Details
-
Type:
Bug
-
Status: Closed
-
Priority:
Critical
-
Resolution: Fixed
-
Affects Version/s: None
-
Fix Version/s: None
-
Component/s: None
-
Labels:
Description
This is MariaDB-5.5.23 (GA, as released). DBT3 data at SF30 loaded into InnoDB tables. Default statistics (fluctuating). Running EXPLAIN for Q13 10 times, restarting the server in between. Yields 10 times the same (bad) plan. Earlier tests resulted in exactly this plan 100 times in 100 tries.
How to reproduce:
login to facebook-maria1
cd ~/benchmark/dbt3/mariadb-tools/dbt3_benchmark
~/benchmark/dbt3/mariadb-tools/dbt3_benchmark> time perl launcher.pl --project-home=/home/mariadb/benchmark/dbt3/ --results-output-dir=/home/mariadb/benchmark/dbt3/results/innodb_s30_test --datadir=/home/mariadb/benchmark/dbt3/db_data/ --test=/home/mariadb/benchmark/dbt3/mariadb-tools/dbt3_benchmark/tests/innodb_test_mariadb_5_5_Q13_for_igor.conf --queries-home=/home/mariadb/benchmark/dbt3/gen_query/ --scale-factor=30
(this is one single line)
after that there will be a new results directory "innodb_s30_test_<timestamp>" below ~/benchmark/dbt3/results.
Attached is the tarred results directory from the run I've done just now.
Gliffy Diagrams
Attachments
Activity
- All
- Comments
- Work Log
- History
- Activity
- Transitions
result from DBT3 run
LPexportBug999251_innodb_s30_test_2012-05-14_210225.tgz