Skip to content

Commit

Permalink
More test runs for more DBMS
Browse files Browse the repository at this point in the history
  • Loading branch information
perdelt committed Oct 22, 2024
1 parent 7f0079b commit 6fd6ca3
Show file tree
Hide file tree
Showing 25 changed files with 3,926 additions and 72 deletions.
6 changes: 5 additions & 1 deletion docs/TestCases.md
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
# Test Cases

There is a variety of combination of options to be tested.
There is a variety of combinations of options to be tested.

We here list some more basic use cases to test the functionality of bexhoma.

Expand All @@ -12,6 +12,10 @@ BEXHOMA_NODE_LOAD="cl-worker19"
BEXHOMA_NODE_BENCHMARK="cl-worker19"
```

See also [more test cases](https://github.com/Beuth-Erdelt/Benchmark-Experiment-Host-Manager/blob/master/test-more.sh) for more and longer running test cases.

See the [log folder](https://github.com/Beuth-Erdelt/Benchmark-Experiment-Host-Manager/tree/master/logs_tests) for some demo test logs.
The folder also contains `*_summary.txt` files containing only the result summary.

## TPC-H

Expand Down
38 changes: 19 additions & 19 deletions logs_tests/test_benchbase_testcase_mariadb_1.log
Original file line number Diff line number Diff line change
Expand Up @@ -2,8 +2,8 @@ Data Directory : is running
Result Directory : is running
Dashboard : is running
Message Queue : is running
Experiment : has code 1728550061
Experiment : starts at 2024-10-10 10:47:40.851313 (3286000.975294404)
Experiment : has code 1729530658
Experiment : starts at 2024-10-21 19:10:57.837122 (4266597.961103178)
Experiment : This experiment compares run time and resource consumption of Benchbase queries in different DBMS.
Benchbase data is generated and loaded using several threads.
Benchmark is 'tpcc'. Scaling factor (e.g., number of warehouses) is 16. Benchmarking runs for 5 minutes. Target is based on multiples of '1024'. Factors for benchmarking are [8].
Expand All @@ -21,10 +21,10 @@ done
MariaDB-1-1-1024 : will start now
- waiting 30s - : done
MariaDB-1-1-1024 : is not loaded yet
MariaDB-1-1-1024 : will start loading but not before 2024-10-10 08:49:41 (that is in 60 secs)
MariaDB-1-1-1024 : will start loading but not before 2024-10-21 17:12:58 (that is in 60 secs)
- waiting 30s - : done
MariaDB-1-1-1024 : is not loaded yet
MariaDB-1-1-1024 : will start loading but not before 2024-10-10 08:49:41
MariaDB-1-1-1024 : will start loading but not before 2024-10-21 17:12:58
done
MariaDB-1-1-1024 : is not loaded yet
MariaDB-1-1-1024 : start asynch loading scripts of type loaded
Expand All @@ -47,17 +47,17 @@ MariaDB-1-1-1024 : is loading
MariaDB-1-1-1024 : is loading
- waiting 30s - : done
MariaDB-1-1-1024 : is loading
- waiting 30s - : done
MariaDB-1-1-1024 : is loading
done
MariaDB-1-1-1024 : showing loader times
MariaDB-1-1-1024 : loader timespan (first to last [s]) = 316
MariaDB-1-1-1024 : loader timespan (first to last [s]) = 285
MariaDB-1-1-1024 : benchmarks done 0 of 1. This will be client 1
MariaDB-1-1-1024 : we will change parameters of benchmark as {'PARALLEL': '1', 'SF': '16', 'BENCHBASE_BENCH': 'tpcc', 'BENCHBASE_PROFILE': 'mariadb', 'BEXHOMA_DATABASE': 'benchbase', 'BENCHBASE_TARGET': 8192, 'BENCHBASE_TERMINALS': 16, 'BENCHBASE_TIME': 300, 'BENCHBASE_ISOLATION': 'TRANSACTION_READ_COMMITTED'}
MariaDB-1-1-1024-1 : start benchmarking
MariaDB-1-1-1024 : benchmarking results in folder /home/perdelt/benchmarks/1728550061
MariaDB-1-1-1024 : benchmarking results in folder /home/perdelt/benchmarks/1729530658
- waiting 10s - : done
MariaDB-1-1-1024 : benchmarking is waiting for job bexhoma-benchmarker-mariadb-1-1-1024-1728550061-1-1-87z5z: found
MariaDB-1-1-1024 : benchmarking is waiting for job bexhoma-benchmarker-mariadb-1-1-1024-1729530658-1-1-zx9zz: found
- waiting 30s - : done
MariaDB-1-1-1024 : has running benchmarks
- waiting 30s - : done
MariaDB-1-1-1024 : has running benchmarks
- waiting 30s - : done
Expand All @@ -79,18 +79,18 @@ MariaDB-1-1-1024 : has running benchmarks
done
MariaDB-1-1-1024 : has running benchmarks
MariaDB-1-1-1024-1 : showing benchmarker times
MariaDB-1-1-1024-1 : benchmarker timespan (start to end single container [s]) = 312
MariaDB-1-1-1024-1 : benchmarker times (start/end per pod and container) = [(1728550519, 1728550821)]
MariaDB-1-1-1024-1 : found and updated times {'benchmarker': [(1728550519, 1728550821)]}
MariaDB-1-1-1024-1 : benchmarker timespan (start to end single container [s]) = 342
MariaDB-1-1-1024-1 : benchmarker times (start/end per pod and container) = [(1729531086, 1729531394)]
MariaDB-1-1-1024-1 : found and updated times {'benchmarker': [(1729531086, 1729531394)]}
done
MariaDB-1-1-1024 : can be stopped
Experiment ends at 2024-10-10 11:01:00.176647 (3286800.300626184): 799.3253317796625s total
Experiment ends at 2024-10-21 19:24:17.313984 (4267397.437962507): 799.4768593292683s total
Experiment : downloading partial results
Experiment : uploading full results
Experiment : uploading workload file
Benchmarking connection ... Latency Distribution.Average Latency (microseconds)
connection_pod ...
MariaDB-1-1-1024-1-1 MariaDB-1-1-1024-1 ... 102151
MariaDB-1-1-1024-1-1 MariaDB-1-1-1024-1 ... 79907

[1 rows x 36 columns]
Workflow {'MariaDB-1-1-1024': [[1]]}
Expand All @@ -102,7 +102,7 @@ Result workflow complete
Benchbase Workload SF=16 (warehouses for TPC-C)
Type: benchbase
Duration: 800s
Code: 1728550061
Code: 1729530658
This includes no queries. Benchbase runs the benchmark
This experiment compares run time and resource consumption of Benchbase queries in different DBMS.
Benchbase data is generated and loaded using several threads.
Expand All @@ -124,14 +124,14 @@ MariaDB-1-1-1024-1 uses docker image mariadb:11.4.2
Cores:64
host:5.15.0-116-generic
node:cl-worker11
disk:251659872
datadisk:1942868
disk:252259184
datadisk:1935700
requests_cpu:4
requests_memory:16Gi

### Execution
experiment_run terminals target pod_count time Throughput (requests/second) Latency Distribution.95th Percentile Latency (microseconds) Latency Distribution.Average Latency (microseconds)
MariaDB-1-1-1024-1 1 16 8192 1 300.0 156.6 31335.0 102151.0
MariaDB-1-1-1024-1 1 16 8192 1 300.0 157.86 30764.0 79907.0

Warehouses: 16

Expand All @@ -145,7 +145,7 @@ DBMS MariaDB-1-1-1024 - Pods [[1]]

### Loading
time_load terminals pods Imported warehouses [1/h]
MariaDB-1-1-1024-1 316.0 1.0 1.0 182.278481
MariaDB-1-1-1024-1 285.0 1.0 1.0 202.105263

### Tests
TEST passed: Throughput (requests/second) contains no 0 or NaN
Expand Down
10 changes: 5 additions & 5 deletions logs_tests/test_benchbase_testcase_mariadb_1_summary.txt
Original file line number Diff line number Diff line change
Expand Up @@ -4,7 +4,7 @@
Benchbase Workload SF=16 (warehouses for TPC-C)
Type: benchbase
Duration: 800s
Code: 1728550061
Code: 1729530658
This includes no queries. Benchbase runs the benchmark
This experiment compares run time and resource consumption of Benchbase queries in different DBMS.
Benchbase data is generated and loaded using several threads.
Expand All @@ -26,14 +26,14 @@ MariaDB-1-1-1024-1 uses docker image mariadb:11.4.2
Cores:64
host:5.15.0-116-generic
node:cl-worker11
disk:251659872
datadisk:1942868
disk:252259184
datadisk:1935700
requests_cpu:4
requests_memory:16Gi

### Execution
experiment_run terminals target pod_count time Throughput (requests/second) Latency Distribution.95th Percentile Latency (microseconds) Latency Distribution.Average Latency (microseconds)
MariaDB-1-1-1024-1 1 16 8192 1 300.0 156.6 31335.0 102151.0
MariaDB-1-1-1024-1 1 16 8192 1 300.0 157.86 30764.0 79907.0

Warehouses: 16

Expand All @@ -47,7 +47,7 @@ DBMS MariaDB-1-1-1024 - Pods [[1]]

### Loading
time_load terminals pods Imported warehouses [1/h]
MariaDB-1-1-1024-1 316.0 1.0 1.0 182.278481
MariaDB-1-1-1024-1 285.0 1.0 1.0 202.105263

### Tests
TEST passed: Throughput (requests/second) contains no 0 or NaN
Expand Down
182 changes: 182 additions & 0 deletions logs_tests/test_benchbase_testcase_mariadb_2.log
Original file line number Diff line number Diff line change
@@ -0,0 +1,182 @@
Data Directory : is running
Result Directory : is running
Dashboard : is running
Message Queue : is running
Experiment : has code 1729531508
Experiment : starts at 2024-10-21 19:25:08.143360 (4267448.267341745)
Experiment : This experiment compares run time and resource consumption of Benchbase queries in different DBMS.
Benchbase data is generated and loaded using several threads.
Benchmark is 'tpcc'. Scaling factor (e.g., number of warehouses) is 16. Benchmarking runs for 1 minutes. Target is based on multiples of '1024'. Factors for benchmarking are [8].
Benchmark is limited to DBMS ['MariaDB'].
Import is handled by 1 processes (pods).
Loading is fixed to cl-worker19.
Benchmarking is fixed to cl-worker19.
SUT is fixed to cl-worker11.
Database is persisted to disk of type shared and size 50Gi.
Loading is tested with [1] threads, split into [1] pods.
Benchmarking is tested with [16] threads, split into [1] pods.
Benchmarking is run as [1] times the number of benchmarking pods.
Experiment is run 2 times.
Cluster monitoring : is running
done
MariaDB-1-1-1024 : will start now
- waiting 30s - : done
MariaDB-1-1-1024 : is not loaded yet
MariaDB-1-1-1024 : will start loading but not before 2024-10-21 17:27:08 (that is in 60 secs)
- waiting 30s - : done
MariaDB-1-1-1024 : is not loaded yet
MariaDB-1-1-1024 : will start loading but not before 2024-10-21 17:27:08
done
MariaDB-1-1-1024 : is not loaded yet
MariaDB-1-1-1024 : start asynch loading scripts of type loaded
MariaDB-1-1-1024 : is loading
- waiting 30s - : done
MariaDB-1-1-1024 : is loading
- waiting 30s - : done
MariaDB-1-1-1024 : is loading
- waiting 30s - : done
MariaDB-1-1-1024 : is loading
- waiting 30s - : done
MariaDB-1-1-1024 : is loading
- waiting 30s - : done
MariaDB-1-1-1024 : is loading
- waiting 30s - : done
MariaDB-1-1-1024 : is loading
- waiting 30s - : done
MariaDB-1-1-1024 : is loading
- waiting 30s - : done
MariaDB-1-1-1024 : is loading
- waiting 30s - : done
MariaDB-1-1-1024 : is loading
- waiting 30s - : done
MariaDB-1-1-1024 : is loading
done
MariaDB-1-1-1024 : showing loader times
MariaDB-1-1-1024 : loader timespan (first to last [s]) = 325
MariaDB-1-1-1024 : benchmarks done 0 of 2. This will be client 1
MariaDB-1-1-1024 : we will change parameters of benchmark as {'PARALLEL': '1', 'SF': '16', 'BENCHBASE_BENCH': 'tpcc', 'BENCHBASE_PROFILE': 'mariadb', 'BEXHOMA_DATABASE': 'benchbase', 'BENCHBASE_TARGET': 8192, 'BENCHBASE_TERMINALS': 16, 'BENCHBASE_TIME': 60, 'BENCHBASE_ISOLATION': 'TRANSACTION_READ_COMMITTED'}
MariaDB-1-1-1024-1-1 : start benchmarking
MariaDB-1-1-1024 : benchmarking results in folder /home/perdelt/benchmarks/1729531508
- waiting 10s - : done
MariaDB-1-1-1024 : benchmarking is waiting for job bexhoma-benchmarker-mariadb-1-1-1024-1729531508-1-1-x4w8g: found
- waiting 30s - : done
MariaDB-1-1-1024 : has running benchmarks
done
MariaDB-1-1-1024 : has running benchmarks
MariaDB-1-1-1024-1-1 : showing benchmarker times
MariaDB-1-1-1024-1-1 : benchmarker timespan (start to end single container [s]) = 71
MariaDB-1-1-1024-1-1 : benchmarker times (start/end per pod and container) = [(1729531965, 1729532031)]
MariaDB-1-1-1024-1-1 : found and updated times {'benchmarker': [(1729531965, 1729532031)]}
done
MariaDB-1-1-1024 : can be stopped
MariaDB-1-1-1024 : still being removed
done
MariaDB-1-1-1024 : starts again
MariaDB-1-1-1024 : storage exists bexhoma-storage-mariadb-benchbase-16
MariaDB-1-1-1024 : loading is set to finished
- waiting 10s - : done
- waiting 30s - : done
MariaDB-1-1-1024 : will start benchmarking but not before 2024-10-21 17:36:36 (that is in 60 secs)
- waiting 30s - : done
MariaDB-1-1-1024 : will start benchmarking but not before 2024-10-21 17:36:36
done
MariaDB-1-1-1024 : benchmarks done 1 of 2. This will be client 1
MariaDB-1-1-1024 : we will change parameters of benchmark as {'PARALLEL': '1', 'SF': '16', 'BENCHBASE_BENCH': 'tpcc', 'BENCHBASE_PROFILE': 'mariadb', 'BEXHOMA_DATABASE': 'benchbase', 'BENCHBASE_TARGET': 8192, 'BENCHBASE_TERMINALS': 16, 'BENCHBASE_TIME': 60, 'BENCHBASE_ISOLATION': 'TRANSACTION_READ_COMMITTED'}
MariaDB-1-1-1024-2-1 : start benchmarking
MariaDB-1-1-1024 : benchmarking results in folder /home/perdelt/benchmarks/1729531508
- waiting 10s - : done
MariaDB-1-1-1024 : benchmarking is waiting for job bexhoma-benchmarker-mariadb-1-1-1024-1729531508-2-1-9q8jk: found
- waiting 30s - : done
MariaDB-1-1-1024 : has running benchmarks
done
MariaDB-1-1-1024 : has running benchmarks
MariaDB-1-1-1024-2-1 : showing benchmarker times
MariaDB-1-1-1024-2-1 : benchmarker timespan (start to end single container [s]) = 71
MariaDB-1-1-1024-2-1 : benchmarker times (start/end per pod and container) = [(1729532201, 1729532263)]
MariaDB-1-1-1024-2-1 : found and updated times {'benchmarker': [(1729532201, 1729532263)]}
done
MariaDB-1-1-1024 : can be stopped
Experiment ends at 2024-10-21 19:38:21.570796 (4268241.69477506): 793.4274333147332s total
Experiment : downloading partial results
Experiment : uploading full results
Experiment : uploading workload file
Benchmarking connection ... Latency Distribution.Average Latency (microseconds)
connection_pod ...
MariaDB-1-1-1024-2-1-1 MariaDB-1-1-1024-2-1 ... 29917
MariaDB-1-1-1024-1-1-1 MariaDB-1-1-1024-1-1 ... 40909

[2 rows x 36 columns]
Workflow {'MariaDB-1-1-1024': [[1], [1]]}
Result workflow complete

## Show Summary

### Workload
Benchbase Workload SF=16 (warehouses for TPC-C)
Type: benchbase
Duration: 794s
Code: 1729531508
This includes no queries. Benchbase runs the benchmark
This experiment compares run time and resource consumption of Benchbase queries in different DBMS.
Benchbase data is generated and loaded using several threads.
Benchmark is 'tpcc'. Scaling factor (e.g., number of warehouses) is 16. Benchmarking runs for 1 minutes. Target is based on multiples of '1024'. Factors for benchmarking are [8].
Benchmark is limited to DBMS ['MariaDB'].
Import is handled by 1 processes (pods).
Loading is fixed to cl-worker19.
Benchmarking is fixed to cl-worker19.
SUT is fixed to cl-worker11.
Database is persisted to disk of type shared and size 50Gi.
Loading is tested with [1] threads, split into [1] pods.
Benchmarking is tested with [16] threads, split into [1] pods.
Benchmarking is run as [1] times the number of benchmarking pods.
Experiment is run 2 times.

### Connections
MariaDB-1-1-1024-1-1 uses docker image mariadb:11.4.2
RAM:541008605184
CPU:AMD Opteron(tm) Processor 6378
Cores:64
host:5.15.0-116-generic
node:cl-worker11
disk:250323480
datadisk:1940452
volume_size:50G
volume_used:1.9G
requests_cpu:4
requests_memory:16Gi
MariaDB-1-1-1024-2-1 uses docker image mariadb:11.4.2
RAM:541008605184
CPU:AMD Opteron(tm) Processor 6378
Cores:64
host:5.15.0-116-generic
node:cl-worker11
disk:250323476
datadisk:1980233
volume_size:50G
volume_used:1.9G
requests_cpu:4
requests_memory:16Gi

### Execution
experiment_run terminals target pod_count time Throughput (requests/second) Latency Distribution.95th Percentile Latency (microseconds) Latency Distribution.Average Latency (microseconds)
MariaDB-1-1-1024-1-1 1 16 8192 1 60.0 376.43 35970.0 40909.0
MariaDB-1-1-1024-2-1 2 16 8192 1 60.0 534.45 34873.0 29917.0

Warehouses: 16

### Workflow

#### Actual
DBMS MariaDB-1-1-1024 - Pods [[1], [1]]

#### Planned
DBMS MariaDB-1-1-1024 - Pods [[1], [1]]

### Loading
time_load terminals pods Imported warehouses [1/h]
MariaDB-1-1-1024-1-1 325.0 1.0 1.0 177.230769
MariaDB-1-1-1024-2-1 325.0 1.0 1.0 177.230769

### Tests
TEST passed: Throughput (requests/second) contains no 0 or NaN
TEST passed: Workflow as planned
Loading

0 comments on commit 6fd6ca3

Please sign in to comment.