Ssd data loss without power

[Pc6777]

I have some secondary PC's that might not be powered on for longer periods of time, do I need to power them on from time to time so the ssd don't loose data/os? And how often? And what's the deal with hard drives do I need to power the of on from time to time so a PC with os on hard drive dosnt loose data?

[Mopy]

Eh around every 6 months you should be fine but you could leave it up to a year without turning on however you should normaly be fine.

[George.]

as far as I know, regular turning on and off it will just keep the mechnical parts in good shape.

If fans don't spin for months (and you live somewhere where there are cold winters and hot summers) the greese that is applied becomes kinda stuffy and you may need to re-greese them. Probably goes like this for hdds as well since they also have spinning parts (may be fine though since they are in a closed vacuum.; if it hurts them tho that'd be a problem since opening and fixing a hdd is pretty much a lost cause if you don't know exactly what you're doing and have the proper equipment).

As far as ssds go I've never heard of persistant eletronic memory suffering from data loss in the event of being shut down for far too long.

 

P.S. I've only had problems with the greese of gpu fans and HDD weird behaviour after long periods of having my pc shut down completely.

[vanished]

21 minutes ago, George. said:

[...]

As far as ssds go I've never heard of persistant eletronic memory suffering from data loss in the event of being shut down for far too long.

[...]

I have heard this mentioned before but I do not have a lot of information on whether it's true or what the timescale might be.  I personally have a drive that I power on only occasionally every few weeks or even months and have not noticed any issues yet.

 

As for HDDs, I agree about not letting them sit for too long.  I have heard that it can be bad for them, and in fact this may be true of any mechanical device.  However, there's something to be said for not cycling them on and off too much either as this is also bad.  Although a lot of runtime would surely take a toll eventually, I think the ideal situation is to just turn them on and leave them on as much as possible.  Of course I know this isn't exactly feasible for every situation (a backup drive for example), but something to keep in mind.  As an anecdote, that's what I've done with the main drive in my system which is rather old now in terms of sheer hours but still works perfectly.

 

 

[porina]

The data integrity of flash does decay slowly over time. But how slow? I dug out this old article: https://www.anandtech.com/show/9248/the-truth-about-ssd-data-retention

Depends on operating and storage temperatures. At the end of the drives rated endurance, a year or more is no problem under typical conditions, but you might want to be careful if you have weird operating or storage conditions.

 

Quote

Remember that the figures presented here are for a drive that has already passed its endurance rating, so for new drives the data retention is considerably higher, typically over ten years for MLC NAND based SSDs. If you buy a drive today and stash it away, the drive itself will become totally obsolete quicker than it will lose its data.

 

Also I'm not sure what it would take to "refresh" the SSD when you do power it again. It is the data on the flash you have to worry about. Powering it for a short time is unlikely to change that data.

[Pc6777]

1 hour ago, George. said:

as far as I know, regular turning on and off it will just keep the mechnical parts in good shape.

If fans don't spin for months (and you live somewhere where there are cold winters and hot summers) the greese that is applied becomes kinda stuffy and you may need to re-greese them. Probably goes like this for hdds as well since they also have spinning parts (may be fine though since they are in a closed vacuum.; if it hurts them tho that'd be a problem since opening and fixing a hdd is pretty much a lost cause if you don't know exactly what you're doing and have the proper equipment).

As far as ssds go I've never heard of persistant eletronic memory suffering from data loss in the event of being shut down for far too long.

 

P.S. I've only had problems with the greese of gpu fans and HDD weird behaviour after long periods of having my pc shut down completely.

I have never heard of gpu fans needing greasing, if i dont touch a gpu for a long time and use it will the fans work still but not as well or have problems? and yea it makes sens if it sits too long a hard drive could loose the grease inside for the head which is much more catastrophic than a fan not having grease because hds are much more complex than a fan. and this stuff obviously dosnt effect flash storage, metal whispering does but i presume that takes a loong time. 

[Pc6777]

1 hour ago, porina said:

The data integrity of flash does decay slowly over time. But how slow? I dug out this old article: https://www.anandtech.com/show/9248/the-truth-about-ssd-data-retention

Depends on operating and storage temperatures. At the end of the drives rated endurance, a year or more is no problem under typical conditions, but you might want to be careful if you have weird operating or storage conditions.

 

 

Also I'm not sure what it would take to "refresh" the SSD when you do power it again. It is the data on the flash you have to worry about. Powering it for a short time is unlikely to change that data.

does data integrity have to do with usable shelf life or just data stored on it? and I thought ssds needed to be powered every so often because they require small amounts of electricity to store data and cant store that small amount of electricity forever so booting it every so often would "refill" its electricity so it could keep storing data. 

[Pc6777]

1 hour ago, Ryan_Vickers said:

I have heard this mentioned before but I do not have a lot of information on whether it's true or what the timescale might be.  I personally have a drive that I power on only occasionally every few weeks or even months and have not noticed any issues yet.

 

As for HDDs, I agree about not letting them sit for too long.  I have heard that it can be bad for them, and in fact this may be true of any mechanical device.  However, there's something to be said for not cycling them on and off too much either as this is also bad.  Although a lot of runtime would surely take a toll eventually, I think the ideal situation is to just turn them on and leave them on as much as possible.  Of course I know this isn't exactly feasible for every situation (a backup drive for example), but something to keep in mind.  As an anecdote, that's what I've done with the main drive in my system which is rather old now in terms of sheer hours but still works perfectly.

 

 

so if you run it 24/7 so it always being lightly used is better than just storing it so everything continues to spin properly and the lubricant dosnt stick/ heads get stuck?

[NewMaxx]

It's known as leakage or detrapping, and is generally tested for 10 years. Actual leakage depends on many factors such as active and stored temperature - typical active of 55C with storage at 25C is rated for 404 weeks (7.77 years). The age/wear of a drive may be a factor among other things. Modern drives will locate and refresh cells that are error-prone, so powering on the drive occasionally will improve flash read lifespan. I would advocate to do it once every 52 weeks (once a year) for client use based on the JEDEC models.

[NewMaxx]

12 hours ago, Pc6777 said:

does data integrity have to do with usable shelf life or just data stored on it? and I thought ssds needed to be powered every so often because they require small amounts of electricity to store data and cant store that small amount of electricity forever so booting it every so often would "refill" its electricity so it could keep storing data. 

Over time charge detrapping will change the voltage threshold distribution of the cell. When reading the data you're checking several reference voltages to determine the overall value so shifting and widening of of these distributions makes data harder to read. Modern controllers are able to check and rewrite blocks that are error-prone when powered on. One method used is to recharge the cells, however due to the nature of NAND you can only increase voltage. Since shifting and widening is not uniform, that is to say different for higher and lower voltage states, this is generally used instead to counter other effects (e.g. program disturb, which is nonetheless related to retention time since it's a factor both of active and power-off conditions).

 

Read here for more - specifically see Figure 18 on pg. 14.

[vanished]

While I appreciate the vote of support, I absolutely have not provided the "best answer" here so OP feel free to change that please