Continue a Markov chain from previous simulation

I would like to be able to continue a Markov chain from previous simulation.
My attempt is to start from the end posterior value of previous run with warmup =0, but am told that this may not be the correct method to follow.

I discussed this previously here,

I am moving the conversation to “rstan” as advised.

Any help is appreciated

Page 46 of the rstan Reference Manual on CRAN address the mass matrix or metric. However it appears that the variable we want to control is only settable as a type using a choice of Strings.

The CmdStan guide on 49 has the same argument metric but there we can pass in values.

What isn’t readily apparent in the Guide is how to extract the mass matrix as described as