SSD failures

SDD or not,
My experience before loadshedding was that towns with poor AC Power distribution systems, poor joints, poor cabling, poor transformers, all sorts of things like earthing, were more prone to failures for the most part till that town was fixed which did happen. One could argue that those failures were worse than loadshedding just disconnecting the power, but electronics does not like fluctuations in power, thats why my pocket calculator is the most reliable item by far, and thats why I keep my battery up to date in modern car.
 
r00igev@@r said:
Does loadshedding contribute to a higher rate of SSD failures?
I've seen 4 SDDs fail after about 3 years use.
Click to expand...
Yes and no.

SSD's much like HDD's, while not having moving part has power loss protection, and only immediate data is lost. Power outage rarely kill hardware, power surge or brown out does. Also SSD's and computer hardware isn't directly connected to the wall socket. The Power supply is more likely to fail as a result of a power failure, and more likely the power supply it self will damage or kill hardware inside a computer, than a simple power failure. A shitty power supply, can also be a contributing factor.

It also depends on the size of the SSD, if you have a 128gig SSD no doubt you going to fill up and reuse cells in the SSD much more overall compared to a larger capacity SSD. I have a 128gig WD SSD that is now 4 years old + with 71% health that was pretty much used daily, along with photoshop I wasn't able to kill it. I now have a standard OS SDD and a dedicated scratch disk for photoshop and still only managed to drop it's health in 8 months by 2%.

It entirely depends how the SSD was used, if trim was active, did some thing stupid like trying to defrag an SSD, malware or viruses can significantly increase write cycles. But power outages killing an SSD, not likely. I think there is more to the story here. What SSD's did you buy, what size, brand and how was it used, ect.
 
wizardofid said:
Yes and no.

SSD's much like HDD's, while not having moving part has power loss protection, and only immediate data is lost. Power outage rarely kill hardware, power surge or brown out does. Also SSD's and computer hardware isn't directly connected to the wall socket. The Power supply is more likely to fail as a result of a power failure, and more likely the power supply it self will damage or kill hardware inside a computer, than a simple power failure. A shitty power supply, can also be a contributing factor.

It also depends on the size of the SSD, if you have a 128gig SSD no doubt you going to fill up and reuse cells in the SSD much more overall compared to a larger capacity SSD. I have a 128gig WD SSD that is now 4 years old + with 71% health that was pretty much used daily, along with photoshop I wasn't able to kill it. I now have a standard OS SDD and a dedicated scratch disk for photoshop and still only managed to drop it's health in 8 months by 2%.

It entirely depends how the SSD was used, if trim was active, did some thing stupid like trying to defrag an SSD, malware or viruses can significantly increase write cycles. But power outages killing an SSD, not likely. I think there is more to the story here. What SSD's did you buy, what size, brand and how was it used, ect.
Click to expand...
They are Innodisk SATA III SATADOM-SV 3ME2 32GB running Buster.
Nothing write intensive on them like a database app.
 
r00igev@@r said:
They are Innodisk SATA III SATADOM-SV 3ME2 32GB running Buster.
Nothing write intensive on them like a database app.
Click to expand...
a 32GB is tiny, probably killing it by overwritting cells
 
r00igev@@r said:
They are Innodisk SATA III SATADOM-SV 3ME2 32GB running Buster.
Nothing write intensive on them like a database app.
Click to expand...

Quick search, reveals, it has no Power loss protection, or dram, failures isn't a result of device failure but due to incorrect usage and such.

supersunbird said:
Who that? ;)
Click to expand...
A server based SSD that plugs directly into the Sata port.
 
So the units had 2GB of RAM and seem to trigger the use of swap space while the units with 4GB RAM never require swap space.

Does that mean the use of swap space is killing it?

Edit: Also moved to msata as well with upgrading to 4GB. Have these older units that don't work with a msata. Have ordered a few replacements, will bump up memory to 4GB and monitor.

This is now. I have no graphs of the times swap space was being used as I have deleted those graphs. My understanding was cached memory should have been dropped in preference to swap space being used but it doesn't seem to have been the case.

memory.png
 
Last edited:
r00igev@@r said:
So the units had 2GB of RAM and seem to trigger the use of swap space while the units with 4GB RAM never require swap space.

Does that mean the use of swap space is killing it?

Edit: Also moved to msata as well with upgrading to 4GB. Have these older units that don't work with a msata. Have ordered a few replacements, will bump up memory to 4GB and monitor.

This is now. I have no graphs of the times swap space was being used as I have deleted those graphs. My understanding was cached memory should have been dropped in preference to swap space being used but it doesn't seem to have been the case.

View attachment 1436525
Click to expand...

It depends. Linux likes to cache less used memory pages to the swap file, the longer the system uptime the larger the cache is likely to increase with memory pages cached to the swap and more active memory pages cached in the memory, just the way linux operates.

With linux you require, double the amount of swap space for 2gig of ram and equal over 2gig of ram, over 8gig half the ram, for it to function properly.
 
Shouldn't *theoretically* be as rough as having the moving parts crashing down when they get rugpulled
 
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X