HDD GURU FORUMS

As the subject says, I have a HDD which reports 65527 reallocated sectors when I run the SMART ATT on the current version of Magic boot MHDD.
Usually, I can safely assume any number over 0 means the drive has, or had a bad sector that was remapped, and usually recycle it as "faulty". Running a scan also usually reveals current bad sectors unremapped.
On this drive, I found no current bad sectors, and a quick Google search of "65527 reallocated sectors" revealed some results, some of which suggested that number is a bug, but also still faulty.

Has anyone had this many sectors been reported as such?

Thanks

workshop wrote:On this drive, I found no current bad sectors

IMHO that's a big clue, although it's just one of many possible tests which could be done.

workshop wrote:a quick Google search of "65527 reallocated sectors" revealed some results, some of which suggested that number is a bug, but also still faulty.

To avoid me wasting time redoing the same detailed search as you did, please can you provide a link to a webpage saying this. After a quick check, I did not find one saying "bug" and "drive faulty".

workshop wrote:Has anyone had this many sectors been reported as such?

I have not seen such a high raw value for that attribute, be "real".

Please supply the full set of SMART data (raw and "cooked" values) from the drive for further review.

Well the customer that uses the laptop has been reporting slowness on this laptop, along with HDD sounds (clunking and clicking as it's usually described), which I haven't noticed yet. This led me to try HDD surface scan as the first test.

Quote from http://ubuntuforums.org/archive/index.p ... 99556.html (I went onto the site and searched for 65527)

Same problem here.

Normalized: 100
Worst: 100
Threshold: 24
Value: 65527 sectors

My value is somewhat higher, so I am wondering if I have indeed a failing harddrive, of that it is a bug like suggested above.

My harddrive is about 4 to 5 years old, so I could imagine that some hardware is indeed failing... No critical data though, just experimenting with Ubuntu a bit.

Can someone tell me how I can check my harddrive with another tool than the standard Disk Utility? I am a Linux noob, so please keep it simple

Thanks in advance,
Mark

Now, I know it's from a Ubuntu support site, but I was just searching and that was the first thing I came across. My point was that for some a random number 65527, to find someone mentioning the same on another post was fairly unlikely, unless it was a known/common SMART error/bug

Having just checked the SMART atts again, it is 65537, not 65527, but on a similar ubuntu link is a poster with the same number http://ubuntuforums.org/archive/index.p ... 05565.html

Under Reallocated sector count I got this result:


Normalized: 100
Worst: 100
Threshold: 24
Value: 65537 sectors [url]


Clearly a bogus warning caused by the same bug as everyone else here got (except SpearZ, who's drive probably really is about to fail). Relax, Your drive is fine, but that doesnt mean you shouldnt make that backup anyhow

I will try and get the full read out of the SMART att list.

I have posted a reply, but it needs to be checked, (I assume because it has several links in it)
I have written down the list of SMART atts, but I used abbreviations rather than writing the full list out, if I can find a list on the documentation or a post, I will copy that and change the numbers. When I have chance

Previously couldn't get into Windows (had no password) but finally got in and installed HD Tune, which gives the same SMART results.

HD Tune Pro: Hitachi HTS541616J9SA00 Health

ID Current Worst ThresholdData Status
(01) Raw Read Error Rate 100 100 62 0 ok
(02) Throughput Performance 100 100 40 0 ok
(03) Spin Up Time 238 100 33 1 ok
(04) Start/Stop Count 99 99 0 1851 ok
(05) Reallocated Sector Count 100 100 5 65537 warning
(07) Seek Error Rate 100 100 67 0 ok
(08) Seek Time Performance 100 100 40 0 ok
(09) Power On Hours Count 91 91 0 4006 ok
(0A) Spin Retry Count 100 100 60 0 ok
(0C) Power Cycle Count 100 100 0 1388 ok
(BF) G-sense Error Rate 100 97 0 0 ok
(C0) Unsafe Shutdown Count 100 100 0 61 ok
(C1) Load Cycle Count 77 77 0 237027 ok
(C2) Temperature 141 100 0 917543 ok
(C4) Reallocated Event Count 100 100 0 1 warning
(C5) Current Pending Sector 100 100 0 0 ok
(C6) Offline Uncorrectable 100 100 0 0 ok
(C7) Ultra DMA CRC Error Count 200 253 0 0 ok
(DF) Load/Unload Retry Count 100 100 0 0 ok

Health Status : warning

workshop wrote:I have posted a reply, but it needs to be checked, (I assume because it has several links in it)

Yes, it'll be held in moderation to check for spam links. That post will probably appear somewhere above mine, as it'll be timestamped with the time it was originally posted.

Thanks for that SMART data. Despite that mention of "warning" from HD Tune pro for attribute 5, I believe it is being fooled by the raw value which you mentioned of 65527. Look at the "cooked" value - it is still at 100.

Drives could still show a "cooked" value of 100 for attribute 5 even with a very small number of reallocated sectors. That could be the situation here, since attribute C5 (hex) has a value of 1, so it is possible that the drive has 1 (G list) reallocated sector in reality (although for complex reasons it could be slightly higher, due to what attribute C5 actually measures,). Attribute C6 (hex) being 0 is a good sign.

So in summary, I believe the raw value shown here for attribute 5 is either a bug or an unusual, but deliberate decision by the drive f/w (which is not prohibited by the spec), but either way, does not seem to be the actual number of reallocated sectors (due to the cooked value being 100). The drive might have 1 (or a few) reallocated sectors, but whether such a drive can be termed "faulty" or not, depends on the definition which is being used of a "fault".

It'll be interesting to see those links (when they appear), to see the other discussions about this.

FYI, if we believe the raw value (!), the load cycle count is high, relative to the number of power-on hours (averaging one load cycle per minute, by my calculation) - that probably indicates that the drive's idle timeout is set too low, although a f/w bug here is also possible. You might want to investigate this further.

Your previous post has now appeared, so I'll read the links in a while - I didn't find those since I included the quotes in the Google search which you mentioned

Now that I see your customer has been talking about the drive clicking and slowness, I suspect that may be related to the load cycle count indication which I mentioned.

Thanks for the responses.
I think I can safely say that the drive isn't in a 100% perfect condition (although how can we say any drive is ), so we'll likely replace it anyway, in an attempt to resolve the problem.

Sorry about the confusion with the quotes on the Google search

workshop wrote:Thanks for the responses.
I think I can safely say that the drive isn't in a 100% perfect condition

Agreed, it's not perfect, but it's also not the most faulty drive that I've seen today It depends what level of confidence the customer is looking for...

workshop wrote:we'll likely replace it anyway, in an attempt to resolve the problem.

Understood - proactive replacement can be a good approach, if cost / justification is not an issue. My policy is always to give new drives a very thorough test first, before trusting them with real data.

workshop wrote:Sorry about the confusion with the quotes on the Google search

No problem

P.S. Perhaps it would be worth monitoring the load cycle count after replacing the drive, to see if it climbs at a more reasonable rate.

Just for information, last week i had an Hitachi HDD with almost half of the sectors reallocated...
You can't always trust SMART !

ici_lemmy wrote:You can't always trust SMART !

Very true, but IMHO it's better than nothing (!) and in my experience, human interpretation and trend analysis, can be useful additions to the basic SMART pre-fail warning. Of course there is never a guarantee of a SMART warning before a drive becomes unusable.

ici_lemmy wrote:Just for information, last week i had an Hitachi HDD with almost half of the sectors reallocated...

That sounds like a sick disk. How is this comment related to SMART? Did you mean to explain something about this drive's SMART attributes, which you missed from your comment?

That sounds like a sick disk. How is this comment related to SMART? Did you mean to explain something about this drive's SMART attributes, which you missed from your comment?

Sorry. I didn't explain well : The SMART told me that almost half of the HDD sectors were reallocated !

Thanks for explaining - I might know what happened there.

You were looking at the raw value of attribute 5, and saw a very large number - yes? A raw value sometimes encodes multiple (e.g. 2 or 3) different numbers, and so it can appear to be a very large number when viewed as a single number, but actually that can be misleading. If you still have the raw SMART data from that drive, we can look at it again...

Vulcan wrote:Thanks for explaining - I might know what happened there.

You were looking at the raw value of attribute 5, and saw a very large number - yes?

Right

Vulcan wrote:A raw value sometimes encodes multiple (e.g. 2 or 3) different numbers, and so it can appear to be a very large number when viewed as a single number, but actually that can be misleading. If you still have the raw SMART data from that drive, we can look at it again...

Sorry the data recovery went well so the HDD is back to the customer.
But thank's for the info ! I'll try to look at this next time a see a big number in attribute 5.

ici_lemmy wrote:Sorry the data recovery went well so the HDD is back to the customer.

Well done! But that means we don't have the drive to examine further, unfortunately.

ici_lemmy wrote:But thank's for the info ! I'll try to look at this next time a see a big number in attribute 5.

Convert the whole raw value into hex, and you may find that it becomes obvious that there are 2 or 3 different numbers encoded within the one "apparent" value (e.g. hex ABCDABCD is the hex value ABCD encoded twice within that one number).

Since the SMART raw values are "vendor unique" (i.e. not defined by the current ATA standard), then vendors can choose to use the raw values in non-obvious ways (e.g. encoding multiple numbers into that one value etc.). In old versions of the SMART spec, the meanings of some of the raw values were defined, but not now... So sometimes we can be surprised / confused when we try to interpret those values