We were showing Dan the h1fw0 problems (after turning on all frame writing again) and after less than an hour h1fw0 became stable again and has been running for 2+ hours. We will leave it in this state for now, but if it becomes very unstable again we have several options:

Turn off LDAS frame comparison jobs (reads every frame on both writers)

Remove all LDAS archiving from h1fw0 over to h1fw1

Reduce h1fw0 back to only writing science frames

Fingers crossed we don't have to do any of these.