the problem is that not only the memory needs ecc functionality. the motherboard and the cpu all need as well. so even if the memory was only 11% more expensive then there is still the extra costs ( inflated or not) of the rest of the components to deal with.
For my data store. I refuse to use hardware that is not assessed to be ECC capable.
since i cant edit my post yet;
I meant headware I have confirmed to be ECC functional.
This leaves me with a bit of a headache.
If I need to buy new hardware in a year or 15. then how to I confirm functionality by then? If all the current vendros seem to be cool with this cult.
I hope risk-v by then might offer better support and if not I will build it myself 
Where you the crazy man who soldered onto dimms to physically introduce errors & actually confirm that ecc actually indeed does something on consumer grade ryzen cpus? If so you were a huge reason I actually went with the system I did. Thank you!
Iāll quibble with this:
ECC errors can corrupt both ZFS and other file systems.
However, ZFS users are expecting more longevity and reliability from their data because they have used RAIDZ, and ZFS has checksums.
Data backups donāt work if the data is corrupted in memory before it is stored on drives.
While ECC isnāt strictly necessary for running ZFS, itās by no means useless. Similarly, you could jump out of an aeroplane without a parachute. It is by no means necessary for jumping. Just donāt think too much about landing.
I probably was not as I never soldered.
but I think I might have been been the pioneer there as I just went in with wires.
and once confirmed it enables others to do the same.
And then some dude might have actually started soldering
and I you are welcome.
Your thanks is the icing on the cake.
The cake it self was for for some us to be able if our systems were actually ECC funcational
truest be told though that the real trophy goes to those guys that that went in from a scientific angle demonstrating how broken and riddled with backdoors we are still are 
they showed how to trigger errors and then went on to a length explentation on how architectures back then were vulnerable.
So all I did is just get frustrated enough and then dive in myself.
And the rest is history.
Was a few years now, so maybe I misremember if it was soldered or notā¦ But I 100% remember reading through the post & seeing some madman with wires sticking out of his dimms to introduce errors physically. So Iām very certain it was you 
Thanks again for the work as Iām sure Iām not the only person who was happy to get proof. Glad to see you on the forums again!
yessss that was me with those wires being in and or near where they were never meant to be.
it was taking one for the team
Now I will try again. years later. That what I set out to do years ago.
Please owners of the software;
How?
The way you official guys told me off over different fora is painfull.
While the real question remains.
is ECC usefull if we dont know ti is working?
Now of course the answer will be NO.
that is because I formulated the question deliberatly.
How do we get past this clear good and valid point?
oops. I messed up.
ECC can be usefull even if we can nto know.
But my point is is the trust valid if we can not make sure it is working
btw this is also me taking a stab at commerce.
I just dont like sales pitches like
Just buy A or B and then you will be bliss.
Syntax Error. how do I asses it for my self please?
This is a stab at marketing
and to avoid confusion.
I will once again state. I want ECC and I beleive I need it.
In run more than 50 Supermicro servers with ECC and every once in a while one of them throws visible ECC warnings from which I infer by statistics that it must be working as intended.
Even TrueNAS CORE shows ECC warnings in the UI.
Do you assess that your CPU implements the entire documented command set? Do you measure output and efficiency of your power supply or do you buy 750 watts when you need that much?
Why do you assume when a motherboard and a memory module both have ECC on the data sheet, the vendors are probably (?) lying to you? Thatās what specs are for.
I assume nothing. In fact I dont trust when vendors tell on their spec sheets it si all working.
you know why?
and now finally you will have to agree.
Did you assess it your self?
This is now longer starting to be productive.
I feel it now is drying up into 2 camps.
I made my case and I have heard no good point yet to wake me up.
I am getting ECC warnings from some of the servers whenever an ECC event occurs. That implies itās working on them.
I have not manually assessed every single identical mainboard and memory module, no. I infer by induction that if some of the servers on which ECC events actually occur properly notify me, it must very very probably be working for all of them and the other ones simply did not experience any events yet.
I am not messing with copper wires in the DIMM slots of machines costing several thousand voiding my warranty.
Actually - no.
so you and I are on the same page after all.
but still you slide by the assess-ability
why do you refuse to see the point?
I fully expect it to be working when I explicitly buy it and experience has shown that it does - so far.
and this is where it all falls down. where it crumbles like a house of cards.
it might have worked one day. but the next? can we assess it is still working?
I am sorry to categorize you as religious. I am pretty certain you are more than that.
But when people actively try to side step my point then what else am I to think of you?
