Intel 320 SSD 8MB Data Recovery

Updated: 1/14/2019

The Intel SSD 8MB bug began appearing in 2011, with Intel confirming the firmware bug’s “sightings” around August of that year. Since its first appearance, the 8MB bug has vexed countless SSD users. This bug affects Intel’s 320 Series SSDs. Not only would it make the SSD show up as unallocated—but to add insult to injury, it apparently reduces the SSD’s useable capacity to a measly eight megabytes. (Victims of the bug may have said to themselves at one point, “I knew SSDs had a lot of catching-up to do with HDDs capacity-wise… but this is ridiculous!”) The Intel SSD 8MB bug is often accompanied by a BAD_CTX error.

Intel was able to swiftly provide a firmware fix to get afflicted SSDs functional again (and with their full capacity) once their engineers had confirmed the 8MB bug’s existence for themselves. But what about the unlucky users’ critical data? Fortunately for them, Gillware has recovered data from plenty of Intel Series 320 SSDs afflicted with the 8MB bug.

In this data recovery case, the client worked for a seismological research program. Seismic monitors helped them collect data from the seafloor, each monitor a custom computer with an Intel 320 series SSD inside. Unlike the data recovery cases we’ve received due to the recent flooding in Louisiana, these computers were designed to be submerged. True to their design, they held up under miles of water pressure. The real failure came not from water damage, but from an all-too-common firmware bug.

Squashing the Intel SSD 8MB Bug

Just like traditional spinning disk hard drives, SSDs rely on complex firmware to manage the flow of data in and out of their NAND flash memory chips. When the 320 Series goes bad, an interesting quirk manifests itself in the firmware. When asked to report its capacity, the SSD responds by listing itself as having a whopping grand total of eight megabytes of storage space in its NAND chips. (For reference, the ancient ST-506 had a five-megabyte capacity—talk about a blast from the past!)

The 8MB bug represents a quirk of the way Intel 320 Series drives react to failure. The 8MB bug makes the serial number of the device appear as “BAD_CTX”. Other error messages define the exact failure mode of the drive. For example, error code 000012B indicates a failure of the flash translation layer.

The Intel 320 Series line of SSDs isn’t the only storage device that can misrepresent its capacity. Traditional hard disk drives will sometimes detect with a capacity of zero, or some patently absurd number (like a 500 GB hard drive saying it has 3 terabytes of disk space). This usually happens when the severely damaged read/write heads cannot properly read the firmware.

Since the Intel SSD 8MB bug first appeared, our engineers have gotten data recovery from affected Intel 320 Series SSDs down to an exact science. These data recovery cases usually tend to be on the “easier” end of the sliding scale of SSD data recovery difficulty for our engineers. And so, of course, in this data recovery case, the universe had to throw us a curveball. Our R&D director and resident flash memory expert Greg Andrezejewski noticed that this firmware bug indicated there was a problem communicating with the NAND memory chips themselves.

How to Talk to NAND Chips

Your SSD only knows one way to talk to the NAND chips that hold your data. And so if that way doesn’t work out, the device fails. Communicating with flash memory chips is a bit like commuting to work. Except if the road the drive take to get to the office is closed, the drive will just sit there, stymied. Our engineers, on the other hand, know that they can just take a detour to get to work.

There are a lot of data recovery cases with SSDs that can be solved without imaging each NAND chip separately and then stringing the individual chip images together in the proper order using custom software. In general, typical 8MB bug data recovery cases tend not to require such extreme measures. This particular case of the Intel SSD 8MB bug was far from typical.

Dumping the raw flash memory and rearranging it properly is a lot of hard work. It’s a bit like assembling a massive jigsaw puzzle. Only all the pieces are nearly the same shape. Putting the NAND chips together properly takes a kind of digital sleuthing and reasoning that would probably give Sherlock Holmes an aneurysm (it also takes custom software to link the chips’ contents together properly, which would also probably give Holmes an aneurysm—only for a completely different reason).

Intel SSD 8MB Bug Data Recovery Case Study: 320 Series

Drive Model: Intel SSD 320 Series SSDSA2NW160G3

Drive Capacity: 160 GB

Operating/File System: Linus Ext3

Data Loss Situation: Intel SSD 8MB bug affected SSD in undersea seismic monitor

Type of Data Recovered: Seismic oceanography data

Binary Read: 100%

Gillware Data Recovery Case Rating: 9


Raw NAND dumps are hardly ideal in the world of data recovery. But when they must be done, they must be done. And our expert SSD data recovery technicians like Greg Andrezejewskiare more than up to the challenge.

After piecing the data together, Greg had a full, correctly-arranged image of the failed Intel 320 Series SSD. After quite a bit of wrangling, we fully recovered the vast majority of the client’s lost seismic data with little file corruption. We rated this Intel SSD 8MB bug data recovery case a 9 out of 10 on our case rating scale.

At Gillware, our SSD data recovery engineers see the Intel SSD 8MB bug fairly frequently. In most cases, data recovery from these misbehaving SSDs is not quite as involved of a process as it was in this case study. The Intel 8MB bug is certainly an interesting one as far as firmware bugs go, mainly because of its shocking symptoms (although if your Intel SSD has fallen victim to it, it may not seem quite so interesting to you). If you’ve lost data because your SSD decided to go back to the stone age, the SSD recovery experts at Gillware are more than capable of helping you out.

Will Ascenzo
Will Ascenzo

Will is the lead blogger, copywriter, and copy editor for Gillware Data Recovery and Digital Forensics, and a staunch advocate against the abuse of innocent semicolons.

Articles: 213