just to point that out:
bash-4.1# smartctl -a /dev/sda | grep Load_Cycle
193 Load_Cycle_Count 0x0032 056 056 000 Old_age Always - 434823
bash-4.1# smartctl -a /dev/sdb | grep Load_Cycle
193 Load_Cycle_Count 0x0032 058 058 000 Old_age Always - 426882
bash-4.1# smartctl -a /dev/sdc | grep Load_Cycle
193 Load_Cycle_Count 0x0032 056 056 000 Old_age Always - 434959
Model Family: Western Digital Caviar Green family
Device Model: WDC WD5000AADS-00S9B0
All three drives have reached the “minimum”, which is 300 000 for these discs. The Power-On-Hours are around 5770 hours (which is, if i calculated correctly, 240 days, 8 months = less than a year) - However, smart checks show no error, the discs are still running fine.
You might want to add settings like this to your /etc/smartd.conf:
/dev/sda -a -o on -S on -m wdp -s (S/…/…/./07|L/…/…/6/08)
/dev/sdb -a -o on -S on -m wdp -s (S/…/…/./07|L/…/…/6/08)
/dev/sdc -a -o on -S on -m wdp -s (S/…/…/./07|L/…/…/6/08)
just google a bit to see what which params are doing; in general these lines are checking your discs (sda, sdb and sdc in this case) regurlarly. You’ll see something like this if you run smartctl -a, near the bottom of the output:
SMART Self-test log structure revision number 1
Num Test_Description Status Remaining LifeTime(hours) LBA_of_first_error
1 Extended offline Completed without error 00% 5780 -
2 Short offline Completed without error 00% 5777 -
3 Extended offline Completed without error 00% 5773 -
4 Short offline Completed without error 00% 5771 -
So, currently nothing to worry about i guess. However - Just to make sure, you might consider setting up a raid with those discs. I did the wdidle3.exe thing yesterday, worked fine with that discs. So, yes, here it did work with the Green Caviar ones. And another side note: This is not dedicated to WD Discs; some other vendors and especially notebook/netbook discs, show the same behaviour to save power, and the linux kernel (probably, im not sure) due to dirty writeback is waking the discs up every 30 seconds. Remember, these discs are made to be energy-saving/efficient, so your real job is to make your OS to let them save as much power as possible. Means: Enable laptop mode, Disable atimes, set good powersaving, disable frequent logging or put it onto a ramdisc, just try to make your discs as much idle as possible.
So things, which you might also try:
take a look at your /etc/fstab and add “noatime” to the disc(s)/partition(s), for example:
/dev/sda5 / ext4 defaults,noatime 0 0
This won’t update the atimes everytime, thus your disc can sleep longer.
This will try to merge some writes, thus the chances for your disc to sleep increase.
echo “1” > /proc/sys/vm/laptop_mode
This causes the system to clear the cache of the discs (if i understood it correctly) every 5 seconds. Higher it to 10 minutes but be aware: If your system crashes, the last 10 minutes arent saved.
echo “x” > /proc/sys/vm/dirty_writeback_centisecs
replace X with your value (remember, its in centisecs)
disable hal-polling, i saw that the wd discs got a jumper to enable/disable powersaving, maybe you have to set the jumper (if hdparm -B and -S aren’t working) to get more powersaving, put your logs onto a ramdisc, which you write back to the real fs when you reboot or shutdown.
Theres much stuff you can do, to have your discs saving power - Just think about it, the real cause for this problem is that the linux kernel is waking the discs up too often and the wd discs are parking the heads too early. So you can argue both: it’s the linux kernel’s/distributions/package-fault and it’s the fault of wd discs with their (way too) low value of 8 seconds. But: You bought those green stuff to save power, so why aren’t you tuning your system to do so?