On 11/5/19 4:33 PM, Song Liu wrote:
On Tue, Nov 5, 2019 at 1:14 PM Song Liu <liu.song.a23@xxxxxxxxx> wrote:
On Mon, Nov 4, 2019 at 12:02 PM Nigel Croxon <ncroxon@xxxxxxxxxx> wrote:
The MD driver for level-456 should prevent re-reading read errors.
For redundant raid it makes no sense to retry the operation:
When one of the disks in the array hits a read error, that will
cause a stall for the reading process:
- either the read succeeds (e.g. after 4 seconds the HDD error
strategy could read the sector)
- or it fails after HDD imposed timeout (w/TLER, e.g. after 7
seconds (might be even longer)
The user can enable/disable this functionality by the following
commands:
To Enable:
echo 1 > /proc/sys/dev/raid/raid456_retry_read_error
To Disable, type the following at anytime:
echo 0 > /proc/sys/dev/raid/raid456_retry_read_error
Signed-off-by: Nigel Croxon <ncroxon@xxxxxxxxxx>
---
drivers/md/md.c | 43 +++++++++++++++++++++++++++++++++++++++++++
drivers/md/md.h | 3 +++
drivers/md/raid5.c | 3 ++-
3 files changed, 48 insertions(+), 1 deletion(-)
diff --git a/drivers/md/md.c b/drivers/md/md.c
index 6f0ecfe8eab2..75b8b0615328 100644
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -125,6 +125,12 @@ static inline int speed_max(struct mddev *mddev)
mddev->sync_speed_max : sysctl_speed_limit_max;
}
+static int sysctl_raid456_retry_read_error = 0;
+static inline void set_raid456_retry_re(struct mddev *mddev, int re)
+{
+ (re ? set_bit : clear_bit)(MD_RAID456_RETRY_RE, &mddev->flags);
Let's just keep this
if (re)
set_bit(...);
else
clear_bit(..);
+}
+
static int rdev_init_wb(struct md_rdev *rdev)
{
if (rdev->bdev->bd_queue->nr_hw_queues == 1)
@@ -213,6 +219,13 @@ static struct ctl_table raid_table[] = {
.mode = S_IRUGO|S_IWUSR,
.proc_handler = proc_dointvec,
},
+ {
+ .procname = "raid456_retry_read_error",
+ .data = &sysctl_raid456_retry_read_error,
+ .maxlen = sizeof(int),
+ .mode = S_IRUGO|S_IWUSR,
+ .proc_handler = proc_dointvec,
+ },
{ }
};
How about we add this to md_redundancy_attrs instead?
Actually, I think raid5_attrs is better.
Thanks,
Song
I'll have to test that, raid5_attrs.
But the ev has to be available before the array is created.
That way when the initial sync starts, a user can already enable/disable
the capability.
-Nigel
