With repro going on, we're getting a few V&V hiccups. It looks like most are due to asp L1 products getting ingested in mica before obspars are available. These PR changes 1) set the obspar ingest to run after the L1 ingest in the task schedule, 2) make the processing more robust to a missing obspar (log and skip).
Testing
[x] Passes unit tests on MacOS, linux, Windows (at least one required)
[ ] Functional testing
I did not put time into creating a good functional test. As this is a race condition situation, one would need to construct a state in the various databases matching the mismatch seen at the time of the V&V hiccup. Given the benign nature of the changes, I think testing that this fixes the problem by just running the code in flight (and looking for warnings) makes sense.
Description
Avoid vv obspar race condition.
With repro going on, we're getting a few V&V hiccups. It looks like most are due to asp L1 products getting ingested in mica before obspars are available. These PR changes 1) set the obspar ingest to run after the L1 ingest in the task schedule, 2) make the processing more robust to a missing obspar (log and skip).
Testing
I did not put time into creating a good functional test. As this is a race condition situation, one would need to construct a state in the various databases matching the mismatch seen at the time of the V&V hiccup. Given the benign nature of the changes, I think testing that this fixes the problem by just running the code in flight (and looking for warnings) makes sense.
Fixes #