fix: ensure result correctness when enable pruning cache (#17524)

* reproduce * reproduce * reproduce * reproduce * fix: ensure result correctness when enable pruning cache
databendlabs · Feb 27, 2025 · 7b27c3a · 7b27c3a
1 parent aeaf443
commit 7b27c3a
Show file tree

Hide file tree

Showing 5 changed files with 65 additions and 3 deletions.
diff --git a/scripts/ci/ci-run-stateful-tests-cluster-minio.sh b/scripts/ci/ci-run-stateful-tests-cluster-minio.sh
@@ -20,6 +20,8 @@ export STORAGE_ALLOW_INSECURE=true
 
 echo "Install dependence"
 python3 -m pip install --quiet mysql-connector-python requests
+sudo apt-get update -yq
+sudo apt-get install -yq lsof
 
 echo "calling test suite"
 echo "Starting Cluster databend-query"

diff --git a/scripts/ci/ci-run-stateful-tests-standalone-minio.sh b/scripts/ci/ci-run-stateful-tests-standalone-minio.sh
@@ -20,6 +20,8 @@ export STORAGE_ALLOW_INSECURE=true
 
 echo "Install dependence"
 python3 -m pip install --quiet mysql-connector-python requests
+sudo apt-get update -yq
+sudo apt-get install -yq lsof
 
 echo "calling test suite"
 echo "Starting standalone DatabendQuery(debug)"

diff --git a/src/query/storages/fuse/src/operations/read_partitions.rs b/src/query/storages/fuse/src/operations/read_partitions.rs
@@ -170,7 +170,7 @@ impl FuseTable {
         plan: &DataSourcePlan,
         source_pipeline: &mut Pipeline,
     ) -> Result<Option<Pipeline>> {
-        let segments_location = plan.statistics.snapshot.clone();
+        let snapshot = plan.statistics.snapshot.clone();
         let table_schema = self.schema_with_stream();
         let dal = self.operator.clone();
         let mut lazy_init_segments = Vec::with_capacity(plan.parts.len());
@@ -180,7 +180,7 @@ impl FuseTable {
                 lazy_init_segments.push(SegmentLocation {
                     segment_idx: lazy_part_info.segment_index,
                     location: lazy_part_info.segment_location.clone(),
-                    snapshot_loc: segments_location.clone(),
+                    snapshot_loc: snapshot.clone(),
                 });
             }
         }
@@ -202,7 +202,7 @@ impl FuseTable {
                 .map(|push_downs| {
                     format!(
                         "{:x}",
-                        Sha256::digest(format!("{:?}_{:?}", segments_location, push_downs))
+                        Sha256::digest(format!("{:?}_{:?}", lazy_init_segments, push_downs))
                     )
                 });
 

diff --git a/tests/suites/1_stateful/02_query/02_0010_issue_17523.result b/tests/suites/1_stateful/02_query/02_0010_issue_17523.result
@@ -0,0 +1,14 @@
+1
+1
+1
+1
+test_text1
+test_text2
+test_text3
+test_text4
+2
+test_text1
+test_text2
+test_text3
+test_text4
+3
diff --git a/tests/suites/1_stateful/02_query/02_0010_issue_17523.sh b/tests/suites/1_stateful/02_query/02_0010_issue_17523.sh
@@ -0,0 +1,44 @@
+#!/usr/bin/env bash
+
+CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
+. "$CURDIR"/../../../shell_env.sh
+
+# Skip in standalone mode
+port_check=$(sudo lsof -i :9093)
+if [ -z "$port_check" ]; then
+    echo -e "1\n1\n1\n1\ntest_text1\ntest_text2\ntest_text3\ntest_text4\n2\ntest_text1\ntest_text2\ntest_text3\ntest_text4\n3"
+    exit 0
+fi
+
+echo "drop table if exists table_test_02_0010;" | $BENDSQL_CLIENT_CONNECT
+
+echo "create table table_test_02_0010(TEXT String);" | $BENDSQL_CLIENT_CONNECT
+
+
+table_inserts=(
+  "INSERT INTO table_test_02_0010 VALUES('test_text1');"
+  "INSERT INTO table_test_02_0010 VALUES('test_text2');"
+  "INSERT INTO table_test_02_0010 VALUES('test_text3');"
+  "INSERT INTO table_test_02_0010 VALUES('test_text4');"
+)
+
+for i in "${table_inserts[@]}"; do
+  echo "$i" | $BENDSQL_CLIENT_CONNECT
+done
+
+
+echo "select * from table_test_02_0010 order by text;" | $BENDSQL_CLIENT_CONNECT
+
+# kill a node
+sudo lsof -i :9093 -t | xargs -r sudo kill -9
+
+echo "select count(*) from system.clusters;" | $BENDSQL_CLIENT_CONNECT
+
+echo "select * from table_test_02_0010 order by text;" | $BENDSQL_CLIENT_CONNECT
+
+# restart the node
+ROOTDIR="$(cd "$(dirname "${BASH_SOURCE[0]}")"/../../../../ && pwd)"
+env "RUST_BACKTRACE=1" nohup "$ROOTDIR"/target/${BUILD_PROFILE}/databend-query -c "$ROOTDIR"/scripts/ci/deploy/config/databend-query-node-3.toml --internal-enable-sandbox-tenant > "$ROOTDIR"/.databend/query-3.out 2>&1 &
+python3 "$ROOTDIR"/scripts/ci/wait_tcp.py --timeout 30 --port 9093 > /dev/null 2>&1
+
+echo "select count(*) from system.clusters;" | $BENDSQL_CLIENT_CONNECT