Is the “idle3” timeout mess still relevant on Western Digital WD Blue hard drives from 2022?

Question

THE QUESTION:

Is the “idle3” timeout still relevant for 2022 Western Digital Blue hard drives, and specifically for their WD40EZAZ line? Or should I take hdparm -J reporting an invalid exchange as proof that the timer was removed and that the disks won't park themselves in a silly short amount of time?

Background:

So, for a long time, Western Digital shipped their “Green”, later (I think) relabeled as “Blue” hard drives with an unusual feature known as “idle3 timeout” (aka “IntelliPark”) which caused the drives to park its heads and enter a low power consumption state after inactivity. This in itself probably isn't a bad thing, but the factory default timeout value was set to the absurdly low value of 8 seconds, meaning that if your drive is accessed every 10 seconds on average, it will keep parking its heads and eventually failing. Symptoms include an abnormally high Load_Cycle_Count SMART attribute and (presumably) eventual premature failure of the drive.

Western Digital (as far as I can tell) told people to use some other of their color lines if they had this kind of activity, but they did release a (shudder) MS-DOS utility called wdidle3 which would allow the timer to be set to a different (i.e., sane) value. (Also, they changed their mind at least once about what the raw values meant, causing more confusion.) This utility was reverse-engineered (I guess — or maybe they decided not to be evil and documented what it did) and its functionality was implemented in the Linux tool idle3ctl (of idle3-tools) and the -J flag of hdparm.

This is not ancient history: I have two 4TB Western Digital Blue hard drives, model number WD40EZRZ-22GXCB0, which were bought in 2020-09, which implemented the idle3 timer and I think the default value was set to a few seconds.

However, a few weeks ago, I bought another 4TB Western Digital Blue drive, model number WD40EZAZ-00SF3B0 this time, and neither hdparm -J nor idle3ctl are able to access the idle3 timer.

• On the 2020 drives having the feature (for comparison):

vega david ~ $ sudo hdparm -J /dev/sdb

/dev/sdb:
 wdidle3      = 300 secs (or 13.8 secs for older drives)
vega david ~ $ sudo idle3ctl -g /dev/sdb
Idle3 timer set to 138 (0x8a)

• On the 2022 hard drive:

pleiades david ~ $ sudo hdparm -J /dev/sdb

/dev/sdb:
SG_IO: bad/missing sense data, sb[]:  70 00 05 00 00 00 00 0a 04 51 a0 10 21 04 00 00 00 4f 00 00 00 00 00 00 00 00 00 00 00 00 00 00
SG_IO: bad/missing sense data, sb[]:  70 00 05 00 00 00 00 0a 04 51 a0 1b 21 04 00 00 00 b0 00 00 00 00 00 00 00 00 00 00 00 00 00 00
 wdidle3      = disabled
pleiades david ~ $ sudo idle3ctl -g /dev/sdb
sg16(VSC_SENDKEY) failed: Invalid exchange

Now I don't have an abnormally high Load_Cycle_Count after a few hours' use, so I'm tempted to believe the idle3 timer was scrapped altogether and I have nothing to worry about, but maybe the disk was just in constant use (as its part of a RAID array it's hard to be sure), and this Amazon reviewer claims to have the same problem (in 2017) of idle3ctl returning “Invalid exchange” despite the drive having “IntelliPark”.

Answer 1

THE QUESTION: Is the “idle3” timeout still relevant for 2022 Western Digital Blue hard drives,

Short answer

Yes, the features still exists and can be modified on a couple but not all WD drives. Monitoring can be done by checking how the raw value of the SMART attribute 193 Load cycle count increases over time given your personal load profile (the way how you stress your drive based on the way you are using it)

Long answer

Yes it is, because the feature was available on two different drives (WD40EZX and WD40PURZ) I bought in 2022.

and specifically for their WD40EZAZ line?

I don't know. WD40EZAZ is a SMR drive. I do not trust this overpriced technology in my environment that does not contain an uninterruptible power supply.

Or should I take hdparm -J reporting an invalid exchange as proof that the timer was removed and that the disks won't park themselves in a silly short amount of time?

The possible removal of that timer does not imply that the drive firmware still contains an aggressive head parking strategy. This is something you have to monitor.

There are some warnings in the hdparm manual about its use:

WD supply a WDIDLE3.EXE DOS utility for tweaking this setting, and you should use that program instead of hdparm if at all possible. The reverse-engineered implementation in hdparm is not as complete as the original official program, even though it does seem to work on at a least a few drives.

Source: https://man7.org/linux/man-pages/man8/hdparm.8.html

Instead of bothering using the WD-Tool in a real DOS environment, I successfully used the linux idle3-tools provided by Christophe Bothamy.

https://idle3-tools.sourceforge.net/

I experienced that the drive has to be connected as close as possible to the mainboard chipset, for instance by connecting it to one of the main ATA connectors. Otherwise error message would lure you into believing that there is no value that could be modified.

List of WD drives where the parking delay time can be modified

size	model	orig. value	time*
3,5''	WD40EFZX	?	?
3,5''	WD40EZRZ	?	?
3,5''	WD40PURZ	?	?

Table to be extended, time figure based on -g105 parameter in idle3ctl