Re: "md/raid10:md5: sdf: redirecting sector 2979126480 to another mirror"

[Guoqing Jiang <guoqing.jiang@xxxxxxxxxxxxxxx>]

Hi,

On 1/7/21 00:27, Andy Smith wrote:
> Hi,
>
> "md/raid10:md5: sdf: redirecting sector 2979126480 to another
> mirror"
>
> I've actually been seeing these messages very occasionally across
> many machines for years but have never found anything wrong so kept
> putting investigation of it to the bottom of my TODO list. I have
> even in the past upon seeing this done a full scrub and check and
> found no issue.
>
> Having just seen one of them again now, and having some spare time I
> tried to look into it.
>
> So, this messages comes from here:
>
>      https://github.com/torvalds/linux/blob/v5.8/drivers/md/raid10.c#L1188
>
> but under what circumstances does it actually happen?
>
> This time, as with the other times, I cannot see any indication of
> read error (i.e. no logs of that) and no problems apparent in SMART
> data.
>
> err_rdev there can only be set inside the block above that starts
> with:
>
>
>      if (r10_bio->devs[slot].rdev) {
>          /*
>           * This is an error retry, but we cannot
>           * safely dereference the rdev in the r10_bio,
>           * we must use the one in conf.
>
> …but why is this an error retry? Nothing was logged so how do I find
> out what the error was?

This is because handle_read_error also calls raid10_read_request, pls 
see commit 545250f2480 ("md/raid10: simplify handle_read_error()").

I guess it is better to distinguish the caller to avoid the normal read 
path prints the message too (if it is the problem).

--- a/drivers/md/raid10.c
+++ b/drivers/md/raid10.c
@@ -1115,7 +1115,7 @@ static void regular_request_wait(struct mddev 
*mddev, struct r10conf *conf,
 }

 static void raid10_read_request(struct mddev *mddev, struct bio *bio,
-                               struct r10bio *r10_bio)
+                               struct r10bio *r10_bio, bool read_err)
 {
        struct r10conf *conf = mddev->private;
        struct bio *read_bio;
@@ -1128,7 +1128,7 @@ static void raid10_read_request(struct mddev 
*mddev, struct bio *bio,
        struct md_rdev *err_rdev = NULL;
        gfp_t gfp = GFP_NOIO;

-       if (slot >= 0 && r10_bio->devs[slot].rdev) {
+       if (read_err && slot >= 0 && r10_bio->devs[slot].rdev) {
                /*
                 * This is an error retry, but we cannot
                 * safely dereference the rdev in the r10_bio,
@@ -1495,7 +1495,7 @@ static void __make_request(struct mddev *mddev, 
struct bio *bio, int sectors)
        memset(r10_bio->devs, 0, sizeof(r10_bio->devs[0]) * conf->copies);

        if (bio_data_dir(bio) == READ)
-               raid10_read_request(mddev, bio, r10_bio);
+               raid10_read_request(mddev, bio, r10_bio, false);
        else
                raid10_write_request(mddev, bio, r10_bio);
 }
@@ -2586,7 +2586,7 @@ static void handle_read_error(struct mddev *mddev, 
struct r10bio *r10_bio)
        rdev_dec_pending(rdev, mddev);
        allow_barrier(conf);
        r10_bio->state = 0;
-       raid10_read_request(mddev, r10_bio->master_bio, r10_bio);
+       raid10_read_request(mddev, r10_bio->master_bio, r10_bio, true);
 }

Thanks,
Guoqing