HP MSA Events Description Reference Guide (762785-001, March 2014)

66 Event descriptions
Troubleshooting steps for leftover disk drives
HP MSA 1040 and 2040 arrays use metadata on hard drives to identify vdisk members and identify other
disk members of the vdisk.
Hard drives enter a Leftover state for several reasons:
• Drive spin up was not completed before a controller polled the drive. When the controller queries the
drive and finds the drive is not in a ready state, the controller may place the drive into a Leftover state.
• Improper power-on sequences.
• Firmware upgrade (due to a timing issue).
• Failover taking longer than expected.
• The drive is swapped from another MSA system, or removed and reinserted in the array.
Metadata on a disk identifies the disk as being a member of a vdisk. Improperly clearing the metadata
from a disk may cause permanent data loss.
CAUTION: Clearing metadata from a leftover drive should be done with extreme care. Only clear
metadata without looking at logs if you are certain the drive has never been associated with a vdisk in this
MSA system or contains no data. This situation most often occurs when inserting a previously used hard
drive into a live system or moving a drive between two systems.
Never clear metadata from a drive if any vdisk in the array is in an Offline, Quarantined, or inaccessible
state. Do not clear metadata from a drive if you are unsure this is the correct step to take. Clearing
metadata from a drive permanently clears all data from the drive. In these types of situations, a backup of
data should be done if possible.
Using the trust command
The CLI trust command should only be used as a last step in a disaster recovery situation. This
command has the potential to cause permanent data loss and unstable operation of the vdisk. If a vdisk
with a single disk is in a leftover or failed condition, the trust command should never be used. The
trust command should only be used if the vdisk is in an Offline state.
If a single disk in a vdisk has failed or been placed into a Leftover state due to UREs, Smart Trip, Sense
Key, or other errors, reintegrating the disk into the same or a different vdisk has the potential to cause
data loss. A hard drive that has failed or been placed into a Leftover state due to multiple errors should be
replaced with a new hard drive. Assign the new hard drive back to the vdisk as a spare and allow
reconstruction to complete in order to return the vdisk to a fault tolerant state
The trust command attempts to resynchronize leftover disks in order to make any leftover disk an active
member of the vdisk again. The user might need to take this step when a vdisk is offline because there is
no data backup, or as a last attempt to try to recover the data on a vdisk. In this case, trust may work,
but only as long as the leftover disk continues to operate. When the "trusted" vdisk is back online, backup
all data on the vdisk and verify all data to ensure it is valid. The user then needs to delete the trusted vdisk,
create a new vdisk, and restore data from the backup to the new vdisk.
IMPORTANT: Using trust on a vdisk is only a disaster-recovery measure; the vdisk has no tolerance for
additional failures and should never be put back into a production environment.
CAUTION: Before trusting a vdisk, carefully read the cautions and procedures for using the trust
command in the CLI Reference Guide and online help.
Once the trust command has been issued on a vdisk, further troubleshooting steps may be limited
towards disaster recovery. If you are unsure of the correct action to take, contact HP Support for further
assistance.