Details
-
Bug
-
Status: Closed (View Workflow)
-
Major
-
Resolution: Duplicate
-
10.6.14, 10.11.6, 10.11.7, 11.4.1
-
SLES 15.5 (Kernel version: Linux version 5.14.21-150500.55.36-default) running as VM
Description
When replaying binary logs (in a point-in-time recovery) that contain a "select nextval(<sequence>)" statement in row format the server crashes.
Steps to recreate (using some example values):
1. "create database seq_test;"
2. "use seq_test"
3. "create sequence test_sequence start with 88000000 minvalue 1 maxvalue 9223372036854775806 increment by 1 nocache nocycle ENGINE=InnoDB;"
4. "select setval(test_sequence, 89454596, 0);"
5. "flush binary logs;"
6. create a backup of DB seq_test with "mysqldump [...] > dump.sql"
7. "select nextval(test_sequence);"
8. "flush binary logs;"
9. create SQL file with binary log from step 5-7 with "mysqlbinlog [...] > binlog.sql"
10. "drop database seq_test;"
11. mysql [...] < dump.sql
12. mysql [...] < binlog.sql
13. server crashes here
Stacktraces for versions 11.4.1 & 10.11.7 are attached as well as the my.cnf used.
Attachments
Issue Links
- duplicates
-
MDEV-31779 Server crash in Rows_log_event::update_sequence upon replaying binary log
- Closed