Ask HN: Do you monitor SSD disk writes regularly to prevent excessive SSD wear? | Hacker News Reader

Ask HN: Do you monitor SSD disk writes regularly to prevent excessive SSD wear?

7 pointsamichail10 comments

It only takes a single software bug to kill your SSD with excessive writes.

Life’s too short for that. Chances are you’ll get a new machine before your SSD wears out unless you’re doing some really disk intensive work.

I had a 11 year old SSD fail recently. The rest of my machine is still working fine at over 12 years old. Not sure it was write related. Could have been related to age of the components (caps in the controller?). The only reason I'll need to get rid of this rig (or turn it into a Linux box) is because of the TPM 2.0 requirement for Win11 and the upcoming EoL for older Win versions.

I check my SMART readout every few months for weirdness but besides that, no.

I check SMART regularly. SSDs aren't cheap!

When I'm writing code, disk wear is pretty much always on my mind. I do RasPi work, and I'm always aware that every write could cause a failure if a power glitch happens at the same time, even if the total isn't excessive.

I think OSes should be doing this by default though. Alarms for unusual conditions should be first class OS objects or built into some common standard out of the box tool.

There's about a dozen issues you'd never noticed unless you manually check or set up a monitor service yourself.

The only SSDs I have personally experienced failing due to wear are ones that I was using in an array for a database. These drives lasted several (5+) years.

While possible (and I am aware of an issue like this killing some SSDs on M1 Macs early on), it's not something I am overly worried about.

You should be backing up anyway.

I haven't been monitoring. I set up backups assuming there will be a failure. The few times I have seen symptoms of a failing drive, SMART didn't show any issues half the time.

Do you know that many SSD's go into read-only mode prior to failure ? If you have a spare on hand; copy && go! Like it or not, but storage has become a consumable.

I've seen many SSD failures in servers and all but one were just gone. No longer visible via the disk interface. The other one just had very slow response for quite some time (I think it was reallocating related to a large cluster of failing flash).

I have heard rumors that SSDs can go into read-only mode, but never experienced it. Our write rate was relatively low though, and we did not include it in our monitoring.

In such a scenario, iostat is your friend. https://manpages.debian.org/unstable/sysstat/iostat.1.en.htm...

Reply on news.ycombinator.com