repating 208 failures

[Dave Chinner]

On Sun, May 04, 2014 at 02:42:18AM -0700, Christoph Hellwig wrote:
> Did anyone see failure like:
> 
> generic/208 201s ... 200s
> umount: /mnt/test: device is busy.
>         (In some cases useful info about processes that use
>          the device is found by lsof(8) or fuser(1))
> 
> which then abort xfstests?  I've seen them occasionally for a long time,
> but these days they seem to happen more than 50% of the time.

I see it occassionally. Most often on a single processor test VM,
but not regularly enough to debug. It's an AIO stress test, so
perhaps there's a problem with delayed AIO reference counts, because
a second attempt to unmount the device will succeed....

Cheers,

Dave.
-- 
Dave Chinner
david at fromorbit.com