Re: [PATCH v2 2/2] md/raid5: use consistency_policy to remove journal feature

[Shaohua Li <shli@kernel.org>]

On Thu, Mar 16, 2017 at 03:14:09PM -0700, Song Liu wrote:
> When journal device of an array fails, the array is forced into read-only
> mode. To make the array normal without adding another journal device, we
> need to remove journal _feature_ from the array.
> 
> This patch allows remove journal _feature_ from an array, For journal
> existing journal should be either missing or faulty.
> 
> To remove journal feature, one can simply echo into the sysfs file:
> 
>  cat /sys/block/md0/md/consistency_policy
>  journal
> 
>  echo resync > /sys/block/md0/md/consistency_policy
>  cat /sys/block/md0/md/consistency_policy
>  resync

please describe the detail steps what you want to do after a journal disk
failure here, like remove the disk. maybe create a new test case too.

> 
> Signed-off-by: Song Liu <songliubraving@fb.com>
> ---
>  drivers/md/raid5.c | 18 +++++++++++++-----
>  1 file changed, 13 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
> index 7a6e7ea..2302ec4 100644
> --- a/drivers/md/raid5.c
> +++ b/drivers/md/raid5.c
> @@ -8369,11 +8369,19 @@ static int raid5_change_consistency_policy(struct mddev *mddev, const char *buf)
>  		if (!err)
>  			raid5_reset_stripe_cache(mddev);
>  		mddev_resume(mddev);
> -	} else if (strncmp(buf, "resync", 6) == 0 && raid5_has_ppl(conf)) {
> -		mddev_suspend(mddev);
> -		log_exit(conf);
> -		raid5_reset_stripe_cache(mddev);
> -		mddev_resume(mddev);
> +	} else if (strncmp(buf, "resync", 6) == 0) {
> +		if (raid5_has_ppl(conf)) {
> +			mddev_suspend(mddev);
> +			log_exit(conf);
> +			raid5_reset_stripe_cache(mddev);
> +			mddev_resume(mddev);
> +		} else if (test_bit(MD_HAS_JOURNAL, &conf->mddev->flags) &&
> +			   r5l_log_disk_error(conf)) {
> +			mddev_suspend(mddev);
> +			clear_bit(MD_HAS_JOURNAL, &mddev->flags);
> +			mddev_resume(mddev);

should this happen after journal disk is removed? I'm wondering if this need
extra checks.

Thanks,
Shaohua