Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Advertise on
Monitoring the health of SSDs
New on LowEndTalk? Please read our 'Community Rules' by clicking on it in the right menu!

Monitoring the health of SSDs


So far, I had servers with HDD, in a near future, I should be able to take a server with SSDs. So I was wondering which S.M.A.R.T. parameters to look careful at, to follow the evolution of the health of these drives?

Thank you,

Good Night and Good Luck :sleeping:


  • usually there is something like total blocks written. depending on the blocksize you can calculate the TB written. every model comes with a guarantued number (read the tech specs from vendor) so you can check, how far it is in its real lifespan.
    also SSDs usually have reserved blocks and s.m.a.r.t. should show if and how much are already in use.

    ofc those numbers are not set in stone, they just should help to get a better idea than just looking at power-on-hours ;-)

    Thanked by 1SteveMC KVM in US/UK/NL/DE: 15% off first 6 month | Netcup VPS/rootDS - 5€ off: 36nc15279180197 (ref)

  • Thank you very much @Falzo !

    Good Night and Good Luck :sleeping:

  • Wear_Leveling_Count

    Thanked by 1SteveMC


  • Thank you @eol .

    Good Night and Good Luck :sleeping:

  • datanoisedatanoise Member
    edited November 2018

    You often have two values regarding the data written on the disk: data sent to the drive and data written to the NAND. The one to take into account is the second one. Some controllers are very good at compressing stuff (like the SandForce SF-2281).

    Available_Reservd_Space will tell you how much of the extra NAND is still available. 100 means nothing has been used yet.

    Don't forget to use trim!

    Thanked by 1SteveMC

    don't buy what you don't need: you'll save money and will end up able to grab a quality VPS when really needed.

  • Don't bother. It dies when it decides to die. No amount of of monitoring can help you escape this.

    Our department recently had an incident where the entire storage node became unresponsive. It was caused by a failed hard drive (SSD is applicable here) where the kernel decided to lock up. A coworker has to wake up in the middle of the night to investigate in the datacenter.

    Although monitoring will mitigate other risks, ultimately you can't monitor your way to freedom.

    Thanked by 1SteveMC

    Now works for a public institution. Opinions expressed here are mine and mine alone.

  • Thank you @datanoise and @zllovesuki for your additional valuable information.

    Good Night and Good Luck :sleeping:

  • The Golden Rule: Keep backups. The more the better :smiley:

    Thanked by 3eol SteveMC datanoise
  • @datanoise said:
    Don't forget to use trim!

    Or just use a proper disk with a decent amount of over-provisioning.. Aka most enterprise grade drives will do :)

    Thanked by 2SteveMC vimalware
Sign In or Register to comment.