SSD freezing with event ID 153 (SRB_STATUS_ERROR)

Question

My 4 years old OCZ Agility 4 is hiccuping after being moved to a different system. This SSD was my main drive until a month ago and I first noticed the "Event ID 153 - The IO Operation at logical block address xxxx for Disk y was retried" 1 year ago.

I'd just get the sporadic warning, one every couple of hours, and no lag associated. The event detail specified either the 00 09 28 or 00 12 28 codes. Based on these articles (1, 2) they mean:

0x00 = SCSISTAT_GOOD

0x09 = SRB_STATUS_TIMEOUT    or    0x12 = SRB_STATUS_DATA_OVERRUN

0x28 = SCSIOP_READ

Nothing alarming, and SMART data was good. It might have nothing to do with my current issue, as I do get the very occasional data_overrun on my brand new SSD.

After I upgraded the computer I moved the OCZ to a laptop. I formatted it and performed a clean install. I've been experiencing freezes since then.

The issue: Whatever program I'm using will become unresponsive for a couple of seconds, but I'll still be able to move the cursor around and even alt+tab. Whatever program becomes active will become unresponsive as well, and that lasts until the SSD responds in 2s-10s. I don't get a single event 153 in hours, I get ~10 of these at once during the freezing incident, and that happens multiple times an hour. If I happen to be running a SSD benchmark at the time it'll output either 0MB/s or fail.

When the drive doesn't hang the benchmark runs smoothly.

I don't get the event 153 if I'm using an external adapter to plug that SSD to an USB port. I wasn't able to determine if Windows just doesn't report that event if the SSD is connected to the USB (the event 153 was only introduced in Win8, it might not support USB data) or if the issue doesn't happen at all.

The SMART data doesn't look too bad to my untrained eyes:

The 153 event details code for these incidents are 00 04 2A and 00 04 28:

0x00 = SCSISTAT_GOOD

0x04 = SRB_STATUS_ERROR

0x28 = SCSIOP_READ    or    0x2A = SCSIOP_WRITE

The SBR error happens when the HBA returns a nonspecific bus error.

Could that be caused by a bad SATA adapter? I didn't try a different adapter yet because the model is difficult to find. If it's not the adapter, what can I do to mitigate the issue?

Already tried:

• Update the SSD firmware. It's already up to date.
• Increased the over provisioning.
• Clean OS install. Updated drivers as possible. It's an old laptop (LGP430) and I'm having trouble finding the best drivers for it.
• Made sure it's using AHCI mode.
• Double-checked TRIM. It's enabled.
• Installed Intel Rapid Storage, based on reports that it was able to stabilize SSDs with similar issues. Only older versions support this laptop specs, and installing it was a bad call. Windows crashed so badly I was forced to ntfsfix and mount the partition on Ubuntu so I could run a chkdsk then perform a system restore.

Solution

You might be a victim of the Intel Cougar Point (6-series) chipset SATA bug.

As far as I can tell, your system uses the HM65 chipset. This, and other 6-series chipsets, were affected by a problem in the SATA 3 Gb/s ports (ports 2 through 5) where a transistor is driven with excessive voltage, resulting in premature failure. The 6 Gb/s ports (0 and 1) are unaffected. See also: How to avoid purchasing Faulty "Cougar Point" Chipset motherboard?

If your SSD is connected to one of the 3 Gb/s ports (and I can't quite tell because CrystalDiskInfo indicates 6 Gb/s while benchmarks suggest 3 Gb/s), then you might be affected by this bug. This would probably be the case if the SSD is installed as a secondary drive on the system. Unfortunately, this means that the 3 Gb/s ports are unusable, leaving you with just the first two SATA ports to work with.

If the drive is in fact connected to a 6 Gb/s port, the issue may be elsewhere on the motherboard, though chipset failure cannot be ruled out.

Comments

[misha256]

So the drive was showing errors on Computer 1 and now it's showing freezes/errors on Computer 2 (your laptop). Two computers, one drive... issues on both. Chances are very high that the drive is faulty.

[Rhaenys]

@misha256 No, I was getting timeouts or data_overruns on computer 1. I only get errors in the laptop (computer 2). P.s. for your edit: It did cross my mind that the drive is faulty, but the fact I also got an overrun for a completely different SSD on computer 1 makes me think this computer's events can be a coincidence—just Windows being Windows.

[bwDraco]

My first guess was the Cougar Point SATA bug but it looks like the drive is plugged into a 6 Gb/s port. On your laptop, is the drive indeed plugged into a 6 Gb/s port (ports 0 and 1) or is it on a 3 Gb/s port (ports 2 through 5)? Despite CrystalDiskInfo suggesting that it's on 6 Gb/s, the benchmarks you're showing suggest that it's running at 3 Gb/s. Were these screenshots taken on different computers?

[bwDraco]

Open the BIOS setup on the computer where the problems are occurring. Does it say which port the SSD is connected to?

[Rhaenys]

@bwDraco The BIOS doesn't tell which port it's using. Makes sense, since it's a laptop and people don't tend to take them apart and mess with SATA. I did a sudo lshw on Ubuntu and it listed only 2 SATA ports, the SSD being in the first (CD-ROM in the second). It's not the surest way to get the SATA number, but it tends to list devices in the right order.

[bwDraco]

Hmm... 2 SATA ports. Is there more than one drive bay in the machine?

[Rhaenys]

I honestly can't explain the discrepancy between reported speed and benchmark speed. The SSD is the only disk in the machine.

[Rhaenys]

You gave me an idea: I'll try to take the laptop further apart and see if there's a way to swap the CD-ROM and disk. I'll let you know if it works. In any case thanks for the answer, it's very detailed and might be my issue.