Wanted to share a bit of excitement from this week. My Proxmox server with three mirrored SSD pairs started sending SMART errors. Lo and behold, two of the drives are showing errors and are part of the same mirror! D’oh. What are the chances/luck?
The RMA process is started and a spare is on the way (since it will be here faster than RMA and it’s good to have spares on hand… like I should have had already).
Drives purchased at the same time and from the same batch wearing out at the same time.
Problem with power/cabling that affected both drives.
Momentary power outage that results in drive errors if both were writing.
Have you been running regular (like monthly) scrubs?
Does the SMART data indicate a problem that usually results from wear, like excess reallocated sectors or is there some other malfunction?
Likely that first one Hank; drives purchased at the same time. It was still a bit of a surprise as they generally last much longer, and I had used hdparm to thin provision them as well. Admittedly, they’re consumer grade SSDs but I’ve used them successfully so far.
I have learned to be very wary of consumer-grade SSDs, in much the same way I’m extremely wary of “desktop” conventional hard drives. The failure rates are much, much higher in my experience–and to be clear, I’m the kinda guy who has to learn that the hard way. Learned it the hard way with Western Digital blue and green drives twenty years ago, learned it the hard way with various models of consumer SSD ten years ago.
I hear you but it’s really hard to make the leap for enterprise drives in my home-lab setting. The price difference is not negligible and these are the first two that I’ve had any issue with (the 14 or so others having eventually reached their wearout level). I have been keeping an eye on the Kingston DC series though, based on your recommendations.