Problem/Justification
I want to spin down disks over night / long periods of inactivity to save power. Energy is really expensive here in Europe, thus saving power over night / weekends is desirable.
The toggle for HDD standby is already available in Truenas Scale, however, it no longer functions since release 2025.10. The issue is heavily discussed on the forum HDD Sleep/Spindown/Standby and a corresponding bug report exists: Jira
Impact
(How is this feature going to impact all TrueNAS users? What are the benefits and advantages? Are there disadvantages?)
Since the toggle to activate/deactivate HDD standby does already exist in the Truenas UI, this feature request has low impact on the existing user base.
Benefits of re-introducing HDD standby is basically energy savings (saving costs and doing positive impact to the environment). Modern HDDs are specified for high load/unload cycles, so there are no real disadvantages if the idle timeout is chosen as a sensitive value, besides longer access times if the disks need to activate from a spun down state.
User Story
(Please give a short description on how you envision some user taking advantage of this feature, what are the steps a user will follow to accomplish it)
A user planning to save power in periods of inactivity enables HDD standby in Truenas Scale disks section (a setting per pool would be worth considering), and selects a timeout to specify the period of inactivity in minutes after which the disks go to standby.
@Jaywalker
Have you generated any scripts that would perform this function and then make a CRON Job for running the scripts? I ask since it “should” be fairly simple to do. You will find out that the main problems are going to be and drive activity would pull them out of the deep sleep and they “could” be spinning up/down quite often.
This could be harmful to the drives life, but generally does not impact the warranty. The warranty being say 3 years, but many drives last over 7 years, not spinning up and down all the time.
If you have not already done so, I would suggest running a BASH script to make the drives sleep. then see how it goes. Assume your Feature Request sits here for 6 months before it is either accepted or declined. You can make your drives sleep in the meantime.
thank you for the hint. However, Truenas 25.10 has introduced an issue (supposedly with temperature monitoring?) that is frequently waking up all disks - even those not being part of any pool. Hence, the solution via scripts / cron no longer works. Manually putting the disks to standby will only last for a few minutes until Truenas wakes the disks up again - without any data activity.
This approach is not feasible for many home lab users. Many opt for an all-in-one solution to reduce costs, meaning the gateway might be located on the TrueNAS, and the network would be interrupted when the system is shut down.
Even without considering whether the gateway is running on TrueNAS, many people run dozens of containers and several virtual machines on TrueNAS. When TrueNAS is shut down, these services will also be interrupted, which is equivalent to amplifying a simple hard drive hibernation problem into a problem of large-scale service interruption.
The above two post are going off-topic: This is a feature to spin down drives with the NAS on, not to shut down the NAS. Thank you for reading.
This is generally advised against anyway: Keep your basic network infrastructure independent.
Do you want to shut down your whole network to deal with failing drives in the NAS?
And what about those who, er, just run storage on their Network-Attached Storage? It’s in the name…
The original post is about the benefit of spinning down the disks to save energy.
I am still on 24.04 and considered the feature (but I am not using it)
spinning down wears out the disks enormously. From what I was able to find - disks are rated to about 150000 start ups. This may seems a lot but on a NAS (again - my understanding is) there are numerous operations that demand access to the disks. This makes the spin down not viable because it is not possible to leave the NAS running and spin down the disks (and mandate them to be in a spun down state). Ultimately they will spin up and down again and contribute negatively to the life of the drives.
The concerns above are significant in data preservation and cost consideration so for me the feature is a no go. Even if it is there - I’d turn it OFF. I see no benefit of it.
I am doubting how big power savings would be because the electrical companies typically have a set charge of power availability. Even if the user do not use any electricity they will still be charged a certain amount a day because the company delivers the power to the home (regardless how much is actually used).
The power savings are not enough to justify a (potential) premature hard disk failure.
If the goal is to save energy - a strategy around power off/on of the compete system would be better because it would be limited to 1 shut down and 1 start up (as an opposite of many - because it is not possible to restrict the server to keep the disks spun down for the entire duration of the time).
365 days per year. 3 times that is a little over 1000, so back-of-the-envelop calculation goes that 150k spin-ups amount to 3 spin-ups a day for 150 years, or 30 spin-ups a day (over one per hour) for 15 years.
The realistic scenario where the backup NAS spins up once or twice a day to receive a replication stream is very safe with respect to specifications.
Three bullet points to say the same thing. No figure.
Back-of-the-envelop again: Saving 5 W per drive, 23 hours a day is 1.15 kWh/day for 10 drives. Your utility company puts a price on that.
I guess this comes down to drive brand and actual number of spins. Some can handle more others less. No point to quote AI here but each person who is interested can do their own research. At the end - it is a big deal for some and not such a big deal for others.
If you can control this - I can see your point. I am doubtful however that we can. Either way - I can’t answer. This assumption can be realistic or rather optimistic.
The way how I see it - 5 W x 10 drives = 50 W/h. Our light bulbs used to be 60 W for a single bulb. Not that you would burn it night and day but … just a thought.
Electricity is not free - I get that. But even if you are to plug a heather or an AC - it all consumes it - at a much bigger rates. And let’s remember that the utility company does charge the fee regardless if you consume or not.
I don’t think it is worth it to dig more in this (in terms of saving vs wear and tear). I do get it that some people can see a potentially a big benefit. I don’t see it - at least for a house hold with 1 NAS and a handful of drives. I can see it in a data center with hundreds or thousand of drives but if that is the case - it is completely different story. And I don’t think we are talking abut a data center here.
@vbs@etorix
My perspective is this: Your drive is under warranty for 3 or 5 years, in general. During that time the company states if the drive fails to spin-up and the count is below 150,000 then it will warranty it. Realistically this is a sales pitch. Reaching 150,000 is likely impossible. You know how these are calculated, you take 20 drives, run them for 342 hours, and calculate the failure rate. It is an example, I didn’t state this is exactly what they do.
Where the spinning up/down hurts is when the warranty is gone. Now you want to protect your assets, leave them running to minimize current surges to the electronics.
Anyway that is all I have to say. I’m muting this thread. Have fun!
a) you need to set a timeout of inactivity only after which the drive will spin down. Set it to 60 min and you are limited to 24 spinup-/down cycles per day maximum.
b) usecase: the example above with a pool receiving one backup every night is exactly my scenario. I run this since years with spin down and monitor the load cycle count of the disks. As one would expect, it is increasing by 1 every day. Once per month a scrub is started which adds one additional spin-up.
I don’t think that I will ever reach the 300.000 load cycles the drive is specified for even closely and have zero concerns w.r.t. the health or lifetime of the drives.
I don’t understand all the concerns here. For people who don’t want spin down, leave it switched off.
Fully agreed. The big difference I see is between homelab and professional usage.
My NAS is idle most of the time since I’m not always at home. Services and frequently accessed data are on SSD drives, while traditional hard drives remain idle most of the time. They are sometimes spun down for days. With 7 drives and an average electricity cost of €0.35 per kWh, this makes a noticeable difference in the power bill for homelab users over the course of a year.
For a homelab, I have no concerns about disk wear. Reaching 150k–300k load cycles is not that easy (at least not if you set reasonable spin-down parameters).
Because of the continuous spin-ups, I reverted to version 25.04.2.6, which does not show the issue mentioned. I’m looking forward to this problem being fixed so I can upgrade!
Everyone should be free to decide whether disks should spin down or not. Especially in homelab environments, requirements can vary significantly from one setup to another.
Yes, it is! If you spin up your drives once a day, you’ll hit 150k load cycles in only 410 years! Don’t you want your drives to last longer than that?
The issue for myself is not just the cost, its the noise. I live in one room, I have two servers with spinning drives, one spins down happily the other does not unless I run an old version that used to work fine or I apply patches to the latest version.
I’m on TrueNAS 25.04 and tinkered around with HDD sleep some time ago. It worked for the most part. However, TrueNAS has some built-in periodic maintenance tasks, executed at system start and every 24h after that, which wake up all the HDDs. See Timing of Periodic Mainteance Tasks on TrueNAS Scale for details.
If HDD sleep is fixed, it would be nice to also take a closer look at these maintenance tasks. I have a setup where my HDDs would only wake up once a week. I’m sure there are a lot of other users who would benefit from this as well.
Edit:
I just learned that this is because of ZFS dataset encryption.
If you are using ZFS dataset encryption, the disks will spin up every 24hours as the sync_db_keys job triggers a write to the pool.
This is just fear mongering. I did not care about power consumption too much but calculating things puts things into perspective. M NAS adds up to about 20 percent of the monthly power consumption and lowering it would provide tangible cost savings.
As for the functionality, I’m also “stuck” on 25.04, initially not intentionally, I just didn’t update, but with all the comments regarding smart reporting and non functioning spindown it made me wait it out y bit more. I actually actively use spindown. My case is well controlled. I spin down the two backup mirrored drives that only receive daily snapshots from my primary data drive which is always on. Why would I want the two drives to always consume power when I access them once per day? It makes no sense.
