PDF(555.9 KB) View with Adobe Reader on a variety of devices
ePub(643.3 KB) View in various apps on iPhone, iPad, Android, Sony Reader, or Windows Phone
Mobi (Kindle)(741.6 KB) View on Kindle device or Kindle app on multiple devices
Updated:April 22, 2020
The documentation set for this product strives to use bias-free language. For the purposes of this documentation set, bias-free is defined as language that does not imply discrimination based on age, disability, gender, racial identity, ethnic identity, sexual orientation, socioeconomic status, and intersectionality. Exceptions may be present in the documentation due to language that is hardcoded in the user interfaces of the product software, language used based on RFP documentation, or language that is used by a referenced third-party product. Learn more about how Cisco is using Inclusive Language.
This document describes different types of disk errors, how to classify them, and tools you can use to identify them.
There are no specific requirements for this document.
The information in this document is based on hard disks in Unified Computing System (UCS).
The information in this document was created from the devices in a specific lab environment. All of the devices used in this document started with a cleared (default) configuration. If your network is live, make sure that you understand the potential impact of any command.
The document also outlines the Hard Disk Drive's (HDD) and Redundant Array of Independent Disks (RAID) controller's role when you identify medium errors on the drives.
Note: Medium errors are also referred to as media errors
Handle HDD Medium Errors
What causes HDD medium errors?
The most common cause of medium errors is poor signal amplitude that results in
Unreliable Logical Bus Address (LBA) read location. Sometimes recoverable with multiple retries.
Transient conditions, high fly writes caused by soft particles.
Transient conditions caused by temporary shock, vibration, or acoustic events that result in off-track writes.
Poor error map function in HDD manufacture that results in padding the current primary defect locations.
How does the HDD detect the medium error?
Step 1.The HDD periodically performs Background Media Scans to detect errors.
Step 2. The HDD tries to read from the media and for some reason is unable to retrieve the data that was written.
Step 3. When the HDD is unable to retrieve data that was written it invokes the HDD recovery code which will try various error recovery steps to successfully read the data from the media.
Step 4. If all the recovery steps fail the drive will generate a 03/11/0x error back to the host and the LBA(s) will be placed on the pending defect list.
How does the Raid controller detect medium errors?
The RAID controller will encounter medium errors while Patrol Reads, Consistency Checks, Normal Reads, Rebuilds, and Read / Modify / Write operations.
Based on the RAID configuration, the controller may be able to handle the medium error reported by the HDD and no further action will be required.
In some cases, the controller will not be able to handle the medium error and will pass the error to the host to handle the error.
When does the Operating System (OS) see medium errors?
If the HDD reports a medium error and the RAID controller cannot handle the recovery, then the host will be notified of the error.
This notification is no longer just an advisory message that would inform the system that the event has occurred, it is a request for the OS to act because the HDD and RAID controller was not able to recover from the medium error.
If the OS has the required context to correctly resolve the medium error, it should be handled by the OS
If disks are in Just a Bunch Of Disk (JBOD), the OS will see errors as they are not corrected by the controller. This is common in HyperFlex (HX)/ Virtual Storage Area Network (VSAN) environments.
Grown Defects (G-list) HDD Level
While a drive is in operation, the head may come across a sector with a weakened magnetic read level. The data is still readable but may fall below the preferred threshold for qualified good sector read levels. This disk drive would consider this a sector that could and would sector spare this data to a new location available in the known good reserve list. Once the data is moved, the old sector address is added to the Grown Defects list, never to be used again. This process is a recoverable media error. The drive will give a SMART trigger once a majority of its known-good spare sectors are exhausted.
RAID Controller Role
Patrol Read is a user-definable option that performs drive reads in the background and maps out any bad areas of the drive.
Patrol Read checks for physical disk errors that could lead to drive failure. These checks usually include an attempt at corrective action. Patrol read can be enabled or disabled with automatic or manual activation.
A Patrol Read periodically verifies all sectors of physical disks that are connected to a controller, that include the system reserved area in the RAID configured drives. Patrol Read works for all RAID levels and all hot spare drives.
This process starts only when the RAID controller is idle for a defined length of time and no other background tasks are active, though it can continue to run at the same time as heavy Input/Output (I/O) processes.
You cannot conduct patrol reads on drives configured in JBOD.
Note:Latent Semantic Indexing (LSI) recommends that you leave the patrol read frequency and other patrol read settings at the default values to achieve the best system performance. If you decide to change the values, record the original default value here so you can restore them later.
Note: Patrol Read does not report on its progress as it runs. The patrol read status is reported in the event log only.
Patrol Read options are as shown in the image:
To see information about the patrol read state and the delay between patrol read runs: # MegaCli64 -AdpPR -Info -aALL
To find out the current patrol read rate, execute: # MegaCli64 -AdpGetProp PatrolReadRate -aALL
To disable automatic patrol read: # MegaCli64 -AdpPR -Dsbl -aALL
To enable automatic patrol read: #MegaCli64 -AdpPR -EnblAuto -aALL
To start a manual patrol read scan: # MegaCli64 -AdpPR -Start -aALL
To stop a patrol read scan: # MegaCli64 -AdpPR -Stop -aALL
In RAID, the Consistency Check verifies the correctness of redundant data in an array. For example, in a system with parity, to check consistency means to compute the parity of the data drives and compare the results to the contents of the parity drive.
JBOD does not support consistency check.
RAID 0 does not support consistency check.
RAID 1 uses a data compare not parity.
RAID 6 computes parity for 2 parity drives and verifies both.
Note: It is recommended you run a consistency check at least once a month.
Consistency Check management options are as shown in the image:
Consistency Check scheduling options are as shown in the image:
To see the next scheduled Consistency Check time: #MegaCli64 -AdpCcSched -Info -aALL
To disable Consistency Check: #MegaCli64 -AdpCcSched -Dsbl -aALL
Conditions when a RAID Controller Cannot Repair a Medium Error
The host OS is responsible for medium errors.
In RAID 0
There is no redundancy, so the controller cannot provide the HDD with the data to write to the LBA.
In RAID 1
When the controller cannot tell which mirror copy contains the correct data. This will only occur if both LBAs can be read, but the data does not match.
If there are 2 or more errors in the same stripe. Most likely to occur when after a rebuild of an array is initiated. The drive that is rebuilt is one error, and a medium error on any other drive rebuild would be the second error. The controller would not be able to reconstruct the data needed to rebuild the LBA on the replacement drive.
If there are 3 or more errors in the same stripe. Most likely to occur when while an array is rebuilt. The drive that is rebuilt is one error, and a medium error on any two other drives while the rebuild is in progress would be second and third errors, or a medium error and a second drive failure. The controller would not be able to reconstruct the data needed to rebuild the LBAs on the drives with the errors.