Hi Marc,
On 02/09/2018 02:29 PM, Marc MERLIN wrote:
> But, I'm confused by what happened. The md check ran to completion.
> It found things and supposedly fixed them:
> [240351.053406] md/raid:md7: read error corrected (8 sectors at
> 9159374528 on sdf1)
> Strangely, it did nothing with this:
> [287271.959779] sd 4:4:0:0: [sdh] tag#6 Add. Sense: Unrecovered read
> error - auto reallocate failed
> Now, the sync is comnplete, and my bad blocks are still there?
> myth:~# smartctl -A /dev/sdh
> 196 Reallocated_Event_Count 0x0032 200 200 000 Old_age
> Always - 0
> 197 Current_Pending_Sector 0x0032 200 200 000 Old_age
> Always - 2
>
> myth:~# smartctl -A /dev/sdf
> 196 Reallocated_Event_Count 0x0032 200 200 000 Old_age
> Always - 0
> 197 Current_Pending_Sector 0x0032 200 200 000 Old_age
> Always - 7
>
> The pending sectors should have been re-written and become
> Reallocated_Event_Count, no?
Yes, and not necessarily. Pending sectors can be non-permanent errors
-- the drive firmware will test a pending sector immediately after write
to see if the write is readable. If not, it will re-allocate while it
still has the write data in its buffers. Otherwise, it'll clear the
pending sector.
> So, mdadm is happy allegedly, but my drives still have the same bad
> sectors they had (more or less).
If you have bad block lists enabled in your array, MD will *never* try
to fix the underlying sectors. Please show your mdadm -E reports for
these devices. If necessary, stop the array and re-assemble with the
options to disable bad block lists. { How this misfeature got into the
kernel and enabled by default baffles me. }
Also, pending sectors that are in dead zones between metadata and array
data will not be accessed by a check scrub, and will therefore persist.
> Yes, I know I should trash (return) those drives,
Well, non-permanent read errors are not considered warranty failures.
They are in the drive specs. When pending is zero and actual
re-allocations are climbing (my threshold is double digits), *then* it's
time to replace.
> but I still want to understand why I can't get basic block remapping
> working Any idea what went wrong?
Invalid expectations, perhaps.
Phil
--
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
