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

[Tuomaz]

Hi

I have 7 Barracuda LP 1TB ST31000520AS. 5 of them are used in a software RAID6 on Ubuntu Server (kernel 2.6.30-02063005-generic). I have set the spindown time to 1h with

hdparm -S242 /dev/sdb

/dev/sdb:
 setting standby to 242 (1 hours)

During use of the drives I have noticed a very frequent clicking sound. When I investigated this further I realized that 4 of the 5 disks in the array had a Start_Stop_Count above 25K! Half of the rated value in 90 days!:

sudo smartctl --all /dev/sdc
smartctl version 5.38 [x86_64-unknown-linux-gnu] Copyright (C) 2002-8 Bruce Allen
Home page is http://smartmontools.sourceforge.net/

=== START OF READ SMART DATA SECTION ===
SMART Attributes Data Structure revision number: 10
Vendor Specific SMART Attributes with Thresholds:
ID# ATTRIBUTE_NAME          FLAG     VALUE WORST THRESH TYPE      UPDATED  WHEN_FAILED RAW_VALUE
  1 Raw_Read_Error_Rate     0x000f   117   099   006    Pre-fail  Always       -       130780386
  3 Spin_Up_Time            0x0003   096   095   000    Pre-fail  Always       -       0
  4 Start_Stop_Count        0x0032   075   075   020    Old_age   Always       -       26557
  5 Reallocated_Sector_Ct   0x0033   100   100   036    Pre-fail  Always       -       0
  7 Seek_Error_Rate         0x000f   069   060   030    Pre-fail  Always       -       9998207
  9 Power_On_Hours          0x0032   098   098   000    Old_age   Always       -       2191
 10 Spin_Retry_Count        0x0013   100   100   097    Pre-fail  Always       -       0
 12 Power_Cycle_Count       0x0032   100   100   020    Old_age   Always       -       26
183 Unknown_Attribute       0x0032   100   100   000    Old_age   Always       -       0
184 Unknown_Attribute       0x0032   100   100   099    Old_age   Always       -       0
187 Reported_Uncorrect      0x0032   100   100   000    Old_age   Always       -       0
188 Unknown_Attribute       0x0032   100   099   000    Old_age   Always       -       1
189 High_Fly_Writes         0x003a   077   077   000    Old_age   Always       -       23
190 Airflow_Temperature_Cel 0x0022   063   044   045    Old_age   Always   In_the_past 37 (0 1 37 37)
194 Temperature_Celsius     0x0022   037   056   000    Old_age   Always       -       37 (0 23 0 0)
195 Hardware_ECC_Recovered  0x001a   036   028   000    Old_age   Always       -       130780386
197 Current_Pending_Sector  0x0012   100   100   000    Old_age   Always       -       0
198 Offline_Uncorrectable   0x0010   100   100   000    Old_age   Offline      -       0
199 UDMA_CRC_Error_Count    0x003e   200   200   000    Old_age   Always       -       0
240 Head_Flying_Hours       0x0000   100   253   000    Old_age   Offline      -       233019155678609
241 Unknown_Attribute       0x0000   100   253   000    Old_age   Offline      -       4099986043
242 Unknown_Attribute       0x0000   100   253   000    Old_age   Offline      -       1907871403

If I read information about one of the 4 problem drives with hdparm:

sudo hdparm -I /dev/sdc

/dev/sdc:

ATA device, with non-removable media
        Model Number:       ST31000520AS
        Serial Number:      5VX--------
        Firmware Revision:  CC32
        Transport:          Serial
Standards:
        Used: unknown (minor revision code 0x0029)
        Supported: 8 7 6 5
        Likely used: 8
Configuration:
        Logical         max     current
        cylinders       16383   16383
        heads           16      16
        sectors/track   63      63
        --
        CHS current addressable sectors:   16514064
        LBA    user addressable sectors:  268435455
        LBA48  user addressable sectors: 1953525168
        device size with M = 1024*1024:      953869 MBytes
        device size with M = 1000*1000:     1000204 MBytes (1000 GB)
Capabilities:
        LBA, IORDY(can be disabled)
        Queue depth: 32
        Standby timer values: spec'd by Standard, no device specific minimum
        R/W multiple sector transfer: Max = 16  Current = 16
        Recommended acoustic management value: 254, current value: 0
        DMA: mdma0 mdma1 mdma2 udma0 udma1 udma2 udma3 udma4 udma5 *udma6
             Cycle time: min=120ns recommended=120ns
        PIO: pio0 pio1 pio2 pio3 pio4
             Cycle time: no flow control=120ns  IORDY flow control=120ns
... 

There is no line "Advanced power management level"! If I do the same on the fifth drive I get:

...
Queue depth: 32
        Standby timer values: spec'd by Standard, no device specific minimum
        R/W multiple sector transfer: Max = 16  Current = 16
        Advanced power management level: 254
        Recommended acoustic management value: 254, current value: 0
        DMA: mdma0 mdma1 mdma2 udma0 udma1 udma2 udma3 udma4 udma5 *udma6
             Cycle time: min=120ns recommended=120ns
...

If I try to set APM on one of the 4 problem drives I get:

hdparm -B254 /dev/sdb

/dev/sdb:
 setting Advanced Power Management level to 0xfe (254)
 HDIO_DRIVE_CMD failed: Input/output error

 and on the fifth, healty drive:

sudo hdparm -B254 /dev/sdf

/dev/sdf:
 setting Advanced Power Management level to 0xfe (254)

This is very frustrating and if the Start_Stop_Count continues to skyrocket the 4 drives will soon die.

It is the same firmware revision on all of the 5 drives ("CC32"). The fifth drive is bought in another batch then the first 4 though.

The system is a Asus M4A78-VM motherboard (AMD 780G+SB700 chipsets) with a 4850e AMD processor. I have tried both the SATA mode and the AHCI mode in the bios.

How do I fix this? Why can't I set APM on 4 of the drives? How is it possible for the Start_Stop_Count to reach 25k in 90 days with a spindown time of 1h?

 / Tuomaz

 

Message Edited by Tuomaz on 09-10-2009 12:31 AM

[fzabkar]

Barracuda LP Series SATA Product Manual, Rev. A:

http://www.seagate.com/staticfiles/support/disc/manuals/desktop/Barracuda%20LP/100564361a.pdf

Page 29 of the manual (page 35 of the PDF) states that "Advanced Power Management (APM) and Automatic Acoustic Management (AAM) features are not supported."

Page 12 (18) has this to say about Standby mode:

"The drive enters Standby mode when the host sends a Standby Immediate command. If the host has set the standby timer, the drive can also enter Standby mode automatically after the drive has been inactive for a specifiable length of time. The standby timer delay is established using a Standby or Idle command. In Standby mode, the drive buffer is enabled, the heads are parked and the spindle is at rest. The drive accepts all commands and returns to Active mode any time disc access is necessary."

See pages 264 (STANDBY) and 171 (Table 40) of the "AT Attachment 8 - ATA/ATAPI Command Set (ATA8-ACS)":

http://www.t13.org/Documents/UploadedDocuments/docs2008/D1699r6a-ATA8-ACS.pdf

[Tuomaz]

Hmm. This just raises two new questions: 1. Why can I set APM on one of the drives? and 2. How do I stop the Start_Stop_Count to increase dramatically?

[fzabkar]

No doubt everyone is just as puzzled as you are. :-(

One thing I would suggest is to capture the 512 bytes (= 256 words) returned by each drive in response to an ATA Identify Device command. Then look for differences.

You can retrieve these data with smartmontools in debug mode, eg ...

smartctl -a -r ioctl,2 -d usbjmicron scsi10

The above example generates the following log file for my AMD K6-2 Win98SE box using a generic Microsoft USB mass storage driver, with a Seagate 320GB drive in an external USB enclosure, behind a JMicron USB-ATA bridge chip:

http://www.users.on.net/~fzabkar/Smartctl/320GB_all.log

The Identify Device command is documented on page 114 of the ATA-8 spec.

[Tuomaz]

Now I have done some further testing. I set the sleep time for one of the first 6 drives (bought in one batch) and not used in my raid array to 1h yesterday. After some hours I checked with

sudo hdparm -C /dev/sda

/dev/sda:
 drive state is:  standby

"standby" means that the spindle has stopped. This drive had a Start_Stop_Count value of 12 before I set the sleep time to 1h. Today I checked it with smartctl (smartctl wakes the drive from standby) :

sudo smartctl -A /dev/sda
smartctl version 5.38 [i686-pc-linux-gnu] Copyright (C) 2002-8 Bruce Allen
Home page is http://smartmontools.sourceforge.net/

=== START OF READ SMART DATA SECTION ===
SMART Attributes Data Structure revision number: 10
Vendor Specific SMART Attributes with Thresholds:
ID# ATTRIBUTE_NAME          FLAG     VALUE WORST THRESH TYPE      UPDATED  WHEN_FAILED RAW_VALUE
  1 Raw_Read_Error_Rate     0x000f   119   099   006    Pre-fail  Always       -       234129350
  3 Spin_Up_Time            0x0003   096   096   000    Pre-fail  Always       -       0
  4 Start_Stop_Count        0x0032   100   100   020    Old_age   Always       -       898
  5 Reallocated_Sector_Ct   0x0033   100   100   036    Pre-fail  Always       -       0

In 24h of sleeping the drive has done 886 load/unload of the heads. Why?

For safety reasons I have now disabled sleep for all my Seagate drives. I have tested the same thing on my 7th drive but not for 24h, only 3h. It just inreased its Start_Stop_Count with 2 (which is the expected value).

I suspect that Seagate has done somehting to correct this issue on newer drives (like my 7th). The question is if the drives actually does all this reported load/unload or if it is just an error in the code that counts the Start_Stop.

[fzabkar]

Interesting experiment.

886 load/unload cycles in 24 hours amounts to about 1 every 100 seconds.

Have any of the other SMART attributes changed in value during that time?

In particular I'd compare the Seek Error Rate before and after. The raw value actually counts the number of seeks -- it is not an error rate.

Also look at the Head Flying Hours.

For example, a raw value of 233019155678609 equates to a hex value of 0xd3ee00000591. I suspect that the actual flying time is 0x00000591 hours (the lower 32 bits), ie 1425. Power_On_Hours is 2191.

BTW, have a look at this [unrelated] thread:

http://forums.seagate.com/stx/board/message?board.id=ata_drives&message.id=14681#M14681

You'll see that you're not the only one with firmware issues.

If you need a hex calculator, try this one:

http://www.mrcalculator.com/hexdec.html

For small numbers you can use Google's calculator, eg ...

http://www.google.com/search?hl=en&client=opera&rls=en&hs=PQP&num=25&newwindow=1&q=0x591+in+decimal&btnG=Search

http://www.google.com/search?hl=en&client=opera&rls=en&hs=9k4&num=25&newwindow=1&q=1425+in+hex&btnG=Search

Message Edited by fzabkar on 10-09-2009 01:24 AM

[fzabkar]

I found the following information in the ATA-8 command spec. Perhaps the APM feature set can be enabled using ATA DCO commands.

"The optional DCO [Device Configuration Overlay] feature set allows a utility program to reduce the capability of the device by modifying some of the optional commands, modes, and feature sets that a device reports as supported in the IDENTIFY DEVICE or IDENTIFY PACKET DEVICE data as well as the capacity reported."

7.10.3 DEVICE CONFIGURATION IDENTIFY - B1h/C2h, PIO Data-In

"The DEVICE CONFIGURATION IDENTIFY command causes a device to return a 512-byte data structure. The content of this data structure indicates the selectable commands, modes, and feature sets that the device is capable of disabling or modifying through processing of a DEVICE CONFIGURATION SET command. If a DEVICE CONFIGURATION SET command reducing a device's capabilities has completed without error, then:

a) the response by a device to an IDENTIFY DEVICE, IDENTIFY PACKET DEVICE, and other commands, except the DEVICE CONFIGURATION IDENTIFY command, shall reflect the reduced set of capabilities; and

b) the response by a device to a DEVICE CONFIGURATION IDENTIFY command shall reflect the entire set of selectable capabilities."

[Tuomaz]

Here are before and after values from 2h of sleep for one of my drives:

cat 090910-1106 090910-1305
smartctl version 5.38 [i686-pc-linux-gnu] Copyright (C) 2002-8 Bruce Allen
Home page is http://smartmontools.sourceforge.net/

=== START OF READ SMART DATA SECTION ===
SMART Attributes Data Structure revision number: 10
Vendor Specific SMART Attributes with Thresholds:
ID# ATTRIBUTE_NAME          FLAG     VALUE WORST THRESH TYPE      UPDATED  WHEN_FAILED RAW_VALUE
  1 Raw_Read_Error_Rate     0x000f   119   099   006    Pre-fail  Always       -       132241
  3 Spin_Up_Time            0x0003   096   096   000    Pre-fail  Always       -       0
  4 Start_Stop_Count        0x0032   100   100   020    Old_age   Always       -       945
  5 Reallocated_Sector_Ct   0x0033   100   100   036    Pre-fail  Always       -       0
  7 Seek_Error_Rate         0x000f   064   060   030    Pre-fail  Always       -       2970203
  9 Power_On_Hours          0x0032   098   098   000    Old_age   Always       -       1956
 10 Spin_Retry_Count        0x0013   100   100   097    Pre-fail  Always       -       0
 12 Power_Cycle_Count       0x0032   100   100   020    Old_age   Always       -       13
183 Unknown_Attribute       0x0032   100   100   000    Old_age   Always       -       0
184 Unknown_Attribute       0x0032   100   100   099    Old_age   Always       -       0
187 Reported_Uncorrect      0x0032   100   100   000    Old_age   Always       -       0
188 Unknown_Attribute       0x0032   100   100   000    Old_age   Always       -       3
189 High_Fly_Writes         0x003a   100   100   000    Old_age   Always       -       0
190 Airflow_Temperature_Cel 0x0022   082   070   045    Old_age   Always       -       18 (Lifetime Min/Max 18/28)
194 Temperature_Celsius     0x0022   018   040   000    Old_age   Always       -       18 (Lifetime Min/Max 0/18)
195 Hardware_ECC_Recovered  0x001a   038   022   000    Old_age   Always       -       132241
197 Current_Pending_Sector  0x0012   100   100   000    Old_age   Always       -       0
198 Offline_Uncorrectable   0x0010   100   100   000    Old_age   Offline      -       0
199 UDMA_CRC_Error_Count    0x003e   200   200   000    Old_age   Always       -       0
240 Head_Flying_Hours       0x0000   100   253   000    Old_age   Offline      -       186118112806821
241 Unknown_Attribute       0x0000   100   253   000    Old_age   Offline      -       2122586099
242 Unknown_Attribute       0x0000   100   253   000    Old_age   Offline      -       2800954763

smartctl version 5.38 [i686-pc-linux-gnu] Copyright (C) 2002-8 Bruce Allen
Home page is http://smartmontools.sourceforge.net/

=== START OF READ SMART DATA SECTION ===
SMART Attributes Data Structure revision number: 10
Vendor Specific SMART Attributes with Thresholds:
ID# ATTRIBUTE_NAME          FLAG     VALUE WORST THRESH TYPE      UPDATED  WHEN_FAILED RAW_VALUE
  1 Raw_Read_Error_Rate     0x000f   119   099   006    Pre-fail  Always       -       132241
  3 Spin_Up_Time            0x0003   096   096   000    Pre-fail  Always       -       0
  4 Start_Stop_Count        0x0032   100   100   020    Old_age   Always       -       1003
  5 Reallocated_Sector_Ct   0x0033   100   100   036    Pre-fail  Always       -       0
  7 Seek_Error_Rate         0x000f   064   060   030    Pre-fail  Always       -       2970208
  9 Power_On_Hours          0x0032   098   098   000    Old_age   Always       -       1958
 10 Spin_Retry_Count        0x0013   100   100   097    Pre-fail  Always       -       0
 12 Power_Cycle_Count       0x0032   100   100   020    Old_age   Always       -       13
183 Unknown_Attribute       0x0032   100   100   000    Old_age   Always       -       0
184 Unknown_Attribute       0x0032   100   100   099    Old_age   Always       -       0
187 Reported_Uncorrect      0x0032   100   100   000    Old_age   Always       -       0
188 Unknown_Attribute       0x0032   100   100   000    Old_age   Always       -       4
189 High_Fly_Writes         0x003a   100   100   000    Old_age   Always       -       0
190 Airflow_Temperature_Cel 0x0022   082   070   045    Old_age   Always       -       18 (Lifetime Min/Max 18/28)
194 Temperature_Celsius     0x0022   018   040   000    Old_age   Always       -       18 (Lifetime Min/Max 0/18)
195 Hardware_ECC_Recovered  0x001a   038   022   000    Old_age   Always       -       132241
197 Current_Pending_Sector  0x0012   100   100   000    Old_age   Always       -       0
198 Offline_Uncorrectable   0x0010   100   100   000    Old_age   Offline      -       0
199 UDMA_CRC_Error_Count    0x003e   200   200   000    Old_age   Always       -       0
240 Head_Flying_Hours       0x0000   100   253   000    Old_age   Offline      -       66052302047143
241 Unknown_Attribute       0x0000   100   253   000    Old_age   Offline      -       2122586099
242 Unknown_Attribute       0x0000   100   253   000    Old_age   Offline      -       2800954763

Commands were:

smartctl -A /dev/sda
hdparm -S120 /dev/sda
hdparm -Y /dev/sda
smartctl -A /dev/sda

 (no hdparm -Y, it seems like it wakes the drive)

Haven't got time right now to carefully read you last post fzabkar, but it seems like you know very much about harddrives and the different ATA standards. Thank you.

  

[fzabkar]

Thanks for the compliment, but I assure you I'm reading the ATA-8 standard for the first time.

I notice this time that the Start_Stop_Count increased by 58 in 2 hours. That's one event every 124 seconds. Assuming the events are regularly spaced, then you would only need to listen for activity for about 2 minutes.

I also notice that the Raw_Read_Error_Rate value does not change, and that the Seek_Error_Rate increases by only 5 counts. The 5 seeks would be incurred when the drive retrieves SMART data from the platters. Otherwise there does not appear to be any sign of software induced activity.

Power_On_Hours increases from 1956 to 1958. At the same time Head_Flying_Hours goes from 186118112806821 to 66052302047143.

In hexadecimal these numbers are 0xa946000007a5 and 0x3c13000007a7.

If I'm right, then the actual flying hours are 0x7a5 (= 1957) and 0x7a7 (= 1959) which of course doesn't make too much sense. It certainly doesn't tally with what we observed for your other drive.

The other curiosity is "188 Unknown_Attribute" which increases from 3 to 4.

According to this article ...

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

... attribute 188 represents Command Timeouts.

[acranox]

Did you find out anything more on this issue?  I used `hdparm -y` to put one of my disks into standby today, and in 8 hours of being spun down and inactive (I kept checking it with hdparm -C, it remained in standby the whole time) my Start_Stop_Count went up by over 150.  This just doesn't make any sense.  I can hear that the drive is inactive, so why is the Start_Stop_Count increasing?