DuckDB inside OLTP databases

PostgreSQL

Versions

Run asImageServerDuckDBDigest
PostgreSQLpostgres:18PostgreSQL 18n/asha256:edb0c08dc33befd0262ab3b5cf1c4f3f8cfd8b09e88c05b45adee29d6761ddb6

Observed in this run

Generated the same way for every system from its result files: record counts per measure and status, and the reasons given for anything that is not ok.

PostgreSQL

Settings that differ from the defaults

SettingValueReasonFile
listen_addresses'*'The runner container connects over the docker bridge.systems/postgres/postgresql.conf
shared_buffers16GB25% of container memory, the usual starting point; the OS page cache holds the rest.systems/postgres/postgresql.conf
effective_cache_size48GBPlanner's estimate of buffer plus page cache: 75% of container memory.systems/postgres/postgresql.conf
max_parallel_workers_per_gather16Allow one query to use every core given to the database.systems/postgres/postgresql.conf
max_parallel_workers16Cap on parallel workers across all sessions: one per core.systems/postgres/postgresql.conf
max_worker_processes16Must be at least max_parallel_workers for those workers to start.systems/postgres/postgresql.conf
work_mem256MBAnalytical sorts and hashes stay in memory instead of spilling at the 4MB default.systems/postgres/postgresql.conf
max_connections200Headroom for the 32-stream throughput run plus probes and cancel connections.systems/postgres/postgresql.conf
max_wal_size16GBWAL of mem_gb/4 keeps checkpoints from throttling bulk load at the 1GB default.systems/postgres/postgresql.conf
checkpoint_timeout30minTimed checkpoints every 30 minutes, so load and query runs are not interrupted by them.systems/postgres/postgresql.conf

README

PostgreSQL 18 baseline

Image postgres:18 (digest recorded in every result as versions.image_digest), started with -c config_file=/etc/postgresql/postgresql.conf, the file rendered from postgresql.conf and mounted read-only. The dataset is mounted read-only at /data, shm_size is 8 GB, and authentication is trust on the private docker network.

Non-default settings

Rendered from the container limits (dbench/systems/baseline_conf.py); at 64 GB and 16 cores:

SettingValueReason
listen_addresses*The runner container connects over the docker bridge (the custom config file replaces the default one).
shared_buffers0.25 x container GB = 16GBThe usual starting point; the OS page cache holds the rest.
effective_cache_size0.75 x container GB = 48GBPlanner's estimate of buffer plus page cache.
max_parallel_workers_per_gathercores in the DB cpuset = 16One query may use every core given to the database.
max_parallel_workers16Cap across sessions: one worker per core.
max_worker_processes16Must be at least max_parallel_workers.
work_mem256MBAnalytical sorts and hashes stay in memory instead of spilling at the 4MB default.
max_wal_sizemem_gb/4 = 16GBKeeps checkpoints from throttling bulk load at the 1GB default.
checkpoint_timeout30minTimed checkpoints do not interrupt load and query runs.
max_connections200Headroom over the 32-stream throughput run plus probe and cancel connections.

Everything else stays at the PostgreSQL 18 default (including JIT).

Loading and measurement notes

Schema

TPC-H tables with primary keys, plus a plain index on every foreign-key column (row-store baseline rule); no other indexes and no declared foreign-key constraints. Strings are VARCHAR(n) with the spec lengths. The default collation of the image is kept. There are no schema deviations.

Dialect

The 22 DuckDB query texts run unchanged.

M11 and M12 are n/a: there is no DuckDB copy to convert or push down to.