Fix restart error on maint-2.0

[peterdschwartz]

Add endwb to ELM restart file in maint-2.0.

[BFB]
Fixes #6195

[peterdschwartz]

Need to check if maint-2.1 will also need PR #5811

[peterdschwartz]

I ran SMS_D_Ld1.ne30pg2_EC30to60E2r2.WCYCLSSP370.pm-cpu_intel.allactive-wcprodssp with no issue, so leaving this as is.

[ndkeen]

I might suggest adding in the fix in 5811. I think I needed that for the tests in maint-2.1 to pass. I had meant to add it here as well anyway, but the e3sm_production tests did pass.

[peterdschwartz]

Went ahead and made that change as well.

[ndkeen]

OK, you likely know more about why this is or is not a good idea.

I did try SMS_Ld1.ne30pg2_EC30to60E2r2.WCYCLSSP585.pm-cpu_intel.allactive-wcprodssp
with and without that finidat and it seemed ok?

[rljacob]

@ndkeen can an ERS test pass on maint-2.0 without these fixes?

[ndkeen]

No, before these changes ERS_Lm3.ne4_oQU240.F2010 would fail. But it also fails with even old maint-2.0 checkouts -- ie I think the test would have always failed.

[rljacob]

Oh that's right, you have to run it for a long time. A standard-length ERS passes. How long does Lm3 take? We should add an ERS test to the prod suite for maint-2.0 to catch this.

[ndkeen]

My test that passes after this fix used a job that ran for a total of 8:45 on 1 node of pm-cpu.

[mingxuanwupnnl]

@peterdschwartz I met the same issue while I was doing production runs. I wonder if this bug fix PR is BFB? Thank you.

[peterdschwartz]

@peterdschwartz I met the same issue while I was doing production runs. I wonder if this bug fix PR is BFB? Thank you.

It's BFB

[rebassoo]

@ndkeen, will this PR be merged into maint-2.0 soon? What additional reviews/checks are necessary? thanks

[rljacob]

@rebassoo merged.

[rebassoo]

Great, thanks @rljacob!