5 ms·

I’ve heard the exact opposite advice (keep the drives running to reduce wear from power cycling). Not sure what to believe, but I like having my ZFS NAS runnin

I’ve heard the exact opposite advice (keep the drives running to reduce wear from power cycling). Not sure what to believe, but I like having my ZFS NAS running so it can regularly run scrubs and check the data. FWIW, I’ve run my 4 drive system for 10 years with 2 drive failures in that time, but they were not enterprise grade drives (WD Green).

louwrentius 2y ago
Hard drives are often configured to spin down when idle for a certain time. This can cause many spinups and spindowns per day. So I don't buy this at all. But I don't have supporting evidence that back up this notion.
- Wowfunhappy 2y ago
  > Hard drives are often configured to spin down when idle for a certain time. This can cause many spinups and spindowns per day. I was under the impression that this was, in fact, known to reduce drive longevity! It is done anyway in order to save power, but the reliability tradeoff is known. No idea where I read that though, I thought it was "common knowledge" so maybe I'm wrong.
  - tekknik 2y ago
    It does, and primarily increases the load / unload cycle count. some WD drives are only rated at a few hundred thousand load / unload cycles. it’s best to buy drives that can handle being on always and leave them on.
- turnsout 2y ago
  I think NAS systems in particular are often configured to prevent drives from spinning down
- max-ibel 2y ago
  There seems to be a huge difference between spin-down while NAS is up vs shutting the whole NAS down and restart. When I start my NAS, it takes a bunch of time to be back up: it seems to do a lot of checking on/ syncing off the drives and puts a fair amount of load on them (same is true for CPU as well, just look at your CPU load right after startup). OTOH, when the NAS spins up a single disk again, I haven't notice any additional load. Presumably, the read operation just waits until the disk is ready.
Dalewyn 2y ago
>Not sure what to believe Keep them running. Why?: * The read/write heads experience literally next to no wear while they are floating above the platters. They physically land onto shelves or onto landing zones on the platters themselves when turned off; landing and takeoff are by far the most wear the heads will suffer. * Following on the above, in the worst case the read/write heads might be torn off during takeoff due to stiction. * Bearings will last longer; they might also seize up if left stationary for too long. Likewise the drive motor. * The rush of current when turning on is an electrical stressor, no matter how minimal. The only reasons to turn your hard drives off are to save power, reduce noise, or transport them.
- tomxor 2y ago
  > Keep them running [...] Bearings will last longer; they might also seize up if left stationary for too long. Likewise the drive motor. All HDD failures I've ever seen in person (5 across 3 decades), were bearing failures, in machine that were almost always on with drives spun up. It's difficult to know for sure without proper A-B comparisons, but I've never seen a bearing failure in a machine where drives were spun down automatically. It also seems intuitive that for mechanical bearings the longer they are spun up the greater the wear and the greater the chance of failure.
  - manwe150 2y ago
    I think I have lost half a dozen hard drives (and a couple DVD-RW drives) over the decades because they sat in a box for a couple years on a shelf (I recall that one recovered working with a higher amperage 12V supply, but only long enough to copy off most of the data)
    - mulmen 2y ago
      My experience with optical drives is similar to ink jet printers. They work once when new then never again.
- telgareith 2y ago
  Citations needed. Counterpoints for each: Heads don't suffer wear when parking. The armature does. If the platters are not spinning fast enough, or the air density is too low: the heads will crash into the sides of the platters. The main wear on platter bearings is vibration, it takes an extremely long time for the lube to "gum up." If its still a thing at all. I suspect it used to happen because they were petroleum distilate lubes. So, shorter chains would evaporate/sublimate leaving longer more viscous chains. Or straight polymerize. With fully synthetic PAO oils, and other options they won't do that anymore. What inrush? They're polyphase steppers. The only reason for inrush is that the engineers didn't think it'd affect lifetime. Counter: turn your drives off, thebsaved power of 8 drives being off half the day easily totals $80 a year- enough to replace all but the highest capacities.
CTDOCodebases 2y ago
I think a lot of the advice around keeping the drives running is about avoiding wear caused by spin downs and startups i.e. keeping the "Start Stop Cycles" low. Theres a difference between spinning a drive up/down once or twice a day and spinning it down every 15 minutes or less. Also WD Green drives are not recommended for NAS usage. I know in the past they used to park the read/write head every few seconds or so which is fine if data is being accessed infrequently but continuously however a server this can result in continuous wear which leads to premature failure.
- larusso 2y ago
  There used to be some tutorials going around to flash the firmware to turn greens into reds I believe. Which simply disables the head parking.
  - tekknik 2y ago
    modern 6tb red drives have 256MB of cache while a green only has 64MB. you cant flash in more drive cache!
  - CTDOCodebases 2y ago
    I think I know what you are talking about. I don’t know if it made the green drive identical to a red but it turned on TLER so the green drives wouldn’t co tangly drop out of the raid array. I’m running a 3 disk ZFS mirror of 8-10yr old greens and are yet to have an issues.
- cm2187 2y ago
  Agree. I do weekly backups, the backup NAS is only switched on and off 52 times a year. After 5-6 years the disks are probably close to new in term of usage vs disks that have been running continuously over that same period. Which leads to another strategy which is to swap the primary and the backup after 5 years to get a good 10y out of the two NAS.
foobarian 2y ago
> regularly run scrubs and check the data Does this include some kind of built-in hash/checksum system to record e.g. md5 sums of each file and periodically test them? I have a couple of big drives for family media I'd love to protect with a bit more assurance than "the drive did not fail".
- Filligree 2y ago
  It's ZFS, so that's built-in. A scrub does precisely that.
- adastra22 2y ago
  Yes, zfs includes file-level checksums.
  - giantrobot 2y ago
    Block level checksums.
  - jclulow 2y ago
    This is not strictly accurate. ZFS records checksums of the records of data that make up the file storage. If you want an end to end file-level checksum (like a SHA-256 digest of the contents of the file) you still need to layer that on top. Which is not to say it's bad, and it's certainly something I rely on a lot, but it's not quite the same!
    - faggotbreath 2y ago
      [flagged]
- ghostly_s 2y ago
  https://en.wikipedia.org/wiki/ZFS?wprov=sfti1#Resilvering_and_scrub_(array_syncing_and_integrity_checking) https://en.wikipedia.org/wiki/ZFS?wprov=sfti1#Resilvering_an...
- Cyph0n 2y ago
  Yep, ZFS reads everything in the array and validates checksums. ZFS (at least on Linux) ships with scrub systemd timers: https://openzfs.github.io/openzfs-docs/man/master/8/zpool-scrub.8.html#PERIODIC_SCRUB https://openzfs.github.io/openzfs-docs/man/master/8/zpool-sc...
bongodongobob 2y ago
This is completely dependant on access frequency. Do you have a bunch of different people accessing many files frequently? Are you doing frequent backups? If so then yes, keeping them spinning may help improve lifespan by reducing frequent disk jerk. This is really only applicable when you're at a pretty consistent high load and you're trying to prevent your disks from spinning up and down every few minutes or something. For a homelab, you're probably wasting way more money in electricity than you are saving in disk maintenance by leaving your disks spin.

louwrentius 2y ago

Hard drives are often configured to spin down when idle for a certain time. This can cause many spinups and spindowns per day. So I don't buy this at all. But I don't have supporting evidence that back up this notion.

Wowfunhappy 2y ago
> Hard drives are often configured to spin down when idle for a certain time. This can cause many spinups and spindowns per day. I was under the impression that this was, in fact, known to reduce drive longevity! It is done anyway in order to save power, but the reliability tradeoff is known. No idea where I read that though, I thought it was "common knowledge" so maybe I'm wrong.
- tekknik 2y ago
  It does, and primarily increases the load / unload cycle count. some WD drives are only rated at a few hundred thousand load / unload cycles. it’s best to buy drives that can handle being on always and leave them on.
turnsout 2y ago
I think NAS systems in particular are often configured to prevent drives from spinning down
max-ibel 2y ago
There seems to be a huge difference between spin-down while NAS is up vs shutting the whole NAS down and restart. When I start my NAS, it takes a bunch of time to be back up: it seems to do a lot of checking on/ syncing off the drives and puts a fair amount of load on them (same is true for CPU as well, just look at your CPU load right after startup). OTOH, when the NAS spins up a single disk again, I haven't notice any additional load. Presumably, the read operation just waits until the disk is ready.

Dalewyn 2y ago

>Not sure what to believe Keep them running. Why?: * The read/write heads experience literally next to no wear while they are floating above the platters. They physically land onto shelves or onto landing zones on the platters themselves when turned off; landing and takeoff are by far the most wear the heads will suffer. * Following on the above, in the worst case the read/write heads might be torn off during takeoff due to stiction. * Bearings will last longer; they might also seize up if left stationary for too long. Likewise the drive motor. * The rush of current when turning on is an electrical stressor, no matter how minimal. The only reasons to turn your hard drives off are to save power, reduce noise, or transport them.

tomxor 2y ago
> Keep them running [...] Bearings will last longer; they might also seize up if left stationary for too long. Likewise the drive motor. All HDD failures I've ever seen in person (5 across 3 decades), were bearing failures, in machine that were almost always on with drives spun up. It's difficult to know for sure without proper A-B comparisons, but I've never seen a bearing failure in a machine where drives were spun down automatically. It also seems intuitive that for mechanical bearings the longer they are spun up the greater the wear and the greater the chance of failure.
- manwe150 2y ago
  I think I have lost half a dozen hard drives (and a couple DVD-RW drives) over the decades because they sat in a box for a couple years on a shelf (I recall that one recovered working with a higher amperage 12V supply, but only long enough to copy off most of the data)
  - mulmen 2y ago
    My experience with optical drives is similar to ink jet printers. They work once when new then never again.
telgareith 2y ago
Citations needed. Counterpoints for each: Heads don't suffer wear when parking. The armature does. If the platters are not spinning fast enough, or the air density is too low: the heads will crash into the sides of the platters. The main wear on platter bearings is vibration, it takes an extremely long time for the lube to "gum up." If its still a thing at all. I suspect it used to happen because they were petroleum distilate lubes. So, shorter chains would evaporate/sublimate leaving longer more viscous chains. Or straight polymerize. With fully synthetic PAO oils, and other options they won't do that anymore. What inrush? They're polyphase steppers. The only reason for inrush is that the engineers didn't think it'd affect lifetime. Counter: turn your drives off, thebsaved power of 8 drives being off half the day easily totals $80 a year- enough to replace all but the highest capacities.

CTDOCodebases 2y ago

I think a lot of the advice around keeping the drives running is about avoiding wear caused by spin downs and startups i.e. keeping the "Start Stop Cycles" low. Theres a difference between spinning a drive up/down once or twice a day and spinning it down every 15 minutes or less. Also WD Green drives are not recommended for NAS usage. I know in the past they used to park the read/write head every few seconds or so which is fine if data is being accessed infrequently but continuously however a server this can result in continuous wear which leads to premature failure.

larusso 2y ago
There used to be some tutorials going around to flash the firmware to turn greens into reds I believe. Which simply disables the head parking.
- tekknik 2y ago
  modern 6tb red drives have 256MB of cache while a green only has 64MB. you cant flash in more drive cache!
- CTDOCodebases 2y ago
  I think I know what you are talking about. I don’t know if it made the green drive identical to a red but it turned on TLER so the green drives wouldn’t co tangly drop out of the raid array. I’m running a 3 disk ZFS mirror of 8-10yr old greens and are yet to have an issues.
cm2187 2y ago
Agree. I do weekly backups, the backup NAS is only switched on and off 52 times a year. After 5-6 years the disks are probably close to new in term of usage vs disks that have been running continuously over that same period. Which leads to another strategy which is to swap the primary and the backup after 5 years to get a good 10y out of the two NAS.

foobarian 2y ago

> regularly run scrubs and check the data Does this include some kind of built-in hash/checksum system to record e.g. md5 sums of each file and periodically test them? I have a couple of big drives for family media I'd love to protect with a bit more assurance than "the drive did not fail".

Filligree 2y ago
It's ZFS, so that's built-in. A scrub does precisely that.
adastra22 2y ago
Yes, zfs includes file-level checksums.
- giantrobot 2y ago
  Block level checksums.
- jclulow 2y ago
  This is not strictly accurate. ZFS records checksums of the records of data that make up the file storage. If you want an end to end file-level checksum (like a SHA-256 digest of the contents of the file) you still need to layer that on top. Which is not to say it's bad, and it's certainly something I rely on a lot, but it's not quite the same!
  - faggotbreath 2y ago
    [flagged]
ghostly_s 2y ago
https://en.wikipedia.org/wiki/ZFS?wprov=sfti1#Resilvering_and_scrub_(array_syncing_and_integrity_checking) https://en.wikipedia.org/wiki/ZFS?wprov=sfti1#Resilvering_an...
Cyph0n 2y ago
Yep, ZFS reads everything in the array and validates checksums. ZFS (at least on Linux) ships with scrub systemd timers: https://openzfs.github.io/openzfs-docs/man/master/8/zpool-scrub.8.html#PERIODIC_SCRUB https://openzfs.github.io/openzfs-docs/man/master/8/zpool-sc...

bongodongobob 2y ago

This is completely dependant on access frequency. Do you have a bunch of different people accessing many files frequently? Are you doing frequent backups? If so then yes, keeping them spinning may help improve lifespan by reducing frequent disk jerk. This is really only applicable when you're at a pretty consistent high load and you're trying to prevent your disks from spinning up and down every few minutes or something. For a homelab, you're probably wasting way more money in electricity than you are saving in disk maintenance by leaving your disks spin.