HDD compatibility and Raw read error rates on "Seagate 4TB SSHD SATA hybrid

Every block on every drive is recorded with additional check bits included in the data. When a block is read, cyclic redundancy check error correction is applied to each and every read from disk, and the error-corrected data are passed to the System, or a re-try sequence is initiated.

If the re-try fails to obtain correctable data after hundreds to thousand of re-tries, an I/O error occurs, and/or your System may halt on a kernel panic.

You are using an unusually large drive, which uses advanced (large blocks) formatting and advanced (large blocks) error correction.

It is perfectly normal to encounter correctable errors from a Hard drive, all day every day. The issue is when a non-trivial number of uncorrectable errors occurs, and you get I/O errors.

Your problems do not rise to the level of replacing the drive. Your vendor is bending over backwards to make sure you are happy.

-------

It would be prudent to Zero all data on such a drive when it is brand new, and again when it is assigned for a different purpose. This takes several hours to complete for a modest sized drive, and possibly overnight for that big drive. This process assures that every block can read and write a known pattern, and any that can not are replaced by internal spare blocks reserved for exactly this purpose.

Apple sets a very low bar (10 block substitutions) for any drive that is being Zeroed. If it does not pass, you get "initialization Failed". If a drive cannot survive that test during its warranty period, it should be returned under warranty.

> Manufacturers use different proprietary formulas for calculating error rates. Some values are logarithmic, some are hexadecimals, etc. That’s why the value is called “Raw”, because there is no method for normalizing it. This is the reason the value is insanely high for some drives. I know on Seagate drives that the Raw Read Error Rate rolls over to 0 once the count reaches about 250 million.

So just ignore it.

<

http://www.cindori.org/forums/

I ran into something similar with a few 3TB Seagate Barracudas and after some searching, came up with this discussion which sheds some light on the obscure character of SMART reporting, especially by Seagate. It doesn't look like there's a problem, and mine have continued to run for a couple of years without issue, despite those awful-looking numbers.

The Seagate rep is making a vague reference to two issues regarding drives used in Macs.

The first is a rather old requirement, no longer an issue, when Apple used to have drives with special Apple firmware, and simply did not support third-party drives (because they had not fully tested them). To use third-party drives, you had to use a third-party format program (or as some users preferred, hack the Apple software). That policy applied to SCSI drives. It has been gone since before the turn of the century.

The second is a requirement for the drives used in iMacs, where specific drives must be used because a temperature sensor output from the drive is required. The sensor must report in a certain standard way through two extra pins (like "option" jumper pins. If not hooked up, the iMac fans race at top speed to avoid possible overheating.

Neither of these applies in your case.

--------

As I tried to explain above, the Seagate rep will not see high numbers of correctable (raw) errors as an issue, because error correction is (and always has been) a part of the way Hard Drives work internally.

Seagate SAS drives for the Pro RAID card would be the only time firmware would be a likely issue today.

I got a Seagate Desktop 4TB SATA HDD (ST4000DM000) as the Time Machine backup drive for my Mac Pro. It seems working well, but similar as above, Mavericks' Disk Utility does show some "abnormal" S.M.A.R.T. attributes as shown below.

Raw Error Rate: 00000D2F5A90

Seek Error Rate: 00000005CEF2

Emergency Retract Count: 000000000197

HDD Temperature: 001400000021

Per the screenshot, the above bold attributes look interesting or even horrible, as the error rates and temperature seems extremely high or abnormal.

However, the attributes Retired Sectors Count and UltraDMA CRC Errors are still zero, which indicates the disk condition is still under control?

Is there an official or authoritative source to learn about the S.M.A.R.T. attributes? I know the Wikipedia article about S.M.A.R.T is informative but it's description regarding the above attributes is not actually helpful.

There are some online discussions like this giving a few hints but lack of supporting documents.

Could someone share an official Seagate document about these attributes?

Thanks,

B

bengmugenr wrote:

...Is there an official or authoritative source to learn about the S.M.A.R.T. attributes? I know the Wikipedia article about S.M.A.R.T is informative but it's description regarding the above attributes is not actually helpful.

Could someone share an official Seagate document about these attributes?

Thanks,

B

Check out the link in my post above. It looks like interpretations of the SMART readings are intentionally obscured and vary among manufacturers. Internally, Seagate surely knows what all their SMART readings mean but it's unlikely they'd tell users. Oddly, the only Seagate SMART value I ever found informative was "190 Airflow_Temperature_Cel" - while the RAW_VALUE was meaningless (40 (0 12 41 24)), the When_Failed column told me "In_the_past." In the past, it was still in its unventilated GoFlex enclosure, so that was no surprise. But it's been running in Bay 4 of my Mac Pro for a couple of years, and with proper ventilation, it's shown no further problems.

Thanks for the reply and the link.

Just don't understand why Seagate won't disclose the meanings of the readings but let end users see the items and their strange values. What's the point for doing that?

BTW, what do you mean by "proper ventilation" against your hard disks? Are you referring to keeping reasonable temperature for the internal drives?

Please see below the temperature readings of my Mac Pro. Any comments? How about yours?

Thanks again.

FatMac\>MacPro wrote:

... Internally, Seagate surely knows what all their SMART readings mean but it's unlikely they'd tell users. .. But it's been running in Bay 4 of my Mac Pro for a couple of years, and with proper ventilation, it's shown no further problems.

bengmugenr wrote:

...Just don't understand why Seagate won't disclose the meanings of the readings but let end users see the items and their strange values. What's the point for doing that?

BTW, what do you mean by "proper ventilation" against your hard disks? Are you referring to keeping reasonable temperature for the internal drives?

Please see below the temperature readings of my Mac Pro. Any comments?...

When Seagate introduced the GoFlex desk 3TB external USB drives, they used Barracuda XT's for the drive mechanisms. The plastic enclosures they were in were essentially sealed so that the drives would tend to overheat when used heavily. Indeed, when running a 700GB backup clone onto them and watching the transfer rate using Activity Monitor, you'd see the transfer initially go above 120MB/sec on a USB 3.0 connection, then drop to a rate in KB/sec! I called tech support about that and was assured ventilation was just fine, the drive wasn't overheating, and Activity Monitor was simply wrong and useless for that purpose (that attitude may explain why Seagate is secretive about its SMART recording codes - I bet they didn't think the SMART results on their GoFlex drives would ever be read since they were USB). At the same time, the GoFlex drives had a built-in sleep function which spun down the drive to allow it to cool when not in use. Seagate offered a utility to remove that function when it caused unintended dismount problems, but still insisted the drives' temperatures were just fine.

I eventually pulled the drives out of all three of the GoFlex externals and put one in a drive bay and the other two in an OWC fan-vented enclosure on which I beefed up the fan. The internal has been running every day for a couple of years without complaint. And that was also a cheaper way of getting Barracudas because the price of the GoFlex on sale was less than the bare drives they contained.

The temperatures your graphs show look just fine; if anything they look pretty cool 😎.