replace cedardb docker image with native cedardb and update measurements

Author: lukasvogel

cedardb-parquet/benchmark.sh

@@ -2,4 +2,8 @@
 # Thin shim — actual flow is in lib/benchmark-common.sh.
 export BENCH_DOWNLOAD_SCRIPT="download-hits-parquet-single"
 export BENCH_DURABLE=yes
+export PGHOST="/tmp"
+export PGUSER=postgres
+export PGDATABASE=postgres
+
 exec ../lib/benchmark-common.sh

cedardb-parquet/check

@@ -1,4 +1,4 @@
 #!/bin/bash
 set -e
 
-PGPASSWORD=test psql -h localhost -U postgres -c 'SELECT 1' >/dev/null
+psql -c 'SELECT 1' >/dev/null

cedardb-parquet/create.sql

@@ -105,4 +105,4 @@ SELECT
     "RefererHash"           AS RefererHash,
     "URLHash"               AS URLHash,
     "CLID"                  AS CLID
-FROM '/data/hits.parquet';
+FROM 'hits.parquet';

cedardb-parquet/data-size

@@ -2,4 +2,4 @@
 set -eu
 
 # No ingestion — reported size is the parquet file itself.
-stat -c%s data/hits.parquet
+stat -c%s hits.parquet

cedardb-parquet/install

@@ -2,9 +2,17 @@
 set -eu
 
 sudo apt-get update -y
-sudo apt-get install -y docker.io postgresql-client
+sudo apt-get install -y postgresql-client
 
-sudo docker pull cedardb/cedardb:latest
+# Stop any running instance before reinstalling
+./stop || true
 
-mkdir -p data db
-chmod -R 777 data db
+# Delete potential previously created database
+rm -rf ./db
+
+mkdir -p "${XDG_CONFIG_HOME:-$HOME/.config}/systemd/user"
+
+curl https://get.cedardb.com | bash -s -- -y \
+    --install-dir "$PWD" \
+    --db-dir "$PWD/db" \
+    --with-systemd=user

cedardb-parquet/load

@@ -1,12 +1,8 @@
 #!/bin/bash
 set -eu
 
-# Stage parquet file under ./data so the docker container sees /data/hits.parquet.
-mkdir -p data
-mv hits.parquet data/
-chmod -R 777 data
-
-# create.sql defines a view over the parquet file — no ingestion needed.
-PGPASSWORD=test psql -h localhost -U postgres -t < create.sql
+# Substitute the absolute path so CedarDB can find the file regardless of
+# its working directory.
+sed "s|'hits\.parquet'|'$PWD/hits.parquet'|" create.sql | psql
 
 sync

cedardb-parquet/query

@@ -8,7 +8,7 @@ set -e
 
 query=$(cat)
 
-raw=$(PGPASSWORD=test psql -h localhost -U postgres -t -c '\timing' -c "$query" 2>&1) && exit_code=0 || exit_code=$?
+raw=$(psql -t -c '\timing' -c "$query" 2>&1) && exit_code=0 || exit_code=$?
 
 if [ "$exit_code" -ne 0 ] || printf '%s\n' "$raw" | grep -qE '^ERROR|psql: error'; then
     printf '%s\n' "$raw" >&2

cedardb-parquet/start

@@ -1,19 +1,4 @@
 #!/bin/bash
 set -eu
 
-if PGPASSWORD=test psql -h localhost -U postgres -c 'SELECT 1' >/dev/null 2>&1; then
-    exit 0
-fi
-
-sudo docker stop cedardb >/dev/null 2>&1 || true
-sudo docker rm cedardb >/dev/null 2>&1 || true
-
-sudo docker run -d --rm -p 5432:5432 \
-    -v "$(pwd)/data:/data" \
-    -v "$(pwd)/db:/var/lib/cedardb/data" \
-    -e CEDAR_PASSWORD=test \
-    --name cedardb cedardb/cedardb:latest >/dev/null
-
-until pg_isready -h localhost --dbname postgres -U postgres >/dev/null 2>&1; do
-    sleep 1
-done
+systemctl --user start cedardb.service

cedardb-parquet/stop

@@ -1,3 +1,3 @@
 #!/bin/bash
 
-sudo docker stop cedardb >/dev/null 2>&1 || true
+systemctl --user stop cedardb.service || true

cedardb/benchmark.sh

@@ -1,5 +1,9 @@
 #!/bin/bash
 # Thin shim — actual flow is in lib/benchmark-common.sh.
-export BENCH_DOWNLOAD_SCRIPT="download-hits-tsv"
+export BENCH_DOWNLOAD_SCRIPT="download-hits-parquet-single"
 export BENCH_DURABLE=yes
+export PGHOST="/tmp"
+export PGUSER=postgres
+export PGDATABASE=postgres
+
 exec ../lib/benchmark-common.sh