Controlling the speed of disk scrubbing

About disk scrubbing

Disk scrubbing means checking the disk blocks of all disks on the filer for media errors and consistency. If Data ONTAP finds media errors or inconsistencies, it fixes them by reconstructing the data from other disks and rewriting the data. Disk scrubbing reduces the chance of a multiple disk failure that is so extensive that RAID reconstruction is impossible.

Data ONTAP enables block checksums to ensure data integrity. If checksums are incorrect, Data ONTAP generates an error message similar to the following:

Scrub found checksum error on /vol/vol0/plex0/rg0/4.0 block 436964

If RAID 4 is enabled, Data ONTAP scrubs a RAID group only when all the group's disks are operational.

If RAID DP is enabled, Data ONTAP can carry out a scrub even if one disk in the RAID group has failed.

Although disk scrubbing slows the filer somewhat, network clients might not notice the change in filer performance because disk scrubbing starts automatically at 1:00 a.m. on Sunday, when most systems are lightly loaded, and stops after six hours.

Modifying disk scrubbing speed

To modify the speed of disk scrubbing, complete the following step.

Step	Action
1	Enter the following command: options raid.scrub.perf_impact impact impact can be high, medium, or low. High means that the filer uses most of the CPU time available for disk scrubbing; this setting can heavily affect filer performance. Low means that the filer uses very little CPU time; this setting lightly affects filer performance. The default speed is low.

Step

Action

Enter the following command:

options raid.scrub.perf_impact impact

impact can be high, medium, or low. High means that the filer uses most of the CPU time available for disk scrubbing; this setting can heavily affect filer performance. Low means that the filer uses very little CPU time; this setting lightly affects filer performance. The default speed is low.

RAID operations affecting disk scrubbing speed

When disk scrubbing and mirror verification are running at the same time, Data ONTAP limits the combined resource utilization to the greatest impact set by either operation. For example, if raid.verify.perf_impact is set to medium and raid.scrub.perf_impact is set to low, the resource utilization by both operations has a medium impact.