Slowdown over time of Crucial P3 SSD (vs Intel P4610)

[hmw]

The P3 is definitely weird. I've seen this drive slow down to a crawl when filled to 1/2 capacity. I tried to overprovision or use Flex Capacity to increase spare space/OP - both Crucial and Micron versions of Storage Executive cannot do much with the drive. I can't even get nvme-cli or msecli to give me proper stats or info on the drive. I was thinking of using the drive in my Pi5 running InfluxDB - but after reading this post I am going to try and get a proper non QLC 2TB SSD. In short this is a piece of junk that I unfortunately paid too much for ;-(

 

[DaveLTX]

Write amplification is likely the missing bit, from what I understand ZFS is often kinda weak there, as are low-end SSD controllers, I would assume that the NAND has taken 2-3x as many writes as the SMART host writes stat indicates. If it's actually the age of the write causing read performance to drop then that's a sign that the NAND is losing charge and requiring re-reads to pass checksum, in which case toss that thing right now.

Early TLC suffered the same or worse(Samsung 840 and evo... ouch) with cold data.
I really would tread carefully with QLC, in fact I vowed never to touch QLC with a 10 foot pole

 

[twin_savage]

That sounds interesting, I'd like to read that! Is this the thread? Ssd data retention

Yes, that's the one. It goes off the rails a couple of times but in general there is some good information on there.

 

[twin_savage]

Early TLC suffered the same or worse(Samsung 840 and evo... ouch) with cold data.
I really would tread carefully with QLC, in fact I vowed never to touch QLC with a 10 foot pole

The Samsung 840 evo situation from years ago was interesting because the drive's controller was just being overly sensitive in applying all the extra degraded read tricks when they weren't truly necessary due to an errant calibration value. Presumably those drives would slow down to the same degree with time with the more functional firmware on them, but that amount of time should have been measured in years.

If the SSD manufactures would just add a weak cell refresh algorithm into garbage collection I'd be satisfied with even QLC....... but most of them don't seem to be doing that and in the minority of cases where the manufactures are, they don't advertise it.