2012-04-02 15:17:20

by Daniel Drake

[permalink] [raw]
Subject: ext4 online resize and robustness to interruption

Hi,

I think this question has been asked before, but I'd like to revisit
it given the recent rework of ext4 online resizing:

How robust is ext4 online resize in the face of unexpected
interruption (i.e. sudden loss of power)?

The context here is that OLPC will now performing an online resize
upon first boot of the system. Our laptops are frequently used in very
remote places with unreliable power, and given that we work in large
numbers, we're going to see some cases where the system lost power
during the resize operation. What is the expected effect here - will
the system boot up again just fine (perhaps with the original
partition size), and is then repeating the resize operation expected
to be safe and reliable?

Thanks,
Daniel


2012-04-02 15:20:53

by Andreas Dilger

[permalink] [raw]
Subject: Re: ext4 online resize and robustness to interruption

On 2012-04-02, at 9:17 AM, Daniel Drake wrote:
> I think this question has been asked before, but I'd like to revisit
> it given the recent rework of ext4 online resizing:
>
> How robust is ext4 online resize in the face of unexpected
> interruption (i.e. sudden loss of power)?

It used to be the case that resize was totally robust against failure,
when resize was done on a group-by-group basis. I can't comment on
the behaviour in this respect with the new code, but there is some
concern in my mind because the new code has removed the transaction
commits in the middle of the resize for performance reasons.

> The context here is that OLPC will now performing an online resize
> upon first boot of the system. Our laptops are frequently used in very
> remote places with unreliable power, and given that we work in large
> numbers, we're going to see some cases where the system lost power
> during the resize operation. What is the expected effect here - will
> the system boot up again just fine (perhaps with the original
> partition size), and is then repeating the resize operation expected
> to be safe and reliable?
>
> Thanks,
> Daniel
> --
> To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html


Cheers, Andreas






2012-04-03 10:58:30

by Yongqiang Yang

[permalink] [raw]
Subject: Re: ext4 online resize and robustness to interruption

On Mon, Apr 2, 2012 at 11:17 PM, Daniel Drake <[email protected]> wrote:
> Hi,
>
> I think this question has been asked before, but I'd like to revisit
> it given the recent rework of ext4 online resizing:
>
> How robust is ext4 online resize in the face of unexpected
> interruption (i.e. sudden loss of power)?
Yep! It is only robust when journal is used. In no journal mode, it is not.

Yongqiang.
>
> The context here is that OLPC will now performing an online resize
> upon first boot of the system. Our laptops are frequently used in very
> remote places with unreliable power, and given that we work in large
> numbers, we're going to see some cases where the system lost power
> during the resize operation. What is the expected effect here - will
> the system boot up again just fine (perhaps with the original
> partition size), and is then repeating the resize operation expected
> to be safe and reliable?
>
> Thanks,
> Daniel



--
Best Wishes
Yongqiang Yang