xfs
[Top] [All Lists]

Re: task xfssyncd blocked while raid5 was in recovery

To: 韩国中 <vincent.han.megan@xxxxxxxxx>
Subject: Re: task xfssyncd blocked while raid5 was in recovery
From: Dave Chinner <david@xxxxxxxxxxxxx>
Date: Thu, 11 Oct 2012 16:09:55 +1100
Cc: xfs@xxxxxxxxxxx
In-reply-to: <CAE-xvygu7j4zTBVRmm93-VbzhwiU2enEqwnFOPy5gbVHBn1CEQ@xxxxxxxxxxxxxx>
References: <CAE-xvygu7j4zTBVRmm93-VbzhwiU2enEqwnFOPy5gbVHBn1CEQ@xxxxxxxxxxxxxx>
User-agent: Mutt/1.5.21 (2010-09-15)
On Thu, Oct 11, 2012 at 11:55:01AM +0800, 韩国中 wrote:
>  Hello, every one:
> 
> Recently, a problem has troubled me for a long time.
> 
> I created a 4*2T (sda, sdb, sdc, sdd) raid5 with XFS file system, 128K
> chuck size and 2048 strip_cache_size. The mdadm 3.2.2, kernel 2.6.38
> and mkfs.xfs 3.1.1 were used. When the raid5 was in recovery and the
> schedule reached 47%, I/O errors occurred in sdb. The following was
> the output:
> 
> ......
> 
> ata2: translated ATA stat/err 0x41/04 to SCSI SK/ASC/ASCQ 0xb/00/00
> 
> ata2: status=0x41 { DriveReady Error }

Looks like you've had a drive fail during rebuild.

> Then, there were lots of error messages about the file system. The
> following was the output:
> 
> 
> 
> ......
> 
> INFO: task xfssyncd/md127:1058 blocked for more than 120 seconds.
> 
> "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
> xfssyncd/md127  D fffffff7000216d0     0  1058      2 0x00000000
>   frame 0: 0xfffffff700020570 __switch_to+0x1b8/0x1c0 (sp 0xfffffe008d7ff900)
>   frame 1: 0xfffffff7000216d0 schedule+0x918/0x1538 (sp 0xfffffe008d7ff9d0)
>   frame 2: 0xfffffff700022a90 schedule_timeout+0x268/0x5b0 (sp 
> 0xfffffe008d7ffd18)
>   frame 3: 0xfffffff700024ee0 __down+0xd8/0x158 (sp 0xfffffe008d7ffda8)
>   frame 4: 0xfffffff70085da78 down.cold+0x8/0x28 (sp 0xfffffe008d7ffe18)
>   frame 5: 0xfffffff700750788 xfs_buf_lock+0xd0/0x120 (sp 0xfffffe008d7ffe38)
>   frame 6: 0xfffffff700821b40 xfs_getsb+0x38/0x78 (sp 0xfffffe008d7ffe50)
>   frame 7: 0xfffffff70077e230 xfs_trans_getsb+0xe0/0x100 (sp 
> 0xfffffe008d7ffe68)
>   frame 8: 0xfffffff7006babc0 xfs_mod_sb+0x88/0x198 (sp 0xfffffe008d7ffe88)
>   frame 9: 0xfffffff7007a6480 xfs_fs_log_dummy+0x68/0xe0 (sp 
> 0xfffffe008d7ffeb8)
>   frame 10: 0xfffffff70079c6c0 xfs_sync_worker+0xe0/0xe8 (sp 
> 0xfffffe008d7ffed8)
>   frame 11: 0xfffffff700570a00 xfssyncd+0x240/0x328 (sp 0xfffffe008d7ffef0)
>   frame 12: 0xfffffff7000f0530 kthread+0xe0/0xe8 (sp 0xfffffe008d7fff80)
>   frame 13: 0xfffffff7000bab38 start_kernel_thread+0x18/0x20 (sp 
> 0xfffffe008d7fffe8)

Which is basically saying that the superblock buffer is under IO -
that's the only reason it ever gets locked.

> The output said “INFO: task xfssyncd/md127:1058 blocked for more than
> 120 seconds? What did that mean? I used “cat /proc/mdstat?to see the
> state of the raid5. The output was:
> 
> Personalities : [raid0] [raid6] [raid5] [raid4]
> 
> md127 : active raid5 sdd[3] sdc[2] sdb[1](F) sda[0]
> 
>       5860540032 blocks super 1.2 level 5, 128k chunk, algorithm 2 [4/3] 
> [U_UU]
> 
>         resync=PENDING
> 
> unused devices: <none>
> 
> 
>      The state of the raid5 was “PENDING? I had never seen such a
> state of raid5 when I used ext4. After that, I wrote a program to access the
> raid5, there was no response any more.

Waiting on IO to complete, but with the MD device down, it will
enver complete.

> Then I used “ps aux| task
> xfssyncd?to see the state of “xfssyncd? Unfortunately, there was no
> response yet. Then I tried “ps aux? There were outputs, but the
> program could exit with “Ctrl+d? or “Ctrl+z? And when I tested the
> write performance for raid5, I/O errors often occurred. I did not know
> why this I/O errors occurred so frequently.
> 
>      What was the problem? Can any one help me?

Broken hardware causing MD to go into a bad state, which causes XFS
to stall because it can't make progress.

Bottom line: replace the broken disk, though given that MD was
already rebuilding a RAID5 when the disk died, you probably have
lost everything on the filesystem....

Cheers,

Dave.
-- 
Dave Chinner
david@xxxxxxxxxxxxx

<Prev in Thread] Current Thread [Next in Thread>