Barracuda XT, Barracuda, Barracuda Green, DiamondMax and other desktop SATA and IDE drives

[longham]

As you can see in image, the soft alerts 3 issues, so should I need to worry about my hdd?

[fzabkar]

The author of your SMART diagnostic software is one of many who do not understand Seagate SMART attributes. S/he also does not appear to be aware that the raw values are 48-bit numbers, not 32-bit. I prefer HD Sentinel or HDDScan. HDDScan is written by a data recovery professional who is familiar with Seagate hardware and firmware.

For example, the raw value of the Seek Error Rate attribute is indicating that the drive has encountered 0 errors in 27,647,373 seeks, ie a perfect 100% score, not 51%.

See this URL for an explanation:

http://en.wikipedia.org/wiki/Talk.M.A.R.T.#Seagate_raw_Seek_Error_Rate_attribute

The Read Error Rate has a normalised value of 118 which is very good. The raw value of the attribute is a sector count, not an error count.

See this URL for an explanation:

http://forums.seagate.com/stx/board/message?board.id=ata_drives&message.id=8700

The number of Command Timeouts is not 1,179,694. If you convert the number into hexadecimal, you get 0x0012002E. AFAIK, the actual number of timeouts is 0x2E, ie 46, which is still not a good number.

The "Airflow Temperature" is actually equal to 100 - "Temperature". Its value of 722,927,659 in hex is 0x002B17002B. This represents a group of 3 temperatures, namely 0x2B, 0x17, and 0x2B. In decimal these values are 43, 23, and 43. They probably represent the max, min, and current temperature readings for the current power cycle. The temperature is not 57C -- it is 43C.

Attribrute 194 is strange. I don't understand why the raw value is 61C.

The raw value of Hardware ECC Recovered also appears to be a sector count, not an error count.

It appears to me that the only attribute that is of concern is the Command Timeouts.

Here are a few alternative SMART diagnostic and benchmarking tools.

HD Sentinel (DOS / Windows/Linux):

http://www.hdsentinel.com/

HDDScan for Windows:

http://hddscan.com/

HD Tune:

http://www.hdtune.com/download.html

CrystalDiskMark:

http://crystalmark.info/software/CrystalDiskMark/index-e.html

smartmontools (Linux/Windows):

http://sourceforge.net/projects/smartmontools/files/

http://sourceforge.net/apps/trac/smartmontools/wiki/Download

See this article for SMART info:

http://en.wikipedia.org/wiki/S.M.A.R.T.

[longham]

As u suggest, I try HD Sentinel, and I'm so scared now:

Is it true that my HDD will be dead in 269 days???

[fzabkar]

I don't know what to think. The author appears to believe that a spin retry count of 22 is serious enough to downgrade the drive's reliability prediction. Maybe you should email him with your results.

BTW, I believe the number of seek errors is probably 6 or 7, not zero as I previously stated. Ashampoo does not report the uppermost 16-bits of the SER attribute. You will probably see them in HD Sentinel's SMART tab. In any case, 7 errors out of 27 million seeks is nothing to be worried about, I would think.

The attributes that are of most concern to most people are reallocated sectors and pending sectors. Your drive is clean in that respect.

[longham]

Here is the SMART result in Sentinel:

Just as u guessed, it is true that the real number of seek error is 6 .

But I don't have to worry about Spin Retry Count, do I? Is hard disk mechanical subsystem ok? When using PC, sometime I hear the sound from the HDD. And is it normal that one hdd goes off when it is idle in long time (It is the first time I have 2 hdds in PC)? I'm so happy when u answer my thread... Thank you very much.

Message Edited by longham on 12-14-2009 12:54 AM

[fzabkar]

Sorry, I don't know how to answer your concerns. The SMART reports don't always tell everything. For example, if your drive is having difficulty reading a particular sector, you may notice freezing or hesitation as it retries several times. If it eventually reads the sector successfully, then the SMART report may not show the problem. I've also seen cases where a drive will "disappear" and then not reappear until it is power cycled. AFAICS, if the drive's MCU has hung, then it will not be able to log the event that caused it to hang. Catch 22.

I think that all you can say for certain is that Ashampoo's author got it wrong. As for the significance of the Spin Retry Count, I would think that this could be caused by a tight spindle motor bearing, in which case you would hear bearing noise (?). Alternatively, it may be the result of stiction, although this is much less of a problem in modern drives. It could be that the drive is having difficulty generating the motor torque required to break any adhesion between the heads and the platters.

See this Wikipedia article:

http://en.wikipedia.org/wiki/Stiction#Hard_disk_drives

If your HD is powering down after a certain idle time limit has been exceeded, then this could be the result of Power Management settings within the OS, or it could be the drive's own internal APM (Advanced Power Management) setings.

The one attribute that would interest me would be Command Timeouts. I don't understand exactly what it means, but I suspect it *may* reflect freezes and pauses during reading and writing. But that's only a guess.

[longham]

I suspect it *may* reflect freezes and pauses during reading and writing. But that's only a guess.

I have done surface test with hddscan and i saw the more time, the lower performance in diagram. Was that normal? I will run test again and upload screenshot for you later. Can full data in the disk make low performance?

[longham]

Is this normal? (Surface Read Test in HDD Scan)

P/S: My HDD is full with data, is that the cause? Or is it hardware reason?

[atx12v]

couldn't it just be from a cheap power supply

[fzabkar]

I can't explain the blips in the curve, but overall the performance appears normal.

According to page 10 of the following document (page 16 of PDF), the maximum sustained data transfer rate is 105 MBytes/sec. That's what your graph is showing at LBA 0.

Barracuda 7200.11 Serial ATA Product Manual, Rev. E:

http://www.seagate.com/staticfiles/support/disc/manuals/desktop/Barracuda%207200.11/100452348e.pdf

The maximum performance occurs at the outer cylinders where there are more bits per track. The performance decreases as you move towards the spindle, at which point the bits per track is at the minimum.

There are about 20 or 30 flat steps in the curve. These correspond to the "zones". See the following articles for an explanation.

http://en.wikipedia.org/wiki/Zone_bit_recording

HDD from inside: Tracks and Zones:

http://hddscan.com/doc/HDD_Tracks_and_Zones.html

You can compare your own test results with these HD Tune results:

http://www.hdtune.com/testresults.html