Re: System hangs with vnode backed md(4) devices -- still a problem?

From: Nik Clayton <nik_at_freebsd.org>
Date: Sun, 7 Mar 2004 21:42:18 +0000
On Wed, Mar 03, 2004 at 10:07:17PM +0000, Nik Clayton wrote:
> Toward the end of last year, through to earlier this year, -current had  
> a problem with vnode backed md(4) devices.  Specifically,
> 
>     truncate -s 6G foo.md
>     mdconfig -a -t vnode -f foo.md -u 0
>     newfs /dev/md0
>     mount /dev/md0 /mnt
>     rsync /large/amount/of/data /mnt
> 
> would cause the rsync process to eventually get stuck in a wdrain  
> state, and other processes in the system would gradually hang --  
> processes were still running, but doing anything that might need to  
> read from a buffer caused the process to hang.  Eventually, everything  
> would be hung.
> 
> If I'm reading the mailing lists right, this was supposed to be solved  
> here:
>      
> http://lists.freebsd.org/pipermail/freebsd-current/2003-November/ 
> 013183.html
> 
> However, I'm still seeing the problem in -current from Jan 14th.  So  
> I'm wondering if anyone else is seeing it, and if not, a rough idea of  
> when it was fixed.  

Following some off-list suggestions, I can report that 

    mount -o sync /dev/md0 /mnt

solves the "copying large amounts of data will hang" problem.  However,
it replaces it with an "unmounting the md0 device hangs" problem.

N
-- 
FreeBSD: The Power to Serve      http://www.freebsd.org/               (__)
FreeBSD Documentation Project    http://www.freebsd.org/docproj/    \\\'',)
                                                                      \/  \ ^
   --- 15B8 3FFC DDB4 34B0 AA5F  94B7 93A8 0764 2C37 E375 ---         .\._/_)

Received on Sun Mar 07 2004 - 12:42:21 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:37:46 UTC