Bad SSD SMART status?

Toggle sidebar Toggle sidebar
T

timmer3

Honorable
Apr 8, 2013
5
0
10,510
hi, i have a 128Gb Samsung SSD. over the past month i've found that Windows has become increasingly unresponsive, normally instant tasks (typing in search query to browser address bar) freeze the application and can take 5-10 seconds. no changes in hardware, software or usage (software development). windows 7 memory diagnostic came up with nothing, CPU temps are fine, re-seated RAM etc., HD Tune disk benchmarks seem normal for the drive. in trying to identify the source of the performance degradation i looked up the SMART status of my SSD which runs Windows. a lot of the values seem to be at maximum. the ones that worry me are: wear levelling @ 93, erase fail count @ 100, runtime bad block @ 100, uncorrectable error count@ 100, ECC error rate @ 200. can anyone tell if these values indicate a failing drive? thanks for any help!

----------------------------------------------------------------------------
CrystalDiskInfo 5.5.1
----------------------------------------------------------------------------

OS : Windows 7 Ultimate Edition SP1 [6.1 Build 7601] (x86)
Date : 2013/04/08 13:39:41

-- Controller Map ----------------------------------------------------------
- ATA Channel 0 (0) [ATA]
- ATA Channel 1 (1) [ATA]
+ Intel(R) ICH10 Family 4 port Serial ATA Storage Controller 1 - 3A20 [ATA]
+ ATA Channel 0 (0)
- WDC WD5000YS-01MPB1 ATA Device
- SAMSUNG MMCRE28G5MXP-0VB ATA Device
+ ATA Channel 1 (1)
- WDC WD10EACS-32ZJB0 ATA Device
+ Intel(R) ICH10 Family 2 port Serial ATA Storage Controller 2 - 3A26 [ATA]
- ATA Channel 0 (0)
- ATA Channel 1 (1)
+ Marvell 61xx RAID Controller [SCSI]
- ASUS DRW-1608P3S SCSI CdRom Device
- Marvell Virtual Device

-- Disk List ---------------------------------------------------------------
(1) SAMSUNG MMCRE28G5MXP-0VB : 128.0 GB [0/0/0, pd1] - sg
(2) WDC WD5000YS-01MPB1 : 500.1 GB [1/0/1, pd1] - wd
(3) WDC WD10EACS-32ZJB0 : 1000.2 GB [2/1/1, pd1] - wd

----------------------------------------------------------------------------
(1) SAMSUNG MMCRE28G5MXP-0VB
----------------------------------------------------------------------------
Model : SAMSUNG MMCRE28G5MXP-0VB
Firmware : VBM1901Q
Serial Number : DC03301013SE013A5986
Disk Size : 128.0 GB (8.4/128.0/128.0/128.0)
Buffer Size : Unknown
Queue Depth : 32
# of Sectors : 250069680
Rotation Rate : ---- (SSD)
Interface : Serial ATA
Major Version : ATA/ATAPI-7
Minor Version : ATA/ATAPI-7 T13 1532D version 1
Transfer Mode : SATA/300
Power On Hours : 17452 hours
Power On Count : 835 count
Wear Level Count : 419
Temparature : Unknown
Health Status : Good (98 %)
Features : S.M.A.R.T., AAM, 48bit LBA, NCQ, TRIM
APM Level : ----
AAM Level : 8000h [OFF]

-- S.M.A.R.T. --------------------------------------------------------------
ID Cur Wor Thr RawValues(6) Attribute Name
09 _96 _96 __0 00000000442C Power-on Hours
0C _99 _99 __0 000000000343 Power-on Count
AF 100 100 _10 000000000000 Program Fail Count (Chip)
B0 100 100 _10 000000000000 Erase Fail Count (Chip)
B1 _93 _93 _17 0000000001A3 Wear Leveling Count
B2 _88 _88 _10 000000000007 Used Reserved Block Count (Chip)
B3 _98 _98 _10 000000000039 Used Reserved Block Count (Total)
B4 _98 _98 _10 000000000EC7 Unused Reserved Block Count (Total)
B5 100 100 _10 000000000000 Program Fail Count (Total)
B6 100 100 _10 000000000000 Erase Fail Count (Total)
B7 100 100 _10 000000000000 Runtime Bad Block (Total)
BB 100 100 __0 000000000000 Uncorrectable Error Count
C3 200 200 __0 000000000000 ECC Error Rate
C6 100 100 __0 000000000000 Off-Line Uncorrectable Error Count
C7 253 253 __0 000000000000 CRC Error Count
E8 _88 _88 _10 000000000035 Vendor Specific
E9 __1 __1 __0 0000FFFFFFFE Vendor Specific

 
T

timmer3

Honorable
Apr 8, 2013
5
0
10,510


hi Thanks for the suggestion. i downloaded that and it showed up with exactly the same figures. interestingly it identifies the SSD as a samsung drive but says "magician was unable to find a Samsung SSD connected to your system". i switched to AHCI mode in the bios but it doesn't seem to have made any difference, computer still feels like a single core CPU with not enough memory. it's a i7 920 @ 3.4Ghz with 3 Gb memory running in the correct XMP profile, and it used to be lightning!

at a loose end now. i'll start with a drive upgrade and see if that solves it. otherwise new mobo/cpu/ram, all components are 3 years old anyway so not the end of the world.
 
T

timmer3

Honorable
Apr 8, 2013
5
0
10,510
turns out the SSD had degraded over the 3 years i've had it, although it wasn't showing up in any benchmarks. i wiped the drive, used AS Cleaner with the FF option, and restored from backup, back to top performance now, no delays anymore. unfortunately there aren't any firmware upgrades available but it's an early model and i've done well enough out of it, looking forward getting one of the current generation SSDs.
 
F

fzabkar

Judicious
Sep 22, 2010
9,290
697
36,390
I can't find a description of attributes E8 and E9 in Samsung's documentation, but perhaps Intel defines the attributes in a similar way ???

http://code.google.com/p/hddguardian/wiki/Samsung_SMART_attributes
http://code.google.com/p/hddguardian/wiki/Intel_SMART_attributes

E8 - Available Reserved Space - This attribute reports the number of reserve blocks remaining.

E9 - Media Wearout Indicator - This attribute reports the number of cycles the NAND media has experienced.

If we examine the following attributes, ...

0C _99 _99 __0 000000000343 Power-on Count
B1 _93 _93 _17 0000000001A3 Wear Leveling Count

... and if we assume that the drive's wear levelling was perfect during its lifetime, then it appears that your drive wrote an amount of data equivalent to ...

total bytes written = capacity x number of times SSD was completely wear levelled

= 128E9 x 0x1A3 bytes = 48.8 TiB

http://www.google.com/search?q=128E9+x+0x1A3+bytes+in+terabytes

The Power On Count is 0x343 hours, so the average write data transfer rate over the drive's entire lifetime appears to be ...

128E9 x 0x1A3 bytes / 0x343 hours = 17 megabytes per second

http://www.google.com/search?q=0x1A3+x+128E9+bytes+%2F+0x343+hours+in+megabytes+per+second

Does that sound right?

BTW, normalised values represent "health" scores. The raw values are the actual data for the attribute. Therefore a normalised value of 100 is a perfect score.

The worst attribute is E9. IMHO B2 and B3 are also of concern.
 
T

timmer3

Honorable
Apr 8, 2013
5
0
10,510


many thanks for the detailed reply and calculation. 17 Mb/s sounds like a lot, but i did have the swap file there the whole time. i mistakenly thought the drive had TRIM enabled, and there were conflicting reports about whether the swap file should be on the SSD or not. this seems to have taken its toll on the drive. i've used the computer for software development, which probably isn't all that intensive in terms of disk writes. the re-format should tide me over for a few months until haswell comes out and i'll upgrade all the components then.
 
F

fzabkar

Judicious
Sep 22, 2010
9,290
697
36,390
Your usage works out to 1.4TB per day:

http://www.google.com.au/search?q=128E9+x+0x1A3+bytes+per+0x343+hours+in+terabytes+per+day

If we examine the following attribute ...

B1 _93 _93 _17 0000000001A3 Wear Leveling Count

... it would appear that the drive has lost 7 points (= 100 - 93) from its "health" score.

The threshold is 17, so it would appear that the number of times that the drive could be wear levelled before hitting its SMART threshold would be about ...

0x1A3 / (100 - 93) x (100 - 17) = 4968

This suggests to me that the NAND flash chips are rated for about 5000 writes.

FWIW, Hynix H27UBG8T2ATR MLC NANDs are rated for 3000 Program/Erase cycles:

http://www.szyuda88.com/uploadfile/cfile/201061714220663.pdf
 
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom