At what point do you retire an SSD?

eeeeesh

BIT Beta Team
Joined
Jan 5, 2017
Messages
403
Reaction score
677
So I have this 1TB SSD in my Blue Iris machine that has a little over 45,000 'power on hours' which is about 5 years. At one point it was the main SSD of my ESXI Home server before it got retired. My current 'home server' has a 500gb nvme ssd for the OS and some storage. Crystal Disk indicates that SSD is still 'good' with a health status of 85%, but this old 1TB is still listed as 'good' but the a health status has dropped down to 56% which has me a little concerned.

I have Blue Iris set up to write to this drive first (new, alerts, etc) and hen it moves them to a 1 TB WD 2.5" hard drive still in the SFF tower and then from their to a 12 gb drive in my Synology NAS.

I am not really noticing any problems but was wondering - just let it run until it starts showing errors or what?

Crystal Disk 1TB.jpg
 

ThomasCamFan

Pulling my weight
Joined
Dec 14, 2020
Messages
134
Reaction score
204
Location
USA
I can't offer any practical advice on when to retire an aging SSD that still has good health status. If your version/release of the 850 Pro has a known good track record then I would continue to use it (classic watch and wait).

Have you enabled Over Provisioning? Doing that usually extends the life of a SSD, sometimes a lot.


- Thomas
 

Perimeter

Getting comfortable
Joined
Feb 18, 2023
Messages
557
Reaction score
581
Location
Europe
The problem with SSD retirement is, that they don't work like HDDs. When my HDDs have reached sufficient age, I will simply relegate them to multiple backup duty on the shelf (powerless). But I would not advise you to do this with SSDs because afaik, they will lose their memory eventually. So you use them or you toss them.
I am considering raid 1 for these, if it makes sense for the application.
 

CCTVCam

Known around here
Joined
Sep 25, 2017
Messages
2,660
Reaction score
3,480
Not sure I'd retire it.

That said, as I understand it these days if they "fail" they simply turn to read only to protect your data so you can still retrieve it.

I'm confused though about why you are moving your data around so much. 3 separate copy functions must be hammering the cpu and drives and for what?

I like many use 2 drives - an SSD for Apps and a hard drive for storage. Everything goes to "New" and the other folders "stored" etc are turned off. Alerts sits on the ssd. If I felt the need to protect the App SSD, I'd get another SSD just for alerts and write them there.
 

SpacemanSpiff

Known around here
Joined
Apr 15, 2021
Messages
1,456
Reaction score
2,431
Location
USA
Looks like "good" is the highest word rating CrystalDiskInfo (CDI) can give. It will not change to "caution" until Remaining Life <= 10%. According to their site, the health status references the S.M.A.R.T. attribute: SAMSUNG : 0xB4, Unused Reserved Block Count (Total) / Used Reserved Block Count (Total). However, there does not seem to be any info regarding whether they are interpreting the data from 0xB4 with other values, or merely displaying the raw data.

How long did it take to drop to 56%? If the 'caution' rating is not until 10%, I would monitor the CDI health status & the time it takes to approach the threshold and decide what risk I want to take with my data
 

duplo

Getting comfortable
Joined
May 26, 2022
Messages
293
Reaction score
364
Location
Berlin, Deutschland
i had intel ssds fail at 95-99% left.

problem of any ssd..
they fail from one moment to another. to my knowledge it is impossible to recover ssds. mostly the controller chip die and there is some crypto key.
 

Perimeter

Getting comfortable
Joined
Feb 18, 2023
Messages
557
Reaction score
581
Location
Europe
i had intel ssds fail at 95-99% left.

problem of any ssd..
they fail from one moment to another. to my knowledge it is impossible to recover ssds. mostly the controller chip die and there is some crypto key.
Never lost just write ability. They were gone 100% from one moment to next.
 

eeeeesh

BIT Beta Team
Joined
Jan 5, 2017
Messages
403
Reaction score
677
Thanks for the reply's. I this point, I think I will just watch and see...

And yes, I already have about 10 different 1 and 2 TB hard drives that have been retired and are now used for cold storage backups
 

bp2008

Staff member
Joined
Mar 10, 2014
Messages
12,666
Reaction score
14,005
Location
USA
Most SSD failures seem to just happen at random and cause the drive to disappear from the system, and those failures are usually very premature, not obviously caused by being written too much. I wouldn't worry about replacing it until it reports being at < 10% of life remaining.
 
Top