A few weeks ago I ordered myself a new eco friendly home server. The machine will be acting as my content server: running samba, sabnzbd, nginx and mysql. Most of the time it will be idling, so my goal was to build a server with energy saving hardware. The old server already had a 1TB Western Digital Green Cavair drive in it which I bought in 2010. So I placed an order for the following parts:
- MSI H61M-P22 motherboard
- Intel Pentium G620
- Transcend JetRam JM1333KLN-2G
- be quiet! Pure Power L7 300W
Within a few days I received the whole shipment, and a few hours later my server was up and running again :). After installing and configuring Debian Squeeze I measured the power usage: 22 ~ 25 watt idle, not bad! I didn’t tweak anything at all, so I started my journey. Spinning down the disk after a period of idling was my next goal. For Linux users, hdparm is the tool which gets the job done. On Debian or Ubuntu all you need to do is:
|
1 |
# apt-get install hdparm |
To configure your drive you have to know its name. One way to figure that out is
|
1 |
# fdisk -l |
Output on my system:
|
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 |
# fdisk -l Disk /dev/sda: 1000.2 GB, 1000204886016 bytes 255 heads, 63 sectors/track, 121601 cylinders Units = cylinders of 16065 * 512 = 8225280 bytes Sector size (logical/physical): 512 bytes / 512 bytes I/O size (minimum/optimal): 512 bytes / 512 bytes Disk identifier: 0x0c8b0915 Device Boot Start End Blocks Id System Disk /dev/sdb: 16.0 GB, 16022241280 bytes 255 heads, 63 sectors/track, 1947 cylinders Units = cylinders of 16065 * 512 = 8225280 bytes Sector size (logical/physical): 512 bytes / 512 bytes I/O size (minimum/optimal): 512 bytes / 512 bytes Disk identifier: 0x000b2232 Device Boot Start End Blocks Id System /dev/sdb1 * 1 1863 14957568 83 Linux /dev/sdb2 1863 1948 686081 5 Extended /dev/sdb5 1863 1948 686080 82 Linux swap / Solaris |
The /dev/sdb device is the USB key running the OS, while /dev/sda is my WD Green Caviar drive.
To set the spindown time all you have to do is:
|
1 2 3 4 |
# hdparm -S 5 /dev/sda /dev/sda: setting standby to 5 (25 seconds) |
Of course you should change the time out if you want. However, after waiting 25 seconds I did a status check:
|
1 2 3 4 |
# hdparm -C /dev/sda /dev/sda: drive state is: active/idle |
What the? 🙁 It obviously didn’t work! After trying some more time outs I concluded it didn’t really work. Googling for my drive and hdparm I quickly found a lot of other people which ran in the same problem. Furthermore I discovered that Linux and the old Western Digital Green Caviar drives don’t play well with each other. To summarise: the drive puts heads into parking position after 8 seconds idle time! This causes a very high Load_Cycle_Count. To check this, you have to install the smartctl utility:
|
1 |
# apt-get install smartmontools |
Then, check the S.M.A.R.T. data for you drive:
|
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 |
# smartctl -a /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 ... Vendor Specific SMART Attributes with Thresholds: ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE 1 Raw_Read_Error_Rate 0x002f 200 200 051 Pre-fail Always - 0 3 Spin_Up_Time 0x0027 105 088 021 Pre-fail Always - 7741 4 Start_Stop_Count 0x0032 100 100 000 Old_age Always - 453 5 Reallocated_Sector_Ct 0x0033 200 200 140 Pre-fail Always - 0 7 Seek_Error_Rate 0x002e 200 200 000 Old_age Always - 0 9 Power_On_Hours 0x0032 088 088 000 Old_age Always - 9144 10 Spin_Retry_Count 0x0032 100 100 000 Old_age Always - 0 11 Calibration_Retry_Count 0x0032 100 100 000 Old_age Always - 0 12 Power_Cycle_Count 0x0032 100 100 000 Old_age Always - 399 192 Power-Off_Retract_Count 0x0032 200 200 000 Old_age Always - 40 193 Load_Cycle_Count 0x0032 187 187 000 Old_age Always - 41449 194 Temperature_Celsius 0x0022 117 100 000 Old_age Always - 30 |
When issuing this command for a couple of minutes, I saw the number growing rapidly: every 3 minutes a new hit. Not good! 🙁
Fix the drive for Linux usage
Western Digital published a advisory about this problem. Basically we as Linux users are left in the dark. They provide a tool to reset of reconfigure the timeout, but it only runs on MS-DOS…
The good news is that the utility is present on the Ultimate Boot CD. Burn the ISO on a disc (or make a bootable USB key) and remove the timeout with:
|
1 |
wdidle3.exe /D |
After that you’re drive will pay attention to the settings provided by hdparm, and the Load_Cycle_Count won’t be growing that rapidly. The count on my server grows by 2 counts per day, instead of ~ 200! 🙂 And when the drives is standby my server consumes 18 ~ 20 watt!