Nice one, thank you. I've tried to simplify knowing if drives have powered down, by plugging an energy monitor socket adaptor into my UPS. It's quite nice to view a power app to see what's running and what's not.
Video helped me to figure out the system dataset part, disks would not spindown and kept being accessed even when power saving settings were applied. Thanks for the help!
Superb video, praying that this spinning down of discs may elongate the life of my HDDs, they are costly to replace in a home use case. Thanks for the guide!!!!
i was specking out a NAS and bulding it using mostly old parts so this was a big help. I am already doing a detune on a ryzen 2700 so that it is more energy efficient but realized most my wattage will come from spinning drives after my tuning.
Nice video, thank you. The terminal command is really handy. BTW, are you concerned about drive longevity? They say always-spinning disks live longer... But standby mode is actually pretty nice from the acoustic point of view. They say power consumption is not that different.
I've yet to see some really concrete data on how much it reduces drive life. Most of what I see and read is peoples personal experience. We're not able to predict when a drive will fail, so it's hard to do an A/B test to see if spinning them up and down wears them out faster. I'm sure it does, but I suspect the amount is probably negligible. I will say that standby mode doesn't always make sense. It really depends on your environment. If you're in an environment where a lot of people are accessing the NAS frequently, it would not make sense to spin them down as they would constantly be spinning back up. I'm the only person using my NAS, and my drives are accessed infrequently enough that I'm comfortable that spinning them down is what I want to do. In general I'm not as concerned about wearing out my drives as others are, as my strategy is to use cheap drives and overcompensate for failure. I have multiple layers of redundancy against total pool failure, and even if it does fail I have multiple known good backups. I can live with shaving a few hours off the life of my drives (if that's even happening). If you're concerned about power consumption, I'm working on a video right now to test if updating my existing power supply helps to save some money. I need to run on the new supply for 30 days to truly see how much of a difference it makes, so expect to see that drop sometime next month. If you're interested, make sure you subscribe so you don't miss it!
@@BlackbeardSupport wow, sounds great. Looking forward to the video! Will you measure how much Amperes your NAS consume during the start-up when all drives start? It seems to be quite critical for a power supply because in general during a NAS operation the load is minimum but first seconds of start-up lead to big power peaks. Would be nice to hear your thought in the video. :)
Unfortunately it's not the direction we'll be going in. I was more interested to see if upgrading my 10 year old power supply to a modern, more efficient supply would improve power consumption. Peak draw wasn't on my radar as to get the most out of the efficiency curve for a power supply, you want to over provision by quite a lot. Then the power supply always operates as efficiently as possible. Once I see if we draw less power from the wall over the next 30 days, I'll be able to determine if it was worth upgrading, and how much I would need to save on electricity before it was worth it. Because you asked, I rebooted my NAS which currently has 14 drives in it. According to the smart plug that measures the draw from the wall, it peaked at about 360w on boot as everything started up. I would not rely too much on this though, as the plug does not have the granularity I would like to see for measuring such as short burst of power. It is more useful to give a reasonably accurate usage over time (or at least, accurate enough for what I was looking at). Related to your earlier question though, the difference between all my drives spinning vs sleeping seems to be somewhere around the 20-30w mark if they are doing nothing but idling (more, if they are actively working). If they ran at idle for a full month vs being spun down, it would mean I save approximately 20kWhs a month. This of course assumes I do not use the NAS at all for the full month, so in reality I am saving much less. Where I live 20kWhs would cost something like €4, or $4.50 in US dollars. Most users argue that this is pittance in comparison to having to buy a new drive because of the 'extra wear' spinning the drives up and down. They are right, the money saved is much less than would be spent on a new drive, but as I've said before, I'm not convinced it increases the chance a drive will fail early. So far, none of mine have had issues in the last ~5 years, and I spin them down all the time.
cheers this helped me greatly as i followed all instructions in video and hdd still wouldnt stop after 20min. changed smart to power mode standby- perfect mate, works!
@@BlackbeardSupport hi mate, ive just checked my disks again and they arent stopping anymore. did all the steps again even though they were all the same still from last time but i cant get the 00 code (just FF everytime) after timeframe has elapsed anymore. any ideas?
@@todd1748 Hmm, are you sure there could be nothing accessing the disk? It can be tricky to track down what's keeping them up, but no jails? No reports? Cron jobs?
Your videos are great and very helpful. But is there any way you could adjust the audio levels in them? I need to have my speakers maxed out to hear you clearly if there is any background noise when I am watching.
Hi Blackbeard, thanks for the video! Good to know about "Force HDD Standby" being a bad idea. That said, were you able to get temperature monitoring working with these settings? Even when my drive are awake, I now can't monitor their temperatures with the settings you described. Any ideas?
Unfortunately if you spin down disk temperature monitoring stops. You must either use 'Force HDD Standby' and accept the consequences, or never set the disks to standby at all. Sorry!
@@BlackbeardSupport How odd! Okay, so if I enable "Force HDD standby", I understand that SMART checks won't occur while the drives are asleep. Will periodic scrubs still occur?
As A home user, my system is only used in the evening mostly. I have been using spindown for a long time and it cut the watt draw of my nas dramatically. Data set is saved on ssd fast pool. Is there really room on the boot ssd for the data set?
Great video and thank you for it! My question is about how much drive space the system dataset will need. I'm using an 8GB SATA DOM as my boot "drive". My storage pool is 14TB spread over 20 1TB drives (big SuperMicro 4U rig but was a free server, obviously not new). 2 hot spares. From what I've read, the system dataset is generally in the megabyte range, tiny. I'm a NAS noob so I don't even know how to check how much room is available on the DOM. I want to say TNC uses about 2GB but I'm pulling that from weak memory. I do have a pair of 128GB SSD's that were used as cache drives when this machine was originally set up with some horrible NAS software called OpenE and a hardware RAID. They aren't being used at all right now, I could add a pool to them. OK to use the DOM for the system dataset or set up a new pool on the SSD's? I'm using this behemoth at home so it sits many hours or even sometimes days without pulling anything from it.
Yes! Definitely, boot drive is always very tiny. The only advantage of having a large one is that it spreads the writes across the memory to decrease wear and increase lifespan. However, 8GB should last you many years no issues (With the caveat that this is only if the drive itself is well manufactured. Cheap and crappy will not last because it is cheap and crappy, not because of a size issue!).
How do you know what size you need for the system datapool set? I have 2 4TB nas drives in raid 1 and a 128 GB nvme drive as boot pool.. would that be enough? What size does the info take?
It can vary depending on a few factors, but generally 32Gigs is plenty, and 64 is way more than enough. I like to put in a Min of 64, but that's just cause the memory is cheap.
I was planning to make NAS home build, but i could not get to spin down disks (1 pool with 3 disks), in the end i got WD home cloud, and i dont like it cause it also doesn't spin down disks. And now i came across this video and the problem was system data set log on that data set! .. So i move the log to my usbnas key, and i notice that it was slow on configuring disks, saving etc.. and then it crashed NAS in to a boot loop. But it was just nas to play around, no data lost. i presume that its USB 2.0 was the problem. So i will try to add ssd cash device to NAS and use it as syslog device.
Hi Ivan, glad I could point you to the system dataset. It's true in the past that TrueNAS boot pool was recommended to use a USB drive, but in the last few years this is not the best option. A small SSD (Even 60Gig would be more than enough, maybe even smaller would work) would be a great option, and hopefully you will have more success. If you can buy two, it can be a great idea so that you can have a boot mirror, in case one drive fails. You can see how to do this on my video here: ruclips.net/video/xTUsJNvLHS8/видео.html. Best of luck with your NAS.
@@BlackbeardSupport Sorry, to clarify im useing a DAS NetApp shelf for my drives. it seems to be doing something as my power consumptions dropped but it seems i cant use the command to get the disk info
Hmm, yes, I can see how that might be an issue. Unfortunately I don't have a DAS NetApp shelf so I can't test myself, but I wonder if the server doesn't support direct access to disk information. Does the server return any disk information from any command at all, or is it JUST this command that fails?
The fact that my DL380P Gen8 uses 65 watts idle without the spinning drives plugged in, and 100-120 Watts idle when they are in the server makes a huge difference in my electricity bill (0.6 Euro Cents / Kwh // Germany!) so i already made the changes but the server never stopps the drives and nothing is using them. I store ripped movies and series on them so dont need to be 24/7 online By the way i use a virtualized TrueNAS Scale (has qemu-guest-agent) with a dedicated PCI HBA. Dataset set to boot pool, and no force spindown option anymore :/
@@BlackbeardSupport in some cases we need to disable smart monitoring on the drives, so they can spin down. Since acl permissions are a mess and cant connect via nfs to the freshly configured share (after two successful mounts on different things) i decided to wipe this out of my system for ever. And instead i install the Core version again and deal with the ssh and scripting things again... Is there a qemu-guest-agent for the bsd version? Im not familiar with bsd at all.
There's a port available www.freshports.org/emulators/qemu-guest-agent. To be honest I have not tried it myself (I run on bare metal), but I've heard some positive comments from those who have. As with any port, your mileage may vary!
@@BlackbeardSupport I got scale wiped completely, useless garbage, went back to Core and yet, drives are spinning down after a while but the power consumption remains the same :D weird .
@@victorshane4134 Well that is unusual, though if you're using newer drives, they may already by very power efficient. A good ballpark for the consumption being used is about 10Watts. If you have a lot going on on the machine and only a few drives, it could be hard to notice the dip. The other thing to check is that the drives STAY spun down, and don't keep spinning back up for things like daily scrubs. No easy way to check, unfortunately. Just have to occasionally run the command to see they are spun down.
Nice one, thank you. I've tried to simplify knowing if drives have powered down, by plugging an energy monitor socket adaptor into my UPS. It's quite nice to view a power app to see what's running and what's not.
I do the same!
@@BlackbeardSupport Bearded minds think alike!
@@bikerchrisukk Hah certainly!
Video helped me to figure out the system dataset part, disks would not spindown and kept being accessed even when power saving settings were applied. Thanks for the help!
Glad it helped!
Superb video, praying that this spinning down of discs may elongate the life of my HDDs, they are costly to replace in a home use case. Thanks for the guide!!!!
Best of luck! It's quite hard to tell how much use it is!
i was specking out a NAS and bulding it using mostly old parts so this was a big help. I am already doing a detune on a ryzen 2700 so that it is more energy efficient but realized most my wattage will come from spinning drives after my tuning.
damn that system info setting was all i was missing!
Thanks for the info, pretty useful tasks!
this command didnt work for me. i getting a "camcontrol: error sending command" message...
A clear explanation. Thanks!
Glad it was helpful!
Nice video, thank you. The terminal command is really handy. BTW, are you concerned about drive longevity? They say always-spinning disks live longer... But standby mode is actually pretty nice from the acoustic point of view. They say power consumption is not that different.
I've yet to see some really concrete data on how much it reduces drive life. Most of what I see and read is peoples personal experience. We're not able to predict when a drive will fail, so it's hard to do an A/B test to see if spinning them up and down wears them out faster. I'm sure it does, but I suspect the amount is probably negligible.
I will say that standby mode doesn't always make sense. It really depends on your environment. If you're in an environment where a lot of people are accessing the NAS frequently, it would not make sense to spin them down as they would constantly be spinning back up. I'm the only person using my NAS, and my drives are accessed infrequently enough that I'm comfortable that spinning them down is what I want to do.
In general I'm not as concerned about wearing out my drives as others are, as my strategy is to use cheap drives and overcompensate for failure. I have multiple layers of redundancy against total pool failure, and even if it does fail I have multiple known good backups. I can live with shaving a few hours off the life of my drives (if that's even happening).
If you're concerned about power consumption, I'm working on a video right now to test if updating my existing power supply helps to save some money. I need to run on the new supply for 30 days to truly see how much of a difference it makes, so expect to see that drop sometime next month. If you're interested, make sure you subscribe so you don't miss it!
@@BlackbeardSupport wow, sounds great. Looking forward to the video! Will you measure how much Amperes your NAS consume during the start-up when all drives start? It seems to be quite critical for a power supply because in general during a NAS operation the load is minimum but first seconds of start-up lead to big power peaks. Would be nice to hear your thought in the video. :)
Unfortunately it's not the direction we'll be going in. I was more interested to see if upgrading my 10 year old power supply to a modern, more efficient supply would improve power consumption. Peak draw wasn't on my radar as to get the most out of the efficiency curve for a power supply, you want to over provision by quite a lot. Then the power supply always operates as efficiently as possible. Once I see if we draw less power from the wall over the next 30 days, I'll be able to determine if it was worth upgrading, and how much I would need to save on electricity before it was worth it.
Because you asked, I rebooted my NAS which currently has 14 drives in it. According to the smart plug that measures the draw from the wall, it peaked at about 360w on boot as everything started up. I would not rely too much on this though, as the plug does not have the granularity I would like to see for measuring such as short burst of power. It is more useful to give a reasonably accurate usage over time (or at least, accurate enough for what I was looking at).
Related to your earlier question though, the difference between all my drives spinning vs sleeping seems to be somewhere around the 20-30w mark if they are doing nothing but idling (more, if they are actively working). If they ran at idle for a full month vs being spun down, it would mean I save approximately 20kWhs a month. This of course assumes I do not use the NAS at all for the full month, so in reality I am saving much less. Where I live 20kWhs would cost something like €4, or $4.50 in US dollars. Most users argue that this is pittance in comparison to having to buy a new drive because of the 'extra wear' spinning the drives up and down. They are right, the money saved is much less than would be spent on a new drive, but as I've said before, I'm not convinced it increases the chance a drive will fail early. So far, none of mine have had issues in the last ~5 years, and I spin them down all the time.
I've also read online that S.M.A.R.T Will always run and stop your drives from sleeping.
Services > S.M.A.R.T. Settings. Set Power Mode to Standby.
Yea that's definitely true. A lot of settings to tweak.
cheers this helped me greatly as i followed all instructions in video and hdd still wouldnt stop after 20min. changed smart to power mode standby- perfect mate, works!
@@todd1748 Great!
@@BlackbeardSupport hi mate, ive just checked my disks again and they arent stopping anymore. did all the steps again even though they were all the same still from last time but i cant get the 00 code (just FF everytime) after timeframe has elapsed anymore. any ideas?
@@todd1748 Hmm, are you sure there could be nothing accessing the disk? It can be tricky to track down what's keeping them up, but no jails? No reports? Cron jobs?
Really good video!
Glad you enjoyed it.
Your videos are great and very helpful. But is there any way you could adjust the audio levels in them? I need to have my speakers maxed out to hear you clearly if there is any background noise when I am watching.
Great Video :)
I'm so glad you enjoyed them!
what is the system data set and is there a disadvantge to moving it to the boot pool ?
Um how similar is scale to this? I can’t find there to change where the dataset is written too on scale
Hi Blackbeard, thanks for the video! Good to know about "Force HDD Standby" being a bad idea. That said, were you able to get temperature monitoring working with these settings? Even when my drive are awake, I now can't monitor their temperatures with the settings you described. Any ideas?
Unfortunately if you spin down disk temperature monitoring stops. You must either use 'Force HDD Standby' and accept the consequences, or never set the disks to standby at all. Sorry!
@@BlackbeardSupport How odd! Okay, so if I enable "Force HDD standby", I understand that SMART checks won't occur while the drives are asleep. Will periodic scrubs still occur?
@@SeanSL Yea, you'll still get scrubbed. That part will be fine,
@@BlackbeardSupport Good to know! Thanks man 🙂
n1ce .. thanks :)
Nice, guess I don't need ngandrass's github spindown script. Thanks
camcontrol does not exist in TrueNAS scale. Is there a way to check it in TrueNAS scale?
Good!!!!!👏👏👏
Thank you!
As A home user, my system is only used in the evening mostly. I have been using spindown for a long time and it cut the watt draw of my nas dramatically. Data set is saved on ssd fast pool. Is there really room on the boot ssd for the data set?
Does this still work if the disk hardware doesn't support APM?
camcontrol identify | grep "advanced power management"
good job
Thank you! Cheers!
thanks
Great video and thank you for it! My question is about how much drive space the system dataset will need. I'm using an 8GB SATA DOM as my boot "drive". My storage pool is 14TB spread over 20 1TB drives (big SuperMicro 4U rig but was a free server, obviously not new). 2 hot spares. From what I've read, the system dataset is generally in the megabyte range, tiny. I'm a NAS noob so I don't even know how to check how much room is available on the DOM. I want to say TNC uses about 2GB but I'm pulling that from weak memory.
I do have a pair of 128GB SSD's that were used as cache drives when this machine was originally set up with some horrible NAS software called OpenE and a hardware RAID. They aren't being used at all right now, I could add a pool to them.
OK to use the DOM for the system dataset or set up a new pool on the SSD's? I'm using this behemoth at home so it sits many hours or even sometimes days without pulling anything from it.
Yes! Definitely, boot drive is always very tiny. The only advantage of having a large one is that it spreads the writes across the memory to decrease wear and increase lifespan. However, 8GB should last you many years no issues (With the caveat that this is only if the drive itself is well manufactured. Cheap and crappy will not last because it is cheap and crappy, not because of a size issue!).
@@BlackbeardSupport The DOM was made by Wintec. Came with the machine.
Thanks for responding!
How do you know what size you need for the system datapool set? I have 2 4TB nas drives in raid 1 and a 128 GB nvme drive as boot pool.. would that be enough?
What size does the info take?
It can vary depending on a few factors, but generally 32Gigs is plenty, and 64 is way more than enough. I like to put in a Min of 64, but that's just cause the memory is cheap.
@@BlackbeardSupport Thanks so I will move my info to the 128GB nvme bootpool drive
Yea, that will definitely work!
I was planning to make NAS home build, but i could not get to spin down disks (1 pool with 3 disks), in the end i got WD home cloud, and i dont like it cause it also doesn't spin down disks. And now i came across this video and the problem was system data set log on that data set! .. So i move the log to my usbnas key, and i notice that it was slow on configuring disks, saving etc.. and then it crashed NAS in to a boot loop. But it was just nas to play around, no data lost. i presume that its USB 2.0 was the problem. So i will try to add ssd cash device to NAS and use it as syslog device.
Hi Ivan, glad I could point you to the system dataset. It's true in the past that TrueNAS boot pool was recommended to use a USB drive, but in the last few years this is not the best option. A small SSD (Even 60Gig would be more than enough, maybe even smaller would work) would be a great option, and hopefully you will have more success. If you can buy two, it can be a great idea so that you can have a boot mirror, in case one drive fails. You can see how to do this on my video here: ruclips.net/video/xTUsJNvLHS8/видео.html.
Best of luck with your NAS.
The check command dose not work on a shelf
Sorry, not sure what you mean. Are you having trouble with it in the shell maybe?
@@BlackbeardSupport Sorry, to clarify im useing a DAS NetApp shelf for my drives. it seems to be doing something as my power consumptions dropped but it seems i cant use the command to get the disk info
Hmm, yes, I can see how that might be an issue. Unfortunately I don't have a DAS NetApp shelf so I can't test myself, but I wonder if the server doesn't support direct access to disk information.
Does the server return any disk information from any command at all, or is it JUST this command that fails?
@@BlackbeardSupport In the metrics area on truenas i can see alot of stats. what other commands could i try?
The fact that my DL380P Gen8 uses 65 watts idle without the spinning drives plugged in, and 100-120 Watts idle when they are in the server makes a huge difference in my electricity bill (0.6 Euro Cents / Kwh // Germany!) so i already made the changes but the server never stopps the drives and nothing is using them. I store ripped movies and series on them so dont need to be 24/7 online
By the way i use a virtualized TrueNAS Scale (has qemu-guest-agent) with a dedicated PCI HBA. Dataset set to boot pool, and no force spindown option anymore :/
Unfortunately I have not played with Scale enough to give you advise. :/ That's annoying.
@@BlackbeardSupport in some cases we need to disable smart monitoring on the drives, so they can spin down. Since acl permissions are a mess and cant connect via nfs to the freshly configured share (after two successful mounts on different things) i decided to wipe this out of my system for ever. And instead i install the Core version again and deal with the ssh and scripting things again...
Is there a qemu-guest-agent for the bsd version? Im not familiar with bsd at all.
There's a port available www.freshports.org/emulators/qemu-guest-agent. To be honest I have not tried it myself (I run on bare metal), but I've heard some positive comments from those who have. As with any port, your mileage may vary!
@@BlackbeardSupport I got scale wiped completely, useless garbage, went back to Core and yet, drives are spinning down after a while but the power consumption remains the same :D weird .
@@victorshane4134 Well that is unusual, though if you're using newer drives, they may already by very power efficient. A good ballpark for the consumption being used is about 10Watts. If you have a lot going on on the machine and only a few drives, it could be hard to notice the dip.
The other thing to check is that the drives STAY spun down, and don't keep spinning back up for things like daily scrubs. No easy way to check, unfortunately. Just have to occasionally run the command to see they are spun down.
Ratz, no camcontrol on my image: TrueNAS-SCALE-22.12-BETA.2, I guess, one should use: hdparm -C /dev/sdb