Old Hard Drives

[LunarMist]

I have a plethora of the old-school 8TB WD RED and the unofficial EMAZ versions with the same firmware. Many are from 2017 and a few from 2018, mostly with 300-1200 hours on the SMART clock and have been in room temperature storage since the viral disease. If I wipe data is there any reason not to reuse these drives simply due to age? The newer small drives like this are noisier, don't make use of the He, and/or suffer from the shingles. I hate to throw out 160TB if they are expiring.

 

[sedrosken]

Nothing immediately strikes me as wrong with that. I wouldn't be using them to put together something meant to be used for production -- not having a warranty anymore makes that a dicey proposition -- but for personal use they ought to be fine.

(If you're looking to offload a few, let me know -- I happen to have a need to upgrade my NAS, and I'm having trouble finding modern models that are still CMR.)

 

[Handruin]

I'm in a similar boat with 20 x 6TB drives I'm about to retire with newer drives. I don't have a specific need for them after retiring them but they all have 46000+ hours power on time and out of warranty.

for x in {b..u}; do sudo smartctl --all /dev/sd${x} | grep Power_On_Hours; done 9 Power_On_Hours 0x0012 094 094 000 Old_age Always - 46761 9 Power_On_Hours 0x0012 094 094 000 Old_age Always - 46883 9 Power_On_Hours 0x0012 092 092 000 Old_age Always - 57537 9 Power_On_Hours 0x0012 094 094 000 Old_age Always - 46882 9 Power_On_Hours 0x0012 094 094 000 Old_age Always - 46883 9 Power_On_Hours 0x0012 094 094 000 Old_age Always - 46738 9 Power_On_Hours 0x0012 094 094 000 Old_age Always - 46738 9 Power_On_Hours 0x0012 094 094 000 Old_age Always - 46761 9 Power_On_Hours 0x0012 094 094 000 Old_age Always - 46761 9 Power_On_Hours 0x0012 094 094 000 Old_age Always - 46761 9 Power_On_Hours 0x0012 094 094 000 Old_age Always - 46738 9 Power_On_Hours 0x0012 094 094 000 Old_age Always - 46748 9 Power_On_Hours 0x0012 094 094 000 Old_age Always - 46747 9 Power_On_Hours 0x0012 094 094 000 Old_age Always - 46748 9 Power_On_Hours 0x0012 094 094 000 Old_age Always - 46738 9 Power_On_Hours 0x0012 094 094 000 Old_age Always - 46748 9 Power_On_Hours 0x0012 094 094 000 Old_age Always - 46738 9 Power_On_Hours 0x0012 094 094 000 Old_age Always - 46761 9 Power_On_Hours 0x0012 094 094 000 Old_age Always - 46731 9 Power_On_Hours 0x0012 094 094 000 Old_age Always - 46731

 

[Mercutio]

The guys at my datacenter swap out and sell drives after three years of 24x7 operation. They've been my main source for drives larger than 8TB and to date I've had good luck with them, although I'm typically getting SAS drives with ~18 months of remaining warranty. I have a couple dozen drives in service that I've bought that way. I had one die and I have one I pulled for making different noises than the others, though it still passes testing. Losing one drive out of 24 in a ~30 month period is within a normal rate of loss. Many of the drives I'm talking about have exceeded 50k operating hours.

I don't think it's a great state of affairs, but to date I'm not excessively concerned. Drives with only 1200 power on hours, though? Those are basically new.

 

[LunarMist]

I'm in a similar boat with 20 x 6TB drives I'm about to retire with newer drives. I don't have a specific need for them after retiring them but they all have 46000+ hours power on time and out of warranty.

I have maybe ten of the 6TB drives with archives, but they are just hanging around until disposal. I never liked that generation (6x1TB platters) of hot drives.

The guys at my datacenter swap out and sell drives after three years of 24x7 operation. They've been my main source for drives larger than 8TB and to date I've had good luck with them, although I'm typically getting SAS drives with ~18 months of remaining warranty. I have a couple dozen drives in service that I've bought that way. I had one die and I have one I pulled for making different noises than the others, though it still passes testing. Losing one drive out of 24 in a ~30 month period is within a normal rate of loss. Many of the drives I'm talking about have exceeded 50k operating hours.
I don't think it's a great state of affairs, but to date I'm not excessively concerned. Drives with only 1200 power on hours, though? Those are basically new.

I was hoping that the He tech drives would be better over time since they run cooler and are sealed to exclude oxygen and water from the moving parts. Due to the virus ruining everything, the drives were out of service for years. Originally I was thinking about two sets of RAIDs stored at CRT in the mid-section of the US, but maybe one closer on the other side of a river. That might be impossible in the event of an emenrgecny.

 

[jtr1962]

Regarding the number of hours, even for Doug's drives that probably wouldn't concern me unless they were in a very hot environment. My only HDD current still in regular service is the 2TB Samsung I bought in 2010. Still seems healthy according to the SMART data:

 

[Handruin]

All the drives have been in my basement for years with mostly cool temps and my basement area is part of my HVAC so in the summer it gets cooled. They're all in a proper 24 bay supermicro chassis with proper fans running at high speed and have been behind a rackmount UPS.

None the less I am growing concerned at their age which is why I'm considering migrating to 18-20TB drives to run fewer of them. I'm currently doing zfs snapshot sends to my other NAS with 16TB drives but it's almost full.

sudo ./drive-temps.sh 05-28-2022-20:47:21 - Device: /dev/sdb | Temperature Celcius: 33 | Health: PASSED 05-28-2022-20:47:21 - Device: /dev/sdc | Temperature Celcius: 37 | Health: PASSED 05-28-2022-20:47:21 - Device: /dev/sdd | Temperature Celcius: 37 | Health: PASSED 05-28-2022-20:47:21 - Device: /dev/sde | Temperature Celcius: 36 | Health: PASSED 05-28-2022-20:47:21 - Device: /dev/sdf | Temperature Celcius: 36 | Health: PASSED 05-28-2022-20:47:21 - Device: /dev/sdg | Temperature Celcius: 35 | Health: PASSED 05-28-2022-20:47:21 - Device: /dev/sdh | Temperature Celcius: 42 | Health: PASSED 05-28-2022-20:47:21 - Device: /dev/sdi | Temperature Celcius: 41 | Health: PASSED 05-28-2022-20:47:21 - Device: /dev/sdj | Temperature Celcius: 40 | Health: PASSED 05-28-2022-20:47:21 - Device: /dev/sdk | Temperature Celcius: 39 | Health: PASSED 05-28-2022-20:47:21 - Device: /dev/sdl | Temperature Celcius: 38 | Health: PASSED 05-28-2022-20:47:21 - Device: /dev/sdm | Temperature Celcius: 34 | Health: PASSED 05-28-2022-20:47:21 - Device: /dev/sdn | Temperature Celcius: 41 | Health: PASSED 05-28-2022-20:47:21 - Device: /dev/sdo | Temperature Celcius: 41 | Health: PASSED 05-28-2022-20:47:21 - Device: /dev/sdp | Temperature Celcius: 40 | Health: PASSED 05-28-2022-20:47:21 - Device: /dev/sdq | Temperature Celcius: 39 | Health: PASSED 05-28-2022-20:47:21 - Device: /dev/sdr | Temperature Celcius: 37 | Health: PASSED 05-28-2022-20:47:21 - Device: /dev/sds | Temperature Celcius: 34 | Health: PASSED 05-28-2022-20:47:21 - Device: /dev/sdt | Temperature Celcius: 38 | Health: PASSED 05-28-2022-20:47:21 - Device: /dev/sdu | Temperature Celcius: 35 | Health: PASSED ========= #!/bin/bash DATE=$(date +"%m-%d-%Y-%H:%M:%S") for d in $(lsblk -S -o NAME -n); do HEALTH=$(sudo smartctl -H /dev/$d | grep "SMART overall-health" | awk '{ print $6 }') sudo smartctl -A /dev/$d | grep "Temperature_Celsius" | awk -v awk_drive="$d" -v awk_date="$DATE" -v awk_health=$HEALTH \ '{ print awk_date " - Device: /dev/" awk_drive " | Temperature Celcius: " $10 " | Health: " awk_health }' done

 

[jtr1962]

I guess with data which is important to you it's better to preemptively replace drives before they fail.

What do you guys store anyway? I've had PCs since the mid 1990s but I probably have under 500 GB of unique data, not including OS and programs.

 

[Handruin]

It's partially a preemptive change and also I would like to reduce energy consumption and heat by spinning fewer HDDs.

I mainly storage my media collection which takes up the bulk of the space. I rip my UHD movies to mkv containers as remux to watch natively. The other part is photography backups for the two of use here.

I originally planned for 100TB usable to last me for about 5-6 years (after parity) but have only managed to fill about 60% of that.

 

[jtr1962]

OK, movies. That makes sense. You can never have too much space with movies.

Yeah, fewer spindles is better. The heat is beneficial in winter but opposite in summers. It looks like it might be a hot June, too. I'll be dying until July 1 if it is. That's when I turn on my ACs. They get turned off on August 31.

 

[Mercutio]

I also have an enormous media collection, though my stretches to include substantial collections of a number of non-video formats as well.

 

[sedrosken]

What do you guys store anyway? I've had PCs since the mid 1990s but I probably have under 500 GB of unique data, not including OS and programs.

Personally, in addition to the media collection -- music in FLAC can take up quite a bit of space as my library continues to grow, and movies and such tend to be space hogs as well -- I keep a software archive for tools, games and drivers for everything I currently own. I fetch stuff as needed and just hold onto it indefinitely. Does that make me a digital hoarder? Maybe, but it's come in handy far too many times for me to stop. I already have a few things -- mostly old games -- that I archived shortly before they disappeared off the internet (save some far-flung reach of the Wayback machine, perhaps) forever. I also like to keep backup images of all of my machines so if I screw something up it's no big deal to restore the image so I hopefully don't have to set everything up from scratch again. The archive is roughly 600GB by itself so far, growing every day, and the backups are another roughly 3-400GB. My collection of movies and other assorted videos -- mostly youtube downloads hedging against them being removed -- comes out to 800GB, and my music currently tops out at 105GB but grows with each album I buy, physical or digital.

I find these numbers hilarious considering less than 10 years ago I was completely living out of a single 250GB drive and thought I'd never need more. Nevermind the fact that the collections of most others here dwarf mine by orders of magnitude.

 

[jtr1962]

Personally, in addition to the media collection -- music in FLAC can take up quite a bit of space as my library continues to grow, and movies and such tend to be space hogs as well -- I keep a software archive for tools, games and drivers for everything I currently own. I fetch stuff as needed and just hold onto it indefinitely. Does that make me a digital hoarder? Maybe, but it's come in handy far too many times for me to stop. I already have a few things -- mostly old games -- that I archived shortly before they disappeared off the internet (save some far-flung reach of the Wayback machine, perhaps) forever. I also like to keep backup images of all of my machines so if I screw something up it's no big deal to restore the image so I hopefully don't have to set everything up from scratch again. The archive is roughly 600GB by itself so far, growing every day, and the backups are another roughly 3-400GB. My collection of movies and other assorted videos -- mostly youtube downloads hedging against them being removed -- comes out to 800GB, and my music currently tops out at 105GB but grows with each album I buy, physical or digital.

I find these numbers hilarious considering less than 10 years ago I was completely living out of a single 250GB drive and thought I'd never need more. Nevermind the fact that the collections of most others here dwarf mine by orders of magnitude.

I tend to keep anything which interests me indefinitely for the same reason. Eventually it's going to disappear from the Internet. Like I said, I probably have under 500 GB of unique data at this point. I remember back when a few GB of storage space was considered huge. You had to be selective about what you keep. Now drives are so huge it's not even a concern for most people. You can get 1TB of SSD storage for under $100. For most people that'll last for years. I just got a 2TB Samsung T5 external SSD. I'm using it to make periodic images of my Windows 7 and XP installs on my primary PC. I can keep at least 2 images of each. That's fine. Generally if I have enough trouble with my PC to warrant wiping the drive and installing an image I only need the most recent one. No need to save months worth of images, although I'd probably do that just for heck of it if I had 8 or 12 or 14 TB for backup. In truth, since I hardly ever boot into my XP install, and nothing changes much, I could image it only whenever I boot into it (and keep just the last image before that). So that's maybe 1 TB for two images. The Windows 7 install is around 400 GB now. With an 8TB backup drive I could keep the last two XP images and over a dozen Windows 7 images. That's plenty.

If ever get into saving movies then I'll need massive amounts of storage. Still would prefer SSD for that, but the cost would have to drop by a factor of ten at least. If the day comes I can buy a 10TB SSD for about 100 bucks I'll be a happy camper.

 

[Mercutio]

I would also be very happy to see 10TB for $100, even if it's stupid SMR.

 

[Adcadet]

I am redoing my basement and server room (closet) and continue to use 4 TB drives from 2017-2018. SSDs are still too expensive for me.

 

[Mercutio]

If you have seldom-accessed data like media files, there's no reason to look at SSDs anyway. Most of your inexpensive high capacity drives are going to be crummy QLC NAND, and anything that isn't that as of this writing is probably an expensive Enterprise-class model. Intel D3s or something.

I have a pair of 4TB Hitachi UltraStar drives running in the storage server in my hosted rack that were made in 2011. They're not doing anything super-critical (they're scratch disks for local Crashplan backups) and I'm morbidly curious about how long they'll run. So far, they still pass every diagnostic I throw at them.

 

[Adcadet]

I loved Crashplan....didn't they pull out of the consumer market?

 

[Mercutio]

I loved Crashplan....didn't they pull out of the consumer market?

Yes, but you can just use Crashplan Small Business instead. That's appropriate for my needs anyway, but people who were on individual plans were migrated over and just asked to pay a little more for it if they were on a family plan.

 

[Adcadet]

IIRC, CrashPlan had a nice data deduplication feature...that seemed remarkably similar to the ZFS feature. Do you know if they actually run on ZFS?

 

[Handruin]

I recall needing to disable CrashPlan deduplication because it was such a RAM hogging resource via their Java client. I don't know for sure if they used ZFS with dedup on the back end but I suspect not. Rarely have I seen or heard of anyone using zfs with dedup because the tradeoff of RAM and speed isn't worth it given the cost of HDDs these days.

However it's worth running ZFS with lz4 compression because it'll barely add any overhead.

 

[LunarMist]

I have a pair of 4TB Hitachi UltraStar drives running in the storage server in my hosted rack that were made in 2011. They're not doing anything super-critical (they're scratch disks for local Crashplan backups) and I'm morbidly curious about how long they'll run. So far, they still pass every diagnostic I throw at them.

Those old Hibachis just kept on going.
I recently retired an olden WD Black 4TB, which was their fastest 3.5" HD at the time. It still worked perfectly, but is just too small. Modern drives are slower for whatever reason, but the He drives are just so much cooler.

 

[Handruin]

I've lost a bundle of my old 4TB Hibachis but at least 8 of the 12 still work. So far all of my 6TBs still work.

 

[LunarMist]

I was all on the Seagates by the time 6TB was maximum. I don't recall if they were first or just more available. They were quite warm also and needed some airflow. Those were the final non-He 3.5" drives I ever bought.

 

[Handruin]

None of mine are helium filled drives either and they do get a bit warm. They're all in a proper server chassis which has adequate fans and airflow in my cooler basement so they've been managed pretty well.