffff819f1d23>] dump_stack+0x19/0x1b
> > [<ffffffffa047da9b>] xfs_error_report+0x3b/0x40 [xfs]
> > [<ffffffffa04d410e>] ? xfs_inode_buf_read_verify+0xe/0x10 [xfs]
> > [<ffffffffa047daf5>] xfs_corruption_error+0x55/0x80 [xfs]
> > [<ffffffffa04d40a4>] xfs_inode_buf_verify+0x94/0xe0 [xfs]
> > [<ffffffffa04d410e>] ? xfs_inode_buf_read_verify+0xe/0x10 [xfs]
> > [<ffffffffa04d410e>] xfs_inode_buf_read_verify+0xe/0x10 [xfs]
> > [<ffffffffa047b305>] xfs_buf_iodone_work+0xa5/0xd0 [xfs]
> > [<ffffffff8106c00c>] process_one_work+0x17c/0x450
> > [<ffffffff8106cebb>] worker_thread+0x11b/0x3a0
> > [<ffffffff8106cda0>] ? manage_workers.isra.26+0x2a0/0x2a0
> > [<ffffffff810737cf>] kthread+0xcf/0xe0
> > [<ffffffff81073700>] ? insert_kthread_work+0x40/0x40
> > [<ffffffff81ad5908>] ret_from_fork+0x58/0x90
> > [<ffffffff81073700>] ? insert_kthread_work+0x40/0x40
> > XFS (vdb): Corruption detected. Unmount and run xfs_repair
> > ffff882b86a34100: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
> > ................
> > ffff882b86a34110: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
> > ................
> > ffff882b86a34120: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
> > ................
> > ffff882b86a34130: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
> > ................
> >
> > I reboot and remount. It works normally. No error again. I guess data from the
> > wrong blocks was returned to XFS.
> >
> > I have no idea how to reproduce. Our workload sometimes triggers the problem.
> > To data, the problem only occurs on 3.10.x in three versions 3.10.x, 4.14.x and
> > 5.4.x.
> >
> > Environment: Containers with workload are running in a kvm vm. The problem
> > occurs in the kvm vm.
> >
> > Any ideas on how to debug? Thanks!
>
> Uh, does xfs_repair -n on the unmounted filesystem complain about this
> corrupt inode?
Output as below:
# xfs_repair -n /dev/vdb
Phase 1 - find and verify superblock...
Phase 2 - using internal log
- scan filesystem freespace and inode maps...
- found root inode chunk
Phase 3 - for each AG...
- scan (but don't clear) agi unlinked lists...
- process known inodes and perform inode discovery...
- agno = 0
directory flags set on non-directory inode 64
directory flags set on non-directory inode 69
directory flags set on non-directory inode 72
directory flags set on non-directory inode 75
directory flags set on non-directory inode 81
directory flags set on non-directory inode 83
directory flags set on non-directory inode 91
directory flags set on non-directory inode 112
directory flags set on non-directory inode 125
directory flags set on non-directory inode 64
would fix bad flags.
directory flags set on non-directory inode 69
would fix bad flags.
directory flags set on non-directory inode 72
would fix bad flags.
directory flags set on non-directory inode 75
would fix bad flags.
directory flags set on non-directory inode 81
would fix bad flags.
directory flags set on non-directory inode 83
......
directory flags set on non-directory inode 3242233222
would fix bad flags.
- process newly discovered inodes...
Phase 4 - check for duplicate blocks...
- setting up duplicate extent list...
- check for inodes claiming duplicate blocks...
- agno = 0
directory flags set on non-directory inode 64
would fix bad flags.
- agno = 2
- agno = 3
directory flags set on non-directory inode 69
would fix bad flags.
- agno = 1
directory flags set on non-directory inode 2149531806
would fix bad flags.
......
directory flags set on non-directory inode 3242233222
would fix bad flags.
No modify flag set, skipping phase 5
Phase 6 - check inode connectivity...
- traversing filesystem ...
- traversal finished ...
- moving disconnected inodes to lost+found ...
Phase 7 - verify link counts...
No modify flag set, skipping filesystem flush and exiting.
Any suggestions on how to debug? Thanks a lot.
--
Haiwei
