ssd endurance in unraid system

[caplam]

Hello to all,

one of my 2 ssd of the cache pool starts to have errors so i'd like to have advice on the situation.

Attached are the smart data of the ssd.

It's a western digital blue ssd 500Gb. I threw it, brand new, in the server one year ago when i installed unraid.

The ssd is sold for an endurance of 200TBW calculated according to jedec jesd219. But what does it mean ? 

Does it refer to nand writes (TLC or SLC) or host writes?

I find completly crazy that my ssd has 731Tb SLC writes and 610 Tb host writes in one year.

Edited May 28, 2020 by caplam

[JorgeB]

I would keep it for now, monitor SMART, if those reported uncorrectable errors continue to increase it's probably a good idea to replace it.

 

43 minutes ago, caplam said:

I find completly crazy that my ssd has 731Tb SLC writes and 610 Tb host writes in one year.

Likely related to this:

 

[PSYCHOPATHiO]

That's one of the things I look at when purchasing SSDs the TBWs. my NVME has about 1800 TBW endurance. I need that as I host most of my important VMs & all of my dockers on IT, caching is a secondary concern.

 

As a future purchase it's best to look at high endurance SSDs I guess.

Edited May 28, 2020 by PSYCHOPATHiO

[caplam]

my cache is not encrypted but i'll read the entire post.

Grafana dashboard tells me 37MB/S write average during the last 6 monts

[JorgeB]

11 minutes ago, caplam said:

my cache is not encrypted

Issue doesn't appear to be encryption related.

[caplam]

i absolutly need to find a workaround.

Since yesterday things are getting worse. 

Report uncorrected are growing on both ssds and i don't want to replace them and kill the new two.

 

Before installing unraid i had aprroximately the same load on a hp N54l and i3 4330 both installed with proxmox.  And i had no problem during 3 years.

The dockers were in a debian vm.