HPE Issues Patch to Stop SSDs From Bugging Out After 40,000 Hours

By Niels Broekhuijsen
published

A bug in the firmware of a handful of HPE's enterprise SAS SSDs will cause them to brick.

(Image credit: Michael Vi/Shutterstock)

Hewlett Packard Enterprise (HPE) sent out a warning (opens in new tab) Friday that some of its enterprise SAS-SSDs will brick themselves after 40,000 hours of operation. Based on the sale dates of the afflicted parts, most are expected to reach the 40,000-hour mark in October 2020 at the earliest.

To fix the problem, system administrators must install the HPD7 (or newer) firmware (opens in new tab)update. This update is considered critical, and failing to install it will result in unrecoverable data loss.

HPE did not mention in its bulletin what the exact cause of the problem is. Some might suspect that HPE implemented a timer-based obsolescence at such a non-random figure, but that would simply be too obvious. The issue is said to have been reported to HPE by one of its unnamed SSD manufacturers.

What's Affected

HPE ProLiant, Synergy, Apollo 4200, Synergy Storage Modules, D3000 Storage Enclosure and StoreEasy 1000 Storage are systems that may contain SSDs with this bug.

Swipe to scroll horizontally
HPE Model NumberHPE SKUHPE SKU DESCRIPTION
EK0800JVYPN846430-B21HPE 800GB 12G SAS WI-1 SFF SC SSD
EO1600JVYPP846432-B21HPE 1.6TB 12G SAS WI-1 SFF SC SSD
MK0800JVYPQ846434-B21HPE 800GB 12G SAS MU-1 SFF SC SSD
MO1600JVYPR846436-B21HPE 1.6TB 12G SAS MU-1 SFF SC SSD

Given that many SSDs in an enterprise environment might be running in RAID arrays, it is especially critical to apply the update: many of the SSDs will be failing simultaneously, and without a non-RAID backup, all data will be unrecoverable surpassing 40,000 hours.

In November, HPE issued a similar update (opens in new tab) for a different set of SSDs that would fail after 32,768 hours of operation.

Niels Broekhuijsen
Niels Broekhuijsen

Niels Broekhuijsen is a Contributing Writer for Tom's Hardware US. He reviews cases, water cooling and pc builds.

3 Comments Comment from the forums
  • alextheblue
    Somewhere out there, an unpatched drive is going to suffer unrecoverable data failure come October.
    Reply
  • USAFRet
    alextheblue said:
    Somewhere out there, an unpatched drive is going to suffer unrecoverable data failure come October.
    Yep.
    And someone will bitch about it.
    "They shoulda told us! Imma sue!"

    They DID tell you, and "enterprise SAS SSD " means you have a backup, right?
    Reply
  • fzabkar
    This bug raises some awkward questions. For example, why was a time counter set for 40K power-on hours? What did the designer intend to happen at this milestone? Was the SSD supposed to go into some kind of "end-of-life" mode?

    I have examined these firmware patches here:
    http://forum.hddguru.com/viewtopic.php?f=3&t=39964
    AFAICT, both the Dell and HP SSDs are OEM versions of SanDisk's Lightning Eco Gen. II SAS SSDs. WD/SanDisk has issued its own patch, but it's not easy to find. In fact when I examined the payload, I discovered numerous references to Huawei. This led me to wonder whether Huawei was the actual manufacturer, or perhaps Huawei was an OEM and WD mistakenly released Huawei OEM firmware into its retail channel.

    Even stranger is the 32768 hour bug which affects HP OEM SSDs but not Dell or Samsung (who is the original manufacturer):

    https://support.hpe.com/hpesc/public/docDisplay?docId=emr_na-a00092491en_us
    Here is a user who has lost 4 x Samsung PM1633a 15.36TB SSDs all at once:

    datarecovery/comments/govejbView: https://www.reddit.com/r/datarecovery/comments/govejb/are_these_sas_ssds_recoverable_showing_ubad_and_0

    (How do I stop this forum from expanding the reddit URL? <angry>)

    I suspect that Samsung's retail versions of the same drive are also affected by the same bug, but Google turns up nothing.
    The SSD in the reddit thread appears to be an OEM HP model. However, it's not clear whether this model number was one of those listed in HP's bulletin.
    Reply