| 1 |
On 01/29/2017 02:09 AM, Peter Humphrey wrote: |
| 2 |
>> Do you have a SSD? That sounds like symptoms of a failing SSD to me |
| 3 |
>> (it's happened more than once to me :/) |
| 4 |
> |
| 5 |
> Yes, nothing but one 256 GB NVMe drive in this box. Is there something like |
| 6 |
> hdparm that will keep an eye on it for me? |
| 7 |
> |
| 8 |
> I hope it isn't that already: the whole system only dates from last May. |
| 9 |
> |
| 10 |
|
| 11 |
I used to run an SSD on my home server (running mythtv, among other |
| 12 |
things), until one day I got home and it was misbehaving. |
| 13 |
|
| 14 |
I couldn't log into the server using ssh, so I logged in on the console. |
| 15 |
Everything appeared to be running but nothing new would load. Tried |
| 16 |
sync'ing to see if I could remerge various packages but got a similar |
| 17 |
weird error and found out half my portage tree was *poof*. On further |
| 18 |
inspection, about 40% of the OS files were gone. No wonder I problems |
| 19 |
logging on. |
| 20 |
|
| 21 |
This happened twice to me, so I no longer use SSDs on machines I use as |
| 22 |
servers in my home, they're just too unreliable. At least with spinning |
| 23 |
rust you get some warning and can usually do some recovery. |
| 24 |
|
| 25 |
As far as monitoring, some vendors have tools for that I believe. They |
| 26 |
usually do run SMART. |
| 27 |
|
| 28 |
Are you running fstrim once in a while like it's recommended? Apparently |
| 29 |
using 'discard' as an option when mounting is no longer recommended. On |
| 30 |
my laptops I use a systemd timer to do this. Before that I used anacron |
| 31 |
(I think it was anacron) which would run missed cronjobs. |
| 32 |
|
| 33 |
Dan |
Archives