NFS over ipv6 performs better?

[mtsm]

hi , i have a solution being built on suses linux and i noticed that the nfs interface is using ipv6 and the rest of the soluton ipv4. The person that did it affirms that there's a significant performance gain using nfs over ipv6 so thats the reason of it. I'm questioning myself what's the technical reason , this is just making the environment more complex until now.

[Irithori]

Tbh, I dont think that ipv6 alone grants a nfs performance gain.
It is just another layer beneath nfs.

But why guess?
I would claim a performance comparison from this person.

[mtsm]

Tbh, I dont think that ipv6 alone grants a nfs performance gain.
It is just another layer beneath nfs.

But why guess?
I would claim a performance comparison from this person.

thats what i think , the thing is its a little bit more political , i bet the person wont give me any results , i would have to setup my own environment which i dont have time. i dont see any reason ipv6 would increase performance. as im not a specialist on ipv6 i cant afirm that for sure , but my understanding is that ipv6 is just the addressing scheme so i dont see any relation , its just making my OS and storage layer more complex.

[Irithori]

Unfortunately I also know this game.
Someone makes a claim and if you challenge this claim, it is somehow your duty to debunk it.

My suggestion is:
Engage in a email -or otherwise written- conversation.
Ask to repeat the statements,
also ask for that performance comparison.

1) If there is no response: Then you at least have proof, that you had concerns early
2) If that person tries to shift work your way: Either deny it (if the work is not within your scope) or use the discussion to justify the additonal work with your manager
3) If there would be a valid and sound explanation, and it actually *does* give more performance: All the better.

But -again- I doubt 3). I see no reason, why it would be faster.


TL;DR: Documented communication as proof is worth gold