HP/HPE INTERNAL USE ONLY

Analysis Code: 63, 64 and 65

Severity: Error

Both analysis codes, 63 and 64, pertain to the SanDisk OPTIMUS model SSD disk drives in particular the ECO and ASCEND models.

While both these models may present the same error the cause of the error is different between the two (2) models.

The following CFIs reference this analysis code:

CFI: 5872

Reference SAW: http://h41302.www4.hp.com/km/saw/view.do?docId=emr_na-c04780380

Also, the following information will be useful:

BZ: 142874

OPTIMUS Drive and "Media Format Corruption

Background:

The Following model numbers are impacted:

OPTIMUS Eco:

"DOPE0480S5xnNMRI" SmrtStor

"DOPE1920S5xnNMRI" SmrtStor

Fix to problem in 3P02 and Higher (3P04 latest).

OPTIMUS Ascend:

"DOPA0480S5xnNMRI" SanDisk

"DOPA0920S5xnNMRI" SanDisk

Fix to problem in 3P03 and higher.

SandDisk acquired SmartStor, so both models are now SanDisk drives.

Both of these product lines can present with the same error/symptom: "Media Corrupted Format",

however the root causes are different.

Problem Description:

ASCEND (DOPA):

The cause of the "Media Format Corrupted" issue with ASCEND drive models is a result of the drive

having been downgraded at some point. If one of these drives is downgraded from the newer versions

of firmware, the older firmware does not understand the changes in the newer firmware.

Firmware 3P02 and higher use different data patterns for read interactions with different types of blocks.

Also the LBA range was increased for the unmap function.

Fix:

Upgrade to 3P03. Due to the changes in the firmware, which are not on the physical drive itself,

the actual process of upgrading the firmware takes into account the changes between versions of

firmware.

Thus, no further action beyond the upgrade is required.

When to reformat:

1. After a drive has been down graded or there suspicion that the drive may have been downgraded

at some point prior to the current upgrade.

2. When the drive encounters the "Media Format Corrupted" error.

ECO (DOPE)

Firmware prior to 3P02 (current is 3P04) could encounter a problem due to the timing of a drive reset,

where an entry in the L2P table would be marked as stale. And if later an IO tries to access the stale

entry in the L2P table we encounter and receive the "Media Format Corrupted" error. Because this involves

the L2P table, it involves meta-data at a physical drive level.

Fix:

Upgrade to 3P04 or higher. Reformatting may be necessary.

When to reformat:

1. If the drives being upgraded are drives currently at firmware revision 3P01 or earlier.

2. When the drive encounters the "Media Format Corrupted" error.

3. If these are new/replacement drives even at 3P04 and they are *NOT* from HP directly.

We do not know the state of the drives from 3rd party supplies/OEMs or from our

logistics.

4. If the customer had ever downgraded the drive.

Only for ECO (DOPE) PDs

SanDisk Failure analysis report states " When 3P02 or later firmware is loaded, the drive must be

formatted to ensure that earlier un-formats/power cycle conditions did not leave page count mismatches in LBA’s causing a format corrupt issue."

Reformatting:

For existing drives that may have data/chunklets allocated to the drive in question:

1. Vacate the drive. Make sure the you use the "-perm" switch because we do not know

what PD ID the OS will assign when the drive is re-admitted. You need to also remove all the spares.

2. Retreive the WWN of the drive to be reformatted.

3. Upgrade firmware if necessary, (upgradepd -w <WWN>)

4. Dismiss the drive.

5. Reformat the drive. (use: controlpd format 520 <wwn_of_disk>)

6. re-admit the drive. You need to make sure that spares are also allocated back onto the drive.

7. Execute a tune to help move data back.

For replacement drives:

1. Use the noautosmag touch file to disable auto admit with the SMAG oprations. #onallnodes 'touch /var/opt/tpd/touchfiles/noautosmag'

2. use showpd -i to find the drives WWN.

3. Upgrade the firmware. (upgradepd -w <WWN>)

4. Reformat the drive. (use: controlpd format 520 <wwn_of_disk>)

5. Remove the aforementioned touch file.

6. Resume the SMAG operation.

For new drives:

1. DO not admit the drives. (use system variable mvar outlined below)

2. use showpd -i to find the drives WWN.

3. Upgrade the firmware. (upgradepd -w <WWN>)

4. Reformat the drive. (use: controlpd format 520 <wwn_of_disk>)

5. admit the drives.