On: Running Postgres as a Unikernel

[arghwhat]

Unikernels are cool, but half this official post is pointlessly throwing rocks at postgres, the other half having to hack around to get this extremely common piece of software working with your own product. No thanks, I will get my unikernel fix from someone else.

Plus, if you often have to LD_PRELOAD getuid, just add a bogus impl to your libc or as a syscall to reduce onboarding friction.

[anonymous_sorry → arghwhat]

Is it throwing rocks?

"I'm not knocking postgres here. You have to keep in mind [a paragraph of contextualising postgres's design decisions]"

[aljarry → anonymous_sorry]

The rant in the middle is very dismissive over voices against in the discussion, and that makes me agree with OP.

[giraffe_lady → anonymous_sorry]

Saying you're not doing something is not actually the same as not doing it.

[richbell → giraffe_lady]

I'm not disagreeing with you, but I think you're completely wrong about that.

[eyberg → arghwhat]

If you think I was throwing rocks that most definitely was not the intention. The whole point of writing the post was to point out all the abstractions that don't quite gel well in unikernel land.

The LD_PRELOAD wasn't used for this particular port but it does get used sometimes . We don't actually have real uid calls but some applications will require specific ones that we stub from time to time. Very very few apps actually require this.