What to do with my old disks?

ethereal

Guru
Joined
Sep 10, 2012
Messages
762
one of my pools has 6 hard drives it is a Raidz2 - 5 are 3tb AVGP with 3 year warranty and a 3tb red.
all my avgp are between 6 and 7 years and my red is 4 years.

what would people recommend when the disks start dying?

i have been getting my disks from WD MyBooks. on my other pool i have 6 x 8tb.
i don't want smr but cannot afford red plus.
does anyone have any idea which mybook sizes are save that have normal drives.

thank you
 
Last edited:

Constantin

Vampire Pig
Joined
May 19, 2017
Messages
1,828
If you don’t need to expand the pool, then keep replacing like for like. Use companies like goharddrive.com to buy known-good SKUs with no-nonsense warranties even for used stuff.
 

ethereal

Guru
Joined
Sep 10, 2012
Messages
762
If you don’t need to expand the pool, then keep replacing like for like. Use companies like goharddrive.com to buy known-good SKUs with no-nonsense warranties even for used stuff.
thanks for your reply but i am looking to get mybook cmr drives
 
Last edited:

Constantin

Vampire Pig
Joined
May 19, 2017
Messages
1,828
I’m confused. You wanted to buy CMR-based drives, right? Godharddrive sells them in both 2.5” as well as 3.5” formats. They also list them by their exact SKUs, which allows you to avoid SMR drives.

Where they come from or used /new status is immaterial, since the GHHD warranty in my experience is even more painless than those of OEMs.
 

ethereal

Guru
Joined
Sep 10, 2012
Messages
762
the my books have safely stored my data so i don't want to change my methodology.
 

Constantin

Vampire Pig
Joined
May 19, 2017
Messages
1,828
Ok so then buy a bunch of them, qualify them, and then set aside as cold spares.
 

ethereal

Guru
Joined
Sep 10, 2012
Messages
762
back to my op

does anyone have any idea which mybook sizes are save that have normal drives.
 

Constantin

Vampire Pig
Joined
May 19, 2017
Messages
1,828
OEMs usually make no representations about what is inside a given external enclosure. If you recall, there was a run on various MyBook and like external enclosures with the hope to find a RED Helium CMR drive inside, with some folk getting lucky for a while and then more and more folk getting White Helium CMRs instead. Even if the external SKU was the same.

You might be able to get better info over at the servethehome forums and the /datahoarder reddit since they shuck a lot of drives. Good luck.
 

kherr

Explorer
Joined
May 19, 2020
Messages
67
For my $$ ... I won't even consider buying WD drives anymore, which I used to buy exclusively. When buying them straight up ... they won't even give you what RPM they are ... it's listed as "ABOUT" or "Class". Couple that with their SMR/CMR slight of hand, makes you wonder what other fraud they're pulling off ....

At least Seagate never listed an SMR drive as "NAS compatible" ...

As far as old disks ..... if they're "Over the Hill" time wise ...... backups or I cut them in half with a chop saw and pitch them. A sledge also works too.
 

Constantin

Vampire Pig
Joined
May 19, 2017
Messages
1,828
Consider donating them also after a 0000 wipe. A local hospital was only too happy to take my used 2+ TB drives rather than having to buy more. In my case, between the previously-resident data having been encrypted AND the drives being subsequently zeroed, I doubt they will be useful to anyone nefarious.
 

ethereal

Guru
Joined
Sep 10, 2012
Messages
762
For my $$ ... I won't even consider buying WD drives anymore, which I used to buy exclusively. When buying them straight up ... they won't even give you what RPM they are ... it's listed as "ABOUT" or "Class". Couple that with their SMR/CMR slight of hand, makes you wonder what other fraud they're pulling off ....

At least Seagate never listed an SMR drive as "NAS compatible" ...

As far as old disks ..... if they're "Over the Hill" time wise ...... backups or I cut them in half with a chop saw and pitch them. A sledge also works too.
i was very disappointed in wd for the reasons you listed. in my small sample the wd hdds have lasted over 6 years. with my books i get a cheap WDC HGST Ultrastar drive with helium and a 3 year warranty
 
Last edited:

joeschmuck

Old Man
Moderator
Joined
May 28, 2011
Messages
10,970
The title of this thread caught my eye. My answer to your first question is to take the drives apart and salvage the magnets. They are the best refrigerator magnets on the planet! :cool: Then break the platters with a hammer, one well placed hit will take care of your data forever.

As for shucked drives, that is a gamble as to what drive you actually get. I know you can save a lot of money going this route if you get what you expect, but if you end up with SMR drives then all you can do is put the drive back into the enclosure and try to return it. It's a risk of course and many people have made it and it has paid off for them, some folks were not so lucky. I don't shuck drives anymore, the risk is not worth the effort for me. I'd rather just buy the drives I desire and be good with it.

Just a little extra to explain why I don't shuck anymore... I have reduced the amount of data I desire to store, I only store about 100 movie titles (video content) and that is being reduced over time. I rarely use Plex anymore in favor of streaming services where it makes sense for my family. I'm running a smaller capacity server these days and am very happy with it. I'm not trying to encourage you to do the same, just letting you know that if you don't need the huge storage capacity then you could reduce the capacity of your NAS and save some money purchasing drives.
 

Constantin

Vampire Pig
Joined
May 19, 2017
Messages
1,828
You might be able to reveal a SMR drive by running a sustained write to it and seeing if there is a repetitive square wave pattern in the throughput. That is a easy tell for SMR drives since they rely on dumping the contents of the CMR cache to the SMR regions of the drive once the CMR cache is full. The drive will become 100% non-responsive during these cache transfers, resulting in zero throughput.
 

ethereal

Guru
Joined
Sep 10, 2012
Messages
762
The title of this thread caught my eye. My answer to your first question is to take the drives apart and salvage the magnets. They are the best refrigerator magnets on the planet! :cool: Then break the platters with a hammer, one well placed hit will take care of your data forever.

As for shucked drives, that is a gamble as to what drive you actually get. I know you can save a lot of money going this route if you get what you expect, but if you end up with SMR drives then all you can do is put the drive back into the enclosure and try to return it. It's a risk of course and many people have made it and it has paid off for them, some folks were not so lucky. I don't shuck drives anymore, the risk is not worth the effort for me. I'd rather just buy the drives I desire and be good with it.

Just a little extra to explain why I don't shuck anymore... I have reduced the amount of data I desire to store, I only store about 100 movie titles (video content) and that is being reduced over time. I rarely use Plex anymore in favor of streaming services where it makes sense for my family. I'm running a smaller capacity server these days and am very happy with it. I'm not trying to encourage you to do the same, just letting you know that if you don't need the huge storage capacity then you could reduce the capacity of your NAS and save some money purchasing drives.
the title may be misleading what i meant was because all the drives are older than 6 years. does anyone have any advice for replacing the drives as they fail. or should i do it before they die because of their age there could be multiple drives die at the same time.
 

Constantin

Vampire Pig
Joined
May 19, 2017
Messages
1,828
Some suggestions:
  1. Multiple backups. Preferably at least one off-site. The more important the data, the more backup copies.
  2. One or more qualified replacements - i.e. you stress-tested these drives for 24hrs with badblocks or whatever, then set them aside as cold, qualified spares. I standardized around 10TB drives for both my backup and the primary server infrastructure.
  3. Buy drives by exact SKU to know what you're getting, ideally from OEMs or resellers that honor warranties. Shucking drives has an appeal from a initial cost POV, but you have to be good at shucking drives w/o causing damage (to be able to re-assemble them if warranty service is required) and you likely have to store the specific external enclosure for each drive until the warranty has expired. I have a big bin here with plastic parts and a label on the outside re: when I get to chuck all that stuff.
 

rogerh

Guru
Joined
Apr 18, 2014
Messages
1,111
No-one seems to have attempted to answer the original question. Which seems to have been to ask whether it is wise to discard disks which are all a certain number of years/hours use old to prevent them expiring all at about the same time? My view, based on zero actual experience, would be that they will not be likely to expire at the same time (unlike SSDs which might reach a finite total write limit close together) but are likely to die with a wide dispersion of lifetimes because of the random processes involved. I would think it reasonable to wait till a disk actually fails, without too much extra risk of the other disks failing due to age while that one is being replaced. But, as I say, I have no practical experience of using large numbers of old disks, so a comment from those that have would be interesting.
 

jgreco

Resident Grinch
Joined
May 29, 2011
Messages
18,681
Neither hard drives nor SSD's are likely to die at the same time, with some exceptions:

If you push an SSD beyond its designed TBW, you do end up in a place where failures are more LIKELY. However, even there, it is highly unlikely that two SSD's will fail nearly simultaneously.

If you run an HDD beyond five years, you also end up in a place where failures are more LIKELY. However, again, it is highly unlikely that two HDD's will fail nearly simultaneously.

If you have either kind of device with a manufacturing or design defect, all bets are off. Famous examples include the HP SSD's that stop operating at exactly 32768 hours, the Hitachi Deathstar fiasco, Seagate's early 2010's issue with 1.5-3TB HDD's, WD's drives where someone forgets to turn off power management and the LCC skyrockets, etc.

Generally, I find that 3.5 bays are valuable enough that I try to keep larger hard drives in them, as our storage requirements here tend to grow with time, so important pools tend to have fresher drives, and old drives find their way to less-important uses.
 

Constantin

Vampire Pig
Joined
May 19, 2017
Messages
1,828
A common stressor event may also cause wider failure - a friend of mine worked at the Internet Archive and when the cooling system in one of their data centers failed, they lost a lot of disks. I still have to institute a script here that will shut down the server automatically if a drive fails, a simple functionality that consumer grade NAS' typically feature. A common stressor (such as a failing AC system) will hopefully hence only lead to one drive failure.

Inadequate cooling over the expected lifetime of the disk may also lead to more drives failing during the additional stress of a resilver (which generates heat, the PSU may have issues also, etc.). Hence, having a TrueNAS / FreeNAS with Zx data pools is not a substitute for backups, of which at least one should be off-site. Qualified spares are for the statistically-normal failure events.
 

ethereal

Guru
Joined
Sep 10, 2012
Messages
762
i will wait until the drive fails and then replace the 3tb with 6tb My Books

thank you for your replies
 
Top