Disks with TLER / ERC / CCTL & LCC [Table of drives]

[mhakali]

I have been looking at some new drives, especially 2TB and 3TB. It can however be very hard to find information whether tools support setting TLER / ERC / CCTL and/or if the ability exists to modify LCC.

I think there might be a general interest to know if these values can be modified. I know there's a lot of scattered information around aswell. A collective place would be very nice to look up these parameters, and I will start by contributing with the information I have. I am hoping that you will post replies to this threads contributing with additional information. I will update the post accordingly.



Known disks / models

Manufacturer - Model - Firmware - Interface - Cache - TLER - LCC - Comment

2TB

Hitachi - Ultrastar A7K2000 HUA722020ALA330 - JKAOA3HA - 3Gb/s - 32MB - FACTORY SET TO 7s - N/A - Disks shipping with Dell servers
* Samsung - EcoGreen F4 HD204UI - ? - 3Gb/s - 32MB - NO - ? - I did not get it working, but parityboy did! Verifications please?
* Samsung - EcoGreen F4 HD204UI - 1AQ10001 - 3Gb/s - 32MB - YES (TLER) - YES (LCC) - Will not survive a restart (parityboy) - More verifications please!
Samsung - EcoGreen F3 HD203WI - 1AN10002 - 3Gb/s - 32MB - YES - YES - Reliable, no permanent setting
Samsung - EcoGreen F3 HD203WI - 1AN10003 - 3Gb/s - 32MB - YES - YES - Reliable, no permanent setting
Seagate - Barracuda Green ST2000DL003 - CC32 - 6Gb/s - 64MB - NO - NO - Seagate confirmed there is no way for modifying the values
Western Digital - Caviar Green WD20EARS - ? - 3Gb/s - 64MB - NO - YES
* Western Digital - Caviar Green WD20EADS - ?- 3Gb/s - 32MB - YES - YES - rmd3003 -- BEWARE: Only earlier hw revisions!

1.5TB

Seagate - Barracuda 7200.11 ST31500341AS - CC1H - 3Gb/s - 32MB - YES - YES - Been running 10x in RAID6 without failure for 2 years.

1TB

* Western Digital - Caviar Green WD10EACS - ?- 3Gb/s - 16MB - YES - YES - rmd3003 -- BEWARE: Only earlier hw revisions!



TLER?

Time Limited Error Recovery - The ability for drives to remap sectors and silently recover from errors. This can take a while, so by being able to set TLER you can improve your chances for disks surviving in RAID setups.

Error Recovery Control (ERC), used by Seagate, and Command Completion Time Limit (CCTL), used by Samsung and Hitachi provide the same functionality.

Some drives have modifiable firmwares and can set hard values for enabling TLER. This is preferable if you are going to use a hardware RAID controller, where the S.M.A.R.T. may not be accessible for modification outside the controller. But this depends on the controller itself.

Some drives have the ability to receive S.M.A.R.T. commands to set TLER to 7s during the operation of the drive. This value is reset on reboot or power loss on drive.



LCC?

Load Cycle Count, an internal counter of the harddrives S.M.A.R.T. table which records how many times a drive parks it's head. Many recent 5400-5900 RPM "Green" drives provides the functionality to increase drive life span and energy consumption.

However, using the drives in a RAID will cause many load/releases of the head as all drives are accessed simultaneously. This can cause premature failure.



Setting TLER

Use smartctl from smartmontools, it exists for both Linux, BSD & Windows.

Usage:

smartctl -l scterc,70,70 <device>

Under Linux you use the standard paths such as "/dev/sda", under Windows you use the device name such as "C:".



Setting head parking parameters

Use hdparm, versions for both *NIX and Windows exists.

hdparm -S 252 <device>

Retrieving firmware version

Use the following to retrieve the firmware version:

# smartctl -a <device> | grep Firmware
Firmware Version: CC1H

Is there some way to create a structured table here btw? I failed using the

input.

Contributions welcomed. I will copy/paste from your replies!

 

[parityboy]

Good call. OK, here's my contribution:

Samsung - F4 EcoGreen HD204UI- 1AQ10001 - 3Gb/s - 32MB - YES (TLER) - YES (LCC) - Will not survive a restart

Tested with smartmontools 5.3.8.

 

[sub.mesa]

Samsung should support CCTL, not TLER, right? Unlike TLER, CCTL does not survive a reboot so must be set on every boot (either by the OS or by the hardware RAID controller).

 

[parityboy]

Correct, but since the OP used TLER as a synonym for CCTL and ERC, so did I.

 

[rmd3003]

Samsung should support CCTL, not TLER, right? Unlike TLER, CCTL does not survive a reboot so must be set on every boot (either by the OS or by the hardware RAID controller).

And how is it done? Let's say I reboot server and then.... ?

 

[Metaluna]

I have been looking at some new drives, especially 2TB and 3TB. It can however be very hard to find information whether tools support setting TLER / ERC / CCTL and/or if the ability exists to modify LCC.

I think a thread like this is a great idea.

You should clarify what "modify LCC" means though. I assume you mean modify the head park timeout value, rather than modify the load cycle count counter itself. That would be kind of useless (akin to turning the odometer back on your car).

 

[nowwhatnapster]

I think a thread like this is a great idea.

+1 to that. I will make a contribution tonight for the western digital drives. I did not onow what the competitors equivalent of tler was.

 

[rmd3003]

Ha, I just tested my green WD drives on my WHS and 4 x 2TB and 4 x 1TB all support TLER. Yeeess.

 

[parityboy]

@rmd3003

Could you specify the model please? Thanks.

 

[rmd3003]

@rmd3003

Could you specify the model please? Thanks.

Western Digital - Caviar Green WD20EADS -3Gb/s - 32MB - YES - YES
Western Digital - Caviar Green WD10EACS -3Gb/s - 16MB - YES - YES

 

[parityboy]

@rmd3003

Cheers for that. For those drives for whom the ERC/CCTL setting does not survive a reboot, you'd have to run a script to perform the task on start up.

On Linux it would be something like:

smartctl -l scterc,70,70 /dev/sd*

Where * is the letter of a given drive.

On Windows it'd be a .bat file - I don't know where you'd put it.

smartctl -l scterc,70,70 C:

Where C: is a drive, or possibly a RAID array. smartctl has pass-through modes for some hardware and firmware RAID controllers. A more natural method would be to tell the RAID controller directly, via its management software - hardware RAID controllers should support this, but I suspect that many do not.

 

[maxm]

For windows the device is named pdx where x is the physical device number.

With 3ware controllers you can use pdx,y where y is the 3ware port number.

 

[mhakali]

Western Digital - Caviar Green WD20EADS -3Gb/s - 32MB - YES - YES
Western Digital - Caviar Green WD10EACS -3Gb/s - 16MB - YES - YES

Cool!

Got a firmware version for them? Is there a tool compatible to set the values permanently?

 

[mhakali]

Copy / pasted some of your infos, and completed some more. Will add firmware versions to my drives soon. Thanks for the encouragement from many!

 

[parityboy]

@mkahali

You should update the entry for the Samsung F4EG 2TB. I've verified that CCTL and head parking values can be set.

 

[mhakali]

@mkahali

You should update the entry for the Samsung F4EG 2TB. I've verified that CCTL and head parking values can be set.

Cool!

Inserted.

 

[drescherjm]

Western Digital - Caviar Green WD20EADS -3Gb/s - 32MB - YES - YES
Western Digital - Caviar Green WD10EACS -3Gb/s - 16MB - YES - YES

I thought WDC disabled this on all drives produced after some date. 2008? 2009?

 

[mhakali]

I thought WDC disabled this on all drives produced after some date. 2008? 2009?

Did some research, and added a note. Only earlier revisions are safe to do this with.

 

[Jamfan16]

On Windows it'd be a .bat file - I don't know where you'd put it.

Could it be placed in the Startup folder?

For windows the device is named pdx where x is the physical device number.

Would this be the correct way to reference drives if I'm using an ICH7 based RAID and Intel's Rapid Storage software? Do you explicitly set the new TLER/CCTL for each drive as a separate smartctl command? Or if I reference C: would the new TLER/CCTL be applied to all drives in the RAID?

Thanks for your help.

 

[maxoptimus]

Could it be placed in the Startup folder?



Would this be the correct way to reference drives if I'm using an ICH7 based RAID and Intel's Rapid Storage software? Do you explicitly set the new TLER/CCTL for each drive as a separate smartctl command? Or if I reference C: would the new TLER/CCTL be applied to all drives in the RAID?

Thanks for your help.

command --scan -d csmi (appear list of drives)
Some like /dev/csmi0,2; /dev/csmi0,3 (according to ports used for raid).

Also latest Seagate Baracuda XT (ST32000641AS; sata3; 64mb); scterc don't supported.

Looks like on next out of space I will bought 8 cheap (Samsung - EcoGreen F4 HD204UI or
Samsung - EcoGreen F3 HD203WI) and cheap 3ware raid card

Currently I use 4 ST32000641AS (raid10 system and raid5 media, games, backup).
Before build raid I made several times full write, verify and read tests. Run more then 6 month currently without problems on ICH9R (x38 Asus P5E), but after read several posts about TLER/ERC when thinking about external raid card, looks like I should expected big problems if some sectors will be reallocated.

 

[parityboy]

command --scan -d csmi (appear list of drives)
Some like /dev/csmi0,2; /dev/csmi0,3 (according to ports used for raid).

What device creates /dev/csmi?

 

[maxoptimus]

http://sourceforge.net/apps/trac/smartmontools/wiki/Supported_RAID-Controllers
http://smartmontools-win32.dyndns.org/smartmontools/

Intel ICHxR RAID
(Intel Rapid/Matrix Storage driver)

Onboard intel RAID. most of mobo use it. (for example my ASUS p5e X38 ICH9R).

 

[Tanel]

There is a lot of uncertainty going around about the Samsung F4 EcoGreen (HD204UI). I thought I'd bring an end to that.


##
# Specifications:
##
Motherboard: ASRock Z68 Pro3-M, BIOS v1.40, BIOS SATA Mode: AHCI
Drives: 3x Samsung F4 EcoGreen 2TB (HD204UI), firmware "1AQ10001" out-of-box, all labeled "2011.07"
Software stack A: Ubuntu 11.04 Desktop Live-CD i386, hdparm v9.32, smartctl 5.40 (r3124)
Software stack B: Ubuntu 11.10 Desktop Live-CD i386 (daily build 2011-09-04), hdparm v9.37, smartctl 5.41 (r3365)


##
# Performing the checks:
##
Setting LLC (Load_Cycle_Count):

root@ubuntu:~# hdparm -S 1 /dev/sd{a,b,c}

/dev/sda:
 setting standby to 1 (5 seconds)

/dev/sdb:
 setting standby to 1 (5 seconds)

/dev/sdc:
 setting standby to 1 (5 seconds)

Checking LLC:

root@ubuntu:~# DRIVE=sda	# Test one drive at a time
root@ubuntu:~# SLEEP=4		# 4 should not increase Load_Cycle_Count, 6 should
root@ubuntu:~# for i in {1..10}; do smartctl -a /dev/$DRIVE | grep Load_Cycle_Count; done; sleep $SLEEP; for i in {1..10}; do smartctl -a /dev/$DRIVE | grep Load_Cycle_Count; done;
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       15
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       15
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       15
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       15
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       15
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       15
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       15
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       15
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       15
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       15
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       15
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       15
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       15
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       15
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       15
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       15
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       15
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       15
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       15
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       15
root@ubuntu:~# SLEEP=6
root@ubuntu:~# for i in {1..10}; do smartctl -a /dev/$DRIVE | grep Load_Cycle_Count; done; sleep $SLEEP; for i in {1..10}; do smartctl -a /dev/$DRIVE | grep Load_Cycle_Count; done;
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       16
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       16
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       16
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       16
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       16
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       16
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       16
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       16
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       16
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       16
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       17
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       17
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       17
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       17
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       17
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       17
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       17
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       17
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       17
225 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       17

LLC: OK!
LLC with soft reset: OK!
LLC with hard reset: Reset to default!


Setting SCT Error Recovery Control, Z68 native 6Gbps:

root@ubuntu:~# smartctl -l scterc,70,70 /dev/sda
smartctl 5.40 2010-07-12 r3124 [i686-pc-linux-gnu] (local build)
Copyright (C) 2002-10 by Bruce Allen, http://smartmontools.sourceforge.net

Error Write SCT Error Recovery Control Command failed: scsi error aborted command
Warning: device does not support SCT (Set) Error Recovery Control command
Suggest common arguments: scterc,70,70 to enable ERC or sct,0,0 to disable

SCT Error Recovery Control w/ 6Gbps: FAIL!
SCT Error Recovery Control w/ 6Gbps and soft reset: FAIL!
SCT Error Recovery Control w/ 6Gbps and hard reset: FAIL!


Setting SCT Error Recovery Control, Z68 native 3Gbps:

root@ubuntu:~# smartctl -l scterc,70,70 /dev/sdb
smartctl 5.40 2010-07-12 r3124 [i686-pc-linux-gnu] (local build)
Copyright (C) 2002-10 by Bruce Allen, http://smartmontools.sourceforge.net

SCT Error Recovery Control:
           Read:     70 (7.0 seconds)
          Write:     70 (7.0 seconds)

SCT Error Recovery Control w/ 3Gbps: OK!
SCT Error Recovery Control w/ 3Gbps and soft reset: OK!
SCT Error Recovery Control w/ 3Gbps and hard reset: Reset to default!


Setting SCT Error Recovery Control, Z68 native 3Gbps eSATA:

root@ubuntu:~# smartctl -l scterc,70,70 /dev/sdc
smartctl 5.40 2010-07-12 r3124 [i686-pc-linux-gnu] (local build)
Copyright (C) 2002-10 by Bruce Allen, http://smartmontools.sourceforge.net


Checking TLER/CCTL/ERC:
SCT Error Recovery Control:
           Read:     70 (7.0 seconds)
          Write:     70 (7.0 seconds)

SCT Error Recovery Control w/ 3Gbps eSATA: OK!
SCT Error Recovery Control w/ 3Gbps eSATA and soft reset: OK!
SCT Error Recovery Control w/ 3Gbps eSATA and hard reset: Reset to default!


##
# Notes:
##
1) According to Samsung, disks manufactured December 2010 or later include the firmware patch that resolves the issue with possible data loss. I wanted to be sure so I still flashed my drives. The updated firmware did not change any of the above results. More info about the patch and what it is to fix: http://sourceforge.net/apps/trac/smartmontools/wiki/SamsungF4EGBadBlocks

2) As I couldn't get the SATA 6Gbps to work with SCT Error Recovery Control I hoped it would be solved with an updated software stack. However it made no difference.

3) I have come across these (valuable) resources for TLER/CCTL/LLC:

• HardForum: link
• StorageForum: link
• Habrahabr (Russian blog?): link

 

[MRAB54]

I read this thread a week too late, purchased three ST300DM001 3TB seagate drives for use in RAID 5 with LSI 9260-41. I've opened two of the drives, plugged them in and smartmon reports that they do no support erc

Makes me sick to my stomach. Is this really going to be a problem for me? I want reliability more than anything. Should I send the one drive back, eat the other ones? Are there any settings in the RAID controller that could minimize any of these issues?

 

[hotcrandel]

If you're serious about reliability, I would. I've had such bad experiences trying to use consumer drives with RAID cards that I always just pay out the nose for enterprise drives..

Did a firmware update to a 3ware card that hosed a bunch of pre-firmware update, supported drives, post firmware, no support and had a big fiasco on the array.

I have very few/no problems with ultrastars and so far even deskstars, as well as WD RE disks.

 

[pingouyn]

I have read on forum that it would be possible to set the CCTL value (and maybe once and for all!) on the HD204UI HDD from Samsung. The people who did this used a DOS software called HDAT2.

Has anyone heard of it or tried to set the CCTL on Samsung's impressive F4EG 2Tb drive in any other way than using HDAT2?
I have read this forum thread and there is mention of solutions for Linux, unfortunately nothing for Windows.

If anyone with an adventurous mind and a passion for problem solving want to try the HDAT2 approach, report the findings and the "how to" on this forum, that would be smashing indeed!

 

[Red Falcon]

Excellent thread, I can't believe I didn't see this until now.
Keep up the good work!