keep_solution_during_restore should be able to also keep the auxiliary system

[vincentlaboure]

Bug Description

For some complicated setups, keep_solution_during_restore can fail, which seems to be related to the auxiliary system not being restored.

Steps to Reproduce

This simple sabertooth example can reproduce the issue.

Impact

This can make the computation of a steady state solution painfully and unnecessarily slow.

Tag @fdkong @jortensi @nicolasmartin3

[GiudGiud]

Tagging myself. This will be part of the checkpoint rework effort.

[YaqiWang]

@loganharbour can this be closed with the latest update in checkpoint restart?

[loganharbour]

I think so - we now only load what we can, so it shouldn't error. That sounds satisfied, right?

We still have the outstanding issue (if desired) of loading data into other names/vectors (i.e., non direct mapping) but I don't think this issue pertains to that.

[loganharbour]

I think so - we now only load what we can, so it shouldn't error. That sounds satisfied, right?

We still have the outstanding issue (if desired) of loading data into other names/vectors (i.e., non direct mapping) but I don't think this issue pertains to that.

[vincentlaboure]

can this be closed with the latest update in checkpoint restart?

I am not sure to understand how this is related to checkpoint restart. Is the claim that that the recent checkpoint work solved this issue?

[loganharbour]

I might need more of an explanation on this one. I actually think that they're not related now

[GiudGiud]

I ll take a look over the next few weeks. There's a VTB model with a workaround but I would be down not to have that