Bug #42657 The backup_engines test suite restarts server too often.

• Submitted: 6 Feb 2009 17:50
• Modified: 17 Aug 2011 13:52
• Reporter: Rafal Somla
• Status: Won't fix
• Category: MySQL Server: Tests
• Severity: S3 (Non-critical)
• Version: 6.0-backup
• OS: Any
• CPU Architecture: Any

[6 Feb 2009 17:50] Rafal Somla

Description:
Tests in the backup_engines test suite are run on server with different default storage engine. This is similar to rpl suite where each test is run in mixed, row and stmt replication mode.

However, in rpl suite first all tests are run in mixed mode, then in row mode and then in stmt mode, which minimizes the number of necessary server restarts. But for backup_engines, each test is run for each different storage engine before mtr proceeds to the next test. This makes the number of server restarts very high.

How to repeat:
Run "./mtr --suite=backup_engines". You will see something like this:

<cut>
backup_engines.backup_ptr_commit_mixed 'myisam' [ skipped ]  "This test does not support Myisam engine"
backup_engines.backup_ptr_commit_mixed 'falcon' [ skipped ]  "This test does not support Falcon engine"
backup_engines.backup_ptr_commit_mixed 'memory' [ skipped ]  "This test does not support memory engine"
backup_engines.backup_ptr_commit_mixed 'innodb' [ pass ]   3045
backup_engines.backup_ptr_commit_row 'myisam' [ skipped ]  "This test does not support Myisam engine"
backup_engines.backup_ptr_commit_row 'falcon' [ skipped ]  "This test does not support Falcon engine"
backup_engines.backup_ptr_commit_row 'memory' [ skipped ]  "This test does not support memory engine"
backup_engines.backup_ptr_commit_row 'innodb' [ pass ]   3076
backup_engines.backup_ptr_commit_stmt 'myisam' [ skipped ]  "This test does not support Myisam engine"
backup_engines.backup_ptr_commit_stmt 'falcon' [ skipped ]  "This test does not support Falcon engine"
backup_engines.backup_ptr_commit_stmt 'memory' [ skipped ]  "This test does not support memory engine"
backup_engines.backup_ptr_commit_stmt 'innodb' [ pass ]   3198
------------------------------------------------------------
The servers were restarted 55 times
Spent 126.357 of 337 seconds executing testcases

All 39 tests were successful.

Note that server was restarted 55 times for 39 tests.

Suggested fix:
Change the order in which tests are run: first all tests for the first storage engine, than all the test for the next storage engine and so on.