u cannot say all 80TB as cricital unless its critical biz like amazon - in that case money price not matters
u can easy make array of 8TB NVMe X 10 pieces - one or two single 50TB is obviously much risk single point of failure - normal personal use max 1TB is enough to be called critical
this commercial profit making biz use - speed is essential n money not matters 
Have you ever had 50MPx camera and just enough time to go for holiday but not enough to trim your growing backlog of 9xHDR photos?
I don’t even mention video cam of your partner?
With 3x redundancy that’s not even 30TB of important shit
professional money making photogrpahy will say
why i always reduce RES on phone cam - default sucks they push u buy more disk
or call it an expensive hobby - still if u spend - better make OWN customer NGFF array LIGHT n less risk than HDD but ar expensive
weight n size far far LESS that what u have now
100% with you. Don’t even mention power consumption.
if u can keep online - i consider (case purpose basis) keeping stick offline (no electric) mount when need - personal normal usercase - stuff like continous high RW IOPS not need much space than 100GB ususally do on NVMe or DRAM replace when degrade to 70%
I think we can agree that ‘normal’ is not who I am
Ditto, but i don’t have that much digital storage
yup 100TB of essential is NOT ACCEPTED as normal for personal user case. LoL
extraordinary requirement needs expensive solution - using HDD in this age HIGHEST RISK
array LVM is MUST - also buying anything > 1TB for normal use super risk
rather 8 pieces than one 8TB - chip or board can fail too.
enterprise system different story
Nah, shitty speed - yes.
With multiple servings nodes still borderline acceptable. (Though running 15 physical server at home is also not really normal or cheap as running costs go.
Risk - nots so sure. 3x redundancy plus backup makes it quite safe. Though offsite backup of this amount is anything but cheap.
1. i consider risk one large disk / stick failure more important n loss hence buy breakup smaller ones given i dodnot need super high parallel capacity
2. ur data is money making n critical - have layered solution - like if u one offsite then HDD in mainsite n offsite both are fine - unlike both will fail - DR check is must
100% with you
Only
- It is hard to not put „critical” tag on memories or on unhappiness of the wife (in case of loss)
- my setup is almost 10y old. Technology moves faster than my finances grow.
- one day I would love to be all on energy’s efficient boards and NVMe ( though I bet as soon as I start buying new shit - technology will move again)
- don’t try to fix what is not broken
also i take shitty speed low cost but high endurance when data is not needed to accessed 24/7 but on demand only manually
