Is this SMART data possible?

[patkim]

I have one old 80GB legacy SATA HDD over 5 years old. I occasionally run the CrystalDiskInfo on Windows. The results are astonishing, specifically the raw values for Current Pending & Uncorrectable Sector count. It reports these as Hex FFFFFFFF that's whooping 4294967295 in Decimal, way way way more than what any 80GB legacy HDD can have as max number of sectors....yet Reallocated Sector count is just too good!!

Could there be some explanation for so high raw values reported?
Thanks.

[guru]

Impossible> Seagate do not support such large G list table

[patkim]

Is the SMART data corrupted or could it simply be some bug in such reporting tools? It's quite confusing for me to interpret this result.

[guru]

Use another software application

[fzabkar]

All tools will report the same 512 bytes of data. However, some might decide to reduce the raw value to 16 bits rather than 32 when they interpret the data. HD Tune reports the value as -1. CrystalDiskInfo actually dumps the raw 512-byte block, as in this example:

http://www.tomshardware.com/answers/id- ... count.html

AISI, it's some kind of bug, or some kind of signal from the drive to flag an internal problem.

I found the following SMART reports, some of which have the same maxed out value for those same two attributes. One of the drives reports "97 97 0 74", which would suggest that the threshold for the attribute is reached when the number of pending sectors is between 1850 (74 / 4 x 100) and 2467 (74 / 3 x 100). Similary, the Reallocated Sector Count (95 95 36 220) would suggest that the threshold corresponds to a number in the range from 2347 (220/6 x 64) to 2816 (220/5 x 64).

http://webcache.googleusercontent.com/s ... 01024.html

HDAT2 v5.0 Kod:
Name Threshold Value Worst RAW
1 Read error rate 6 12 112
5 Reallocated sectors count 36 100 100 0
*187 Uncorrectable Error Count 0 89 89 11
!197 Current pending sectors 0 1 1 4294967295
!198 Offline scan UNC sectors 0 1 1 4294967295

Victoria 4.46b pod XP Kod:
ID Name Value Worst Tresh Raw Health
-------------------------------------------------------------------------
1 Raw read error rate 114 92 6 65311696 •••••
5 Reallocated sector count 100 100 36 0 •••••
187 Reported UNC error 89 89 0 11 ••••
195 Hardware ECC recovered 85 63 0 159050368 ••••
197 Current pending sectors 1 1 0 65535 •
198 Offline scan UNC sectors 1 1 0 65535 •

http://www.tomshardware.com/forum/27249 ... 200-failed

001 Raw Read Error Rate 112 099 00000002AF-57D3 006
005 Reallocation Sector Count 098 098 0000000000-0071 036
187 Reported Uncorrectable Error 001 001 0000000000-21B1 000
195 Error Rate 120 055 0000000E70-8240 000
197 Current Pending Errors Count 099 006 0000000000-001F 000
198 Uncorrectable Errors Count 099 006 0000000000-001F 000

http://webcache.googleusercontent.com/s ... rakes.html

HD Tune: ST3320620AS Health
ID Current Worst ThresholdData Status
(01) Raw Read Error Rate 117 65 6 156835423 Ok
(03) Spin Up Time 96 95 0 0 Ok
(04) Start/Stop Count 100 100 20 721 Ok
(05) Reallocated Sector Count 95 95 36 220 Ok
(07) Seek Error Rate 84 60 30 241371008 Ok
(09) Power On Hours Count 93 93 0 6283 Ok
(0A) Spin Retry Count 100 100 97 0 Ok
(0C) Power Cycle Count 100 100 20 714 Ok
(BB) (unknown attribute) 1 1 0 120 Ok
(BD) (unknown attribute) 100 100 0 0 Ok
(BE) Airflow Temperature 54 47 45 790954030 Ok
(C2) Temperature 46 53 0 46 Ok
(C3) Hardware ECC Recovered 57 45 0 32279873 Ok
(C5) Current Pending Sector 97 97 0 74 Ok
(C6) Offline Uncorrectable 97 97 0 74 Ok

http://webcache.googleusercontent.com/s ... page2.html

HD Tune Pro: ST380215A Health

ID Current Worst Threshold Data Status
(01) Raw Read Error Rate 112 100 6 43563204 Ok
(05) Reallocated Sector Count 100 100 36 0 Ok
(BB) Reported UNC error 13 13 0 87 Ok
(C3) Hardware ECC Recovered 82 66 0 91764525 Ok
(C5) Current Pending Sector 1 1 0 -1 Ok
(C6) Offline Uncorrectable 1 1 0 -1 Ok

[patkim]

Thanks for such a detailed explanation! I tried with HD Tune as well and it reports these as –1.
Apparently Seagate’s Seatools for windows reports SMART test outcome as PASS. It does not report actual SMART data as such.

[fzabkar]

SeaTools is useless as a SMART reporting tool. It only fails an attribute if its normalised value hits the threshold. In this case it appears that the two problematic attributes are not permitted to drop below 1, so they will never trigger a SMART failure.

[guru]

http://hddscan.com

Have you tried it?

[fzabkar]

http://hddscan.com

Have you tried it?

Why would that make any difference, or are you suggesting that the OP scan the drive?

[fzabkar]

Scaning the drive with Remap on or erase delay might be a good idea as the problematic sector will either be re-written and the ECC will be ok or will be considered "bad" and re-located (added to G-List) so the attributes "Current Pending Sector Count" should change.

That makes sense, but simply using a different tool and expecting it to report different values for those same SMART attributes would be absurd.

[guru]

indeed

[guru]

Seagate probably deems those attributes (197 ( 0xC5 )/198 ( 0xC6 )to be unreliable predictive indicators of drive failure.

SeaTools is useless as a SMART reporting tool. It only fails an attribute if its normalised value hits the threshold. In this case it appears that the two problematic attributes are not permitted to drop below 1, so they will never trigger a SMART failure.

[fzabkar]

Here is just one example of why SeaTools' SMART reporting is at best useless, and at worst deceptive:

http://www.tomshardware.com/answers/id- ... se-on-page

Notice that if that drive had 4687 seek errors, it would have passed with no warning that anything is wrong, and a surface scan would have turned up nothing.

[guru]

Guess they want to "Limit" RMA's

[data-medics]

As far as I care the only purpose for SMART to exist is just to keep track of whether the drive is reallocating sectors. The rest of the SMART data usually just reports an error after the drive has already failed (which isn't so "smart" in my opinon).

[fzabkar]

I'm not sure whether their conclusions would be valid for today's drives, but Google published a study in 2007 that evaluated the usefulness of SMART as a failure predictor in their server farms.

http://static.googleusercontent.com/med ... ilures.pdf

In reading their paper, it's clear that the authors did not understand the meanings of several of the SMART attributes.

For example ...

When examining our population, we find that seek errors are widespread within drives of one manufacturer only, while others are more conservative in showing this kind of errors. For this one manufacturer, the trend in seek errors is not clear, changing from one vintage to another. For other manufacturers, there is no correlation between failure rates and seek errors.

Obviously they were referring to Seagate.

[patkim]

Scaning the drive with Remap on or erase delay might be a good idea

would like to know if there's any tool that can do this kind of scan. Thanks.

[patkim]

I have scanned the same drive with MHDD32 and here's the outcome.