Warmed SSD Init Skip¶
Motivation¶
Dashboard-managed sharding and node warmup can already prove that every local-SSD shard cache is ready before submit. The sibling ElasticBLAST runtime still recreated and waited on init-ssd-* setup Jobs, making submit appear stuck even when the warmed cache was complete.
User-facing change¶
Sharded BLAST submits that pass the dashboard warmup readiness gate now write an opt-in ElasticBLAST config flag so the sibling runtime can skip redundant local-SSD DB initialization. If the sibling runtime cannot verify the dashboard warmup Jobs, it falls back to the existing setup Job path.
API/IaC diff summary¶
- Add
[cluster] exp-skip-warmed-ssd-init = trueto generated configs only when the submit warmup gate returned Ready. - No API route or IaC changes.
- Requires the sibling
elastic-blast-azurechange that verifiesapp=elb-db-warmupJobs before skippinginit-ssd-*.
Validation evidence¶
PYTHONPATH=src pytest -q tests/azure/test_db_partitioning.py::TestPartitionConfig::test_cluster_skip_warmed_ssd_init_default tests/azure/test_db_partitioning.py::TestInitializeLocalSsdSharded::test_skips_init_jobs_when_dashboard_warmup_readyin sibling repo: 2 passed.PYTHONPATH=src pytest -q tests/azure/test_db_partitioning.pyin sibling repo: 40 passed.uv run pytest -q api/tests/test_blast_config_sharding.py api/tests/test_blast_tasks.py: 138 passed.uv run ruff check api/services/blast_config.py api/tasks/blast/__init__.py api/tests/test_blast_config_sharding.py: passed.