Re: Fio could damage some disk/partitions already used

On Thu, Jul 19 2007, Erwan Velu wrote:
> Jens Axboe wrote:
>> Not much I can do about that... O_EXCL doesn't really help very much, as
>> it only applies to others that use O_EXCL. And you really do want
>> several fio jobs to access a raw partition as well, so it would rule out
>> some interesting tests.
>>
> But for example e2fsck is able to say my partition is already mounted and
> shouldn't be fscked.
> I tried to see how does e2fsck to handle this and it sounds it tries to
> open the devices as RO O_EXCL to check if the partition is already accessed
> in RW..
>
> [...]
> open("/dev/sda8", O_RDONLY|O_EXCL) = -1 EBUSY (Device or resource
> busy)
>
> As I'm not a kernel expert, I surely missed something ;))

But it's a difference use case. What if I do:

[job1]
filename=/dev/sda8
rw=randread

[job2]
filename=/dev/sda8
rw=randread

?

My point is that the fio write tests are destructive, that's just the
way it works. Trying to prevent all possible ways that a user can shoot
himself in the foot just ends up crippling fio and there will still be
ways that he can screw up :-)

So I appreciate your concerns in this area, however I'm not sure I can
do much about it...

-- 
Jens Axboe