Need help with "reallocated sector count"?
in [Storage]
|
Prev: INT13/02h Read error (AX=0100, DL=80) trying xp installation in SATA
Next: Need to create a CDFS partition on a USB key
From: Arno Wagner on 12 Nov 2007 19:57 Previously Franc Zabkar <fzabkar(a)iinternode.on.net> wrote: > On Mon, 12 Nov 2007 06:43:08 -0800, kimiraikkonen > <kimiraikkonen85(a)gmail.com> put finger to keyboard and composed: >>Here is Everest ones about "reallocated sectors count": >> >>ID Attribute Description Threshold Value Worst >>Data Status >>05 Reallocated Sector Count 36 98 98 >>98 OK: Value is normal >> >>I also check with ActiveSMART saying the raw value is: 98 >> >>Arno said it counts down, i had 2 bad-sectors at the past which i >>fixed using Seatools. Since that, i haven't had any bad-blocks shown >>in chkdsk or Seatools full surface scan. >> >>So what does that values mean? >>ID Attribute Description Threshold Value Worst >>Data Status >>05 Reallocated Sector Count 36 98 98 >>98 OK: Value is normal >> >>SmartUDM from Dos: Raw: 000000000062h >> reallocated sectors: 98 (but how >>reliable is it?) > OK, I see the reason for my confusion. In your case the actual raw > value of 98 sectors (=62 hex) coincides with the "percentage" value or > "normalized" value of 98. Pure coincidence. That looks very likely to me too now. Quite confusing, I agree. 98 bad sectors is a high number. If it does not increase, the drive may still be fine (there are those that discard a drive at the first reallocated secotr, I prefer RAID1 and backups). Arno
From: Rod Speed on 12 Nov 2007 22:49 kimiraikkonen <kimiraikkonen85(a)gmail.com> wrote: > "Rod", No need to quote that, that is my real name. > here is full SMART report from Everest, It looks fine. > it say everyvalue is normal. It always does unless the drive is very close to imminent death, ignore that and focus on the Data numbers. > Meanwhile, this reallocated sectors may remain from past Yes, they certainly are, but you dont get that many reallocated sectors unless the drive is dying. > and i don't want to conclude unless they dramatically continue to populate. Doesnt need to be dramatic, if they keep increasing, the drive is dying. > ID Attribute Description Threshold Value Worst Data Status > > 01 Raw Read Error Rate 34 64 53 171388353 > OK: Value is normal > 03 Spin Up Time 0 70 70 > 0 OK: Always passes > 04 Start/Stop Count 20 100 100 > 692 OK: Value is normal > 05 Reallocated Sector Count 36 98 98 > 98 OK: Value is normal > 07 Seek Error Rate 30 81 60 > 158600840 OK: Value is normal Thats normal for a seagate. That value does vary with the manufacturer. > 09 Power-On Time Count 0 93 93 > 6581 OK: Always passes > 0A Spin Retry Count 97 100 100 > 0 OK: Value is normal > 0C Power Cycle Count 20 98 98 > 2590 OK: Value is normal > C2 Temperature 0 28 51 > 28 OK: Always passes It has been a bit high in the past, but not enough to be the cause of the reallocated sectors. Thats the main reason I wanted the full report to see if there was any other problem but the one you showed. > C3 Hardware ECC Recovered 0 64 53 > 171388353 OK: Always passes Ditto with seagate drives. > C5 Current Pending Sector Count 0 100 100 > 0 OK: Always passes > C6 Off-Line Uncorrectable Sector Count 0 100 100 > 0 OK: Always passes Those two say that there arent any bads that havent been reallocated. > C7 Ultra ATA CRC Error Rate 0 200 200 > 0 OK: Always passes Thats the ribbon cable, its fine. > C8 Write Error Rate 0 100 253 > 0 OK: Always passes > CA TA Increase Count 0 100 253 > 0 OK: Always passes
From: Rod Speed on 12 Nov 2007 22:50 Arno Wagner <me(a)privacy.net> wrote: > Previously kimiraikkonen <kimiraikkonen85(a)gmail.com> wrote: >> On Nov 12, 4:25 am, Arno Wagner <m...(a)privacy.net> wrote: >>> Previously kimiraikkonen <kimiraikkone...(a)gmail.com> wrote: >>>> On 11 Kas m, 22:55, Arno Wagner <m...(a)privacy.net> wrote: >>>>> Previously kimiraikkonen <kimiraikkone...(a)gmail.com> wrote: >>>>>> Hello, >>>>>> I want to ask a question about my Seagate drives SMART attribute >>>>>> "reallocated sector count". >>>>>> "reallocated sector count" is at the limit. The values are i >>>>>> look with my Smart utility: >>>>>> Current: 98, Worst: 98, Threshold: 36, Data: 98 >>> >>>>> That is not at the limit. These attributes count down. >>>>> The limit would be 36, it is currently at 98, likely down >>>>> from 100. >>> >>>>>> I checked my drive a lot of times with SEATOOLS with "complete >>>>>> FULL scan" also checked with regular chkdsk /f function. No bad >>>>>> blocks are found (zero) 0 kb. >>> >>>>> That is why these are "re"-allocated. The bad blocks are not >>>>> visible anymore. >>> >>>>>> So, what does the warning about "reallocated sector count"? >>>>>> Are they really bad sectors which are hidden or what is it? >>>>>> If there are bad blocks hidden, why is there popular utilities >>>>>> like Seatools which can find bad blocks and replace them >>>>>> (zero-fill replacing)? >>> >>>>> Because the disk can not allways reallocated bad blocks. >>>>> It basically can if a) the block is bad but still readable or b) >>>>> it found the bad block in a surface scan and it gets overwritten >>>>> befiore it gets read. >>> >>>>>> Please help. >>> >>>>> Now, there are no raw numbers given by the SMART utility you use. >>>>> Or this disk does not give you a raw reallocation count. One >>>>> thing you should do is to tru to get a raw realocation count. >>>>> This may be two sectors or 200, hard to tell. The other thing is >>>>> that while your disk may have a problem, it might also be fine. >>>>> The key to determining this is to observe the disk carefully. If >>>>> it gets more reallocated sectors over time, replace it. If not, >>>>> it may be fine. For this you need the raw value again. Also run a >>>>> long SMART selftest every week or so for some time. And keep your >>>>> backups current. >>> >>>>> Other poossibel sources of bad secors: Bad PSU, mechanical shock >>>>> or vibration. >>> >>>>> Arno >>>> Hi Arno, >>>> Thanks for replying. It was very helpful and relaxing. I want to >>>> tell its short history: >>>> At past i had 2 bad sectors on that(same) disk which were not >>>> physical(logical) and replaced (zero-filled) via Seatools(officiall >>>> diagnostic utility) easily. >>>> Since then, i frequently scan full surface of my drive i don't see >>>> any >>>> bad sectors reported since 2 bad ones have been repaired by >>>> Seatools. So what is "98" mean at this case? >>> >>> It just means 98 out of 100, with no fixed relation. Unless it >>> ist the "raw" value, then it means 98 defective and reallocated >>> secors. >>> That would be a lot. In your case it could mean that the vendor has >>> choosen to decrese it by 1 for each reallocated secor and therefore >>> allow 64 reallocated secors before a bad SMART status is reached. > >> I don't think its raw value. How will i know? There are only: >> current:98 worst:98 threshold:36 data:98. I don't think there are >> much bad blocks (98 is so much) and never had any serious problem >> that may point 98 bad-blocks. > >>>> So, is there anything than i must concern at the moment? Is there >>>> any present bad sectors although i fixed(zero-filled) those 2 ones >>>> before? >>> >>> After long SMART selftest/surface scan, there should be no >>> unrecognized bad secors. BTW, recognized bad secors that >>> could not (yet) be reallocated are listed unter the >>> "Pending sectors" (or the like) attribute. >>> >>>> Also "reallocated sectors count" is descibed as "lower value is >>>> better" so if threshold is 36, >>>> isn't going to lower value better? Confusing? >>> >>> Huh? Where did you find "lower value is better"? All >>> SMART attributes have a "value decresed on problem" semantics >>> in the "coocked" form. Of course in the raw form, it can be >>> different, but the threshold and ordinary attibute display is >>> cooked. Let me give you an example from one of my disks: > >>> From here an other sites: >> http://en.wikipedia.org/wiki/Self-Monitoring%2C_Analysis%2C_and_Reporting_Technology > >> Says "reallocated sectors count" value lower is better. > >>> ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE >>> UPDATED WHEN_FAILED RAW_VALUE 1 Raw_Read_Error_Rate 0x000f >>> 100 100 051 Pre-fail Always - 0 3 >>> Spin_Up_Time 0x0007 100 100 025 Pre-fail >>> Always - 6080 4 Start_Stop_Count 0x0032 099 >>> 099 000 Old_age Always - 1115 5 >>> Reallocated_Sector_Ct 0x0033 253 253 010 Pre-fail >>> Always - 0 >>> >>> "Value", "Worst" and "Threshold" are cooked values, and lower is >>> always worse. "Raw_Value" is the register value, and here hogher >>> is indeed worse for reallocated sector count. Note that this disk >>> has zero reallocated sectors. > > > >> So, as summary you advise to consider / care "raw value" ? > > For reallocated sectors, yes! Especially since the raw > value is usually a direct count. (Except for some > notepook HDD I have, which seems to count down from > 4096000000....) > >> Which >> programs will show "raw value"? Could you give an Windows-based >> sample? I tried about 3 programs saying only: current, value, >> threshold and worst... > > > The above is from the smartmontools, which have been ported > to Windows as well. Commandline only, but easily the best > functionality. http://smartmontools.sourceforge.net/ But not the most readable report around.
From: Franc Zabkar on 13 Nov 2007 01:42 On 13 Nov 2007 00:57:43 GMT, Arno Wagner <me(a)privacy.net> put finger to keyboard and composed: >Previously Franc Zabkar <fzabkar(a)iinternode.on.net> wrote: >> OK, I see the reason for my confusion. In your case the actual raw >> value of 98 sectors (=62 hex) coincides with the "percentage" value or >> "normalized" value of 98. Pure coincidence. > >That looks very likely to me too now. Quite confusing, I agree. > >98 bad sectors is a high number. If it does not increase, the >drive may still be fine (there are those that discard a drive >at the first reallocated secotr, I prefer RAID1 and backups). > >Arno I've been living with a dying drive for at least two years. This last week was the last straw, though. I find that Seagate's threshold value of 36 is somewhat optimistic. If I have correctly interpreted my logs, then each percentage (?) point corresponds to a loss of approximately 40 sectors. So a value of 36 represents a loss of 64 points, which in turn corresponds to about 2560 reallocated sectors. - Franc Zabkar -- Please remove one 'i' from my address when replying by email.
From: Arno Wagner on 13 Nov 2007 02:34
Previously Franc Zabkar <fzabkar(a)iinternode.on.net> wrote: > On 13 Nov 2007 00:57:43 GMT, Arno Wagner <me(a)privacy.net> put finger > to keyboard and composed: >>Previously Franc Zabkar <fzabkar(a)iinternode.on.net> wrote: >>> OK, I see the reason for my confusion. In your case the actual raw >>> value of 98 sectors (=62 hex) coincides with the "percentage" value or >>> "normalized" value of 98. Pure coincidence. >> >>That looks very likely to me too now. Quite confusing, I agree. >> >>98 bad sectors is a high number. If it does not increase, the >>drive may still be fine (there are those that discard a drive >>at the first reallocated secotr, I prefer RAID1 and backups). >> >>Arno > I've been living with a dying drive for at least two years. Gutsy! ;-) > This last week was the last straw, though. > I find that Seagate's threshold value of 36 is somewhat optimistic. If > I have correctly interpreted my logs, then each percentage (?) point > corresponds to a loss of approximately 40 sectors. So a value of 36 > represents a loss of 64 points, which in turn corresponds to about > 2560 reallocated sectors. Well possible. I once had a Maxtor (in a cluster of compute servers) that was incredible slow and has about 1100 reallocated sectors. This thing was dying pretty fast (had been dropped and it took some weeks to develop problems). The thing was that the SMART status still read good, i.e. above the threshold. At that time I started monitoring the raw reallocated sector count and installed email notification on changes of that.. Some vendors are extremely optimisticc with regard to SMART thresholds. Kind of makes the SMART status alone pretty worthless. No wonder so many people are asking in this group for help interpreting SMART data. Arno |