Continuous media scrub

About continuous media scrubbing

By default, Data ONTAP runs continuous background media scrubbing for media errors on filer disks. The purpose of the continuous media scrub is to detect and scrub media errors in order to minimize the chance of filer disruption due to media error while a filer is in degraded or reconstruction mode.

Negligible performance impact

Because continuous media scrubbing searches only for media errors, the impact on system performance is negligible.

Note
Media scrubbing is a continuous background process. Therefore, you might observe disk LEDs blinking on an apparently idle system. You might also observe some CPU activity even when no user workload is present. The media scrub attempts to exploit idle disk bandwidth and free CPU cycles to make faster progress. However, any client workload results in aggressive throttling of the media scrub resource.

Not a substitute for a scheduled disk scrub

Because the continuous process described in this section scrubs only media errors, Network Appliance strongly recommends that you continue to run the filer's scheduled complete disk scrub operation, which is described in Scheduling or manually running disk scrubbing. The complete disk scrub carries out parity and checksum checking and repair operations, in addition to media checking and repair operations, on a scheduled rather than a continuous basis.

Adjusting maximum time for a media scrub cycle

You can lessen the CPU resource that would be consumed by the continuous media scrub under a heavy client workload by increasing the maximum time allowed for a media scrub cycle to complete from the default 72 hours.

By default, one cycle of a filer's continuous media scrub can take a maximum of 72 hours to complete. In most situations, one cycle of the continuous media scrub will complete in a much shorter time; however, under heavy client workload conditions, the default 72-hour maximum ensures that whatever the client load on the filer, enough CPU resource will be allotted to the media scrub to ensure that one cycle will complete in no more than 72 hours.

If you want the media scrub operation to consume even less CPU resource under heavy filer client workload, you can increase the maximum number of hours for the media scrub cycle, thus lessening the CPU resource consumed by the media scrub in times of heavy filer usage.

To extend or shorten the maximum time for a media scrub cycle, enter the following command at the filer command line:

options raid.media_scrub.deadline max_hrs_per_cycle

max_hrs_per_cycle is the maximum number of hours that you want to allow for one cycle of the continuous media scrub. Valid options range from 72 to 336 hours.

Disabling continuous media scrubbing

Network Appliance recommends that you keep continuous media error scrubbing enabled, but you might decide to disable your continuous media scrub in situations in which your filer is carrying out operations with heavy performance impact and in which you have alternative measures (such as volume SyncMirror replication or RAID DP configuration) in place that prevent data loss due to filer disruption or double disk failure.

To disable continuous media scrubbing, enter the following command on the filer command line:

options raid.media_scrub.enable off

Note
To restart continuous media scrubbing after you have disabled it, enter the following command:

options raid.media_scrub.enable on

Checking media scrub activity

You can confirm media scrub activity on your filer by entering the following command:

vol media_scrub status [/vol[/plex][/raidgroup]] [-v]

[/vol[/plex] [/raidgroup]] is the optional pathname to the volume, plex, or RAID group on which you want to confirm media scrubbing activity.

[-v] is the optional parameter for the verbose version of the media scrubbing activity status. The verbose version of status displays the percentage of the current scrub that is complete, the start time of the current media scrub, and the completion time of the last media scrub.

Example 1. Checking of filer-wide media scrubbing

The following command displays media scrub status information for all the volumes on the filer.

vol media_scrub status

vol media_scrub /vol0/plex0/rg0 is 28% complete

vol media_scrub /vol2/plex0/rg0 is 3% complete

vol media_scrub /vol3/plex0/rg0 is 3% complete

vol media_scrub /vol3/plex0/rg1 is 3% complete

vol media_scrub /vol3/plex0/rg2 is 3% complete

vol media_scrub /vol3/plex0/rg3 is 3% complete

Example 2. Verbose checking of filer-wide media scrubbing

The following command displays verbose media scrub status information for all the volumes on the filer.

vol media_scrub status -v

vol media_scrub: status of /vol0/plex0/rg0 :

  Current instance of media_scrub is 32% complete.

  Media scrub started at Tue Aug 12 19:24:00 GMT 2003

  Last full media_scrub completed: Tue Aug 12 19:23:39 GMT 2003

vol media_scrub: status of /vol2/plex0/rg0 :

  Current instance of media_scrub is 8% complete.

  Media scrub started at Tue Aug 12 19:30:00 GMT 2003

  Last full media_scrub completed: Tue Aug 12 19:27:37 GMT 2003

Example 3. Checking for media scrubbing on a specific volume

The following command displays media scrub status information for the volume vol2.

vol media_scrub status /vol2

  vol media_scrub /vol2/plex0/rg0 is 4% complete

  vol media_scrub /vol2/plex0/rg1 is 10% complete