On Fri, Nov 06 2015, rgoldwyn@xxxxxxx wrote:
> From: Goldwyn Rodrigues <rgoldwyn@xxxxxxxx>
>
> If a RESYNCING message with (0,0) has been sent before, do not send it
> again. This avoids a resync ping pong between the nodes. We read
> the bitmap lockresource's LVB to figure out the previous value
> of the RESYNCING message.
This seems like a bandaid more than a real fix, but maybe I'm wrong.
I'll apply it for now. If I end of figuring out that there is some
other issue I discuss it then.
>
> Signed-off-by: Goldwyn Rodrigues <rgoldwyn@xxxxxxxx>
> ---
> drivers/md/md-cluster.c | 8 ++++++++
> 1 file changed, 8 insertions(+)
>
> diff --git a/drivers/md/md-cluster.c b/drivers/md/md-cluster.c
> index a93734e..3daa464 100644
> --- a/drivers/md/md-cluster.c
> +++ b/drivers/md/md-cluster.c
> @@ -898,8 +898,16 @@ static int resync_start(struct mddev *mddev)
> static int resync_info_update(struct mddev *mddev, sector_t lo, sector_t hi)
> {
> struct md_cluster_info *cinfo = mddev->cluster_info;
> + struct resync_info ri;
> struct cluster_msg cmsg = {0};
>
> + /* do not send zero again, if we have sent before */
> + if (hi == 0) {
> + memcpy(&ri, cinfo->bitmap_lockres->lksb.sb_lvbptr, sizeof(struct resync_info));
> + if (le64_to_cpu(ri.hi) == 0)
> + return 0;
I feel this would look so much better as:
if (hi == 0) {
struct resync_info ri;
memcpy(&ri, cinfo->bitmap_lockres->lksb.sb_lvbptr,
sizeof(ri));
}
so if you happened to resend it like that I'd probably remove what I
have already applied and use this instead.
(make variable more local, use sizeof(var) instead of sizeof(type), and
wrap the long line)
Thanks,
NeilBrown
> + }
> +
> add_resync_info(mddev, cinfo->bitmap_lockres, lo, hi);
> /* Re-acquire the lock to refresh LVB */
> dlm_lock_sync(cinfo->bitmap_lockres, DLM_LOCK_PW);
> --
> 1.8.5.6
Attachment:
signature.asc
Description: PGP signature
