| Bug #51648 | DBUG_SYNC_POINT is not defined on all platforms and mtr can't pre-check that | ||
|---|---|---|---|
| Submitted: | 2 Mar 2010 16:47 | Modified: | 18 Jun 2010 1:10 |
| Reporter: | Andrei Elkin | Email Updates: | |
| Status: | Closed | Impact on me: | |
| Category: | MySQL Server | Severity: | S3 (Non-critical) |
| Version: | 5.1+ | OS: | Any |
| Assigned to: | Andrei Elkin | CPU Architecture: | Any |
| Tags: | synchronization | ||
[3 Mar 2010 16:51]
Bugs System
A patch for this bug has been committed. After review, it may be pushed to the relevant source trees for release in the next version. You can access the patch from: http://lists.mysql.com/commits/102224 3369 Andrei Elkin 2010-03-03 Bug #51648 DBUG_SYNC_POINT is not defined on all platforms and mtr cant pre-check that a prototype of fixes
[17 Mar 2010 13:53]
Bugs System
A patch for this bug has been committed. After review, it may be pushed to the relevant source trees for release in the next version. You can access the patch from: http://lists.mysql.com/commits/103600 3400 Andrei Elkin 2010-03-17 Bug #51648 DBUG_SYNC_POINT is not defined on all platforms and mtr cant pre-check that DBUG_SYNC_POINT has at least one strong limitation that it's not defined on all platforms. It has issues cooperating with @@debug. All in all its functionality is superseded by DEBUG_SYNC facility and there is no reason to maintain the old also less flexible one. Fixed with adding debug_sync_set_action() function as a facility to set up a sync-action in the server sources code and re-writing existing simulations (found 3) to use it. Couple of tests required changes as well. The patch can be served as a template of setting sync-points in slave threads where the standard DEBUG_SYNC does not suffice to read goals. @ mysql-test/extra/rpl_tests/rpl_get_master_version_and_clock.test rewriting the test from GET_LOCK()-based to DEBUG_SYNC-based. @ mysql-test/suite/rpl/r/rpl_get_master_version_and_clock.result results are changed. @ mysql-test/suite/rpl/t/rpl_get_master_version_and_clock.test rewriting the test from GET_LOCK()-based to DEBUG_SYNC-based; limitting the test to run only with MIXED binlog-format as the test last some 10 secs sentitively contributing to the total of tests run. @ mysql-test/suite/rpl/t/rpl_show_slave_running.test rewriting the test from GET_LOCK()-based to DEBUG_SYNC-based. @ sql/debug_sync.cc adding debug_sync_set_action() function as a facility to set up a sync-action in the server sources code. @ sql/debug_sync.h externalizing debug_sync_set_action(). @ sql/item_func.cc purging sources from DBUG_SYNC_POINT. @ sql/mysql_priv.h purging sources from DBUG_SYNC_POINT. @ sql/slave.cc rewriting failure simulations to base on DEBUG_SYNC rather than GET_LOCK()-based DBUG_SYNC_POINT. @ sql/sql_repl.cc removing an orphan failure simulation line because no couterpart in tests existing.
[18 Mar 2010 16:31]
Ingo Strüwing
Approved with small comments. Please see email.
[18 Mar 2010 18:28]
Bugs System
A patch for this bug has been committed. After review, it may be pushed to the relevant source trees for release in the next version. You can access the patch from: http://lists.mysql.com/commits/103734 3400 Andrei Elkin 2010-03-18 Bug #51648 DBUG_SYNC_POINT is not defined on all platforms and mtr cant pre-check that DBUG_SYNC_POINT has at least one strong limitation that it's not defined on all platforms. It has issues cooperating with @@debug. All in all its functionality is superseded by DEBUG_SYNC facility and there is no reason to maintain the old also less flexible one. Fixed with adding debug_sync_set_action() function as a facility to set up a sync-action in the server sources code and re-writing existing simulations (found 3) to use it. Couple of tests required changes as well. The patch provides a template of setting sync-points in replication threads where the standard DEBUG_SYNC does not suffice to reach goals. @ mysql-test/extra/rpl_tests/rpl_get_master_version_and_clock.test rewriting the test from GET_LOCK()-based to DEBUG_SYNC-based. @ mysql-test/suite/rpl/r/rpl_get_master_version_and_clock.result results are changed. @ mysql-test/suite/rpl/t/rpl_get_master_version_and_clock.test rewriting the test from GET_LOCK()-based to DEBUG_SYNC-based; limiting the test to run only with MIXED binlog-format as the test last some 10 secs sensitively contributing to the total of tests run. @ mysql-test/suite/rpl/t/rpl_show_slave_running.test rewriting the test from GET_LOCK()-based to DEBUG_SYNC-based. @ sql/debug_sync.cc adding debug_sync_set_action() function as a facility to set up a sync-action in the server sources code. @ sql/debug_sync.h externalizing debug_sync_set_action(). @ sql/item_func.cc purging sources from DBUG_SYNC_POINT. @ sql/mysql_priv.h purging sources from DBUG_SYNC_POINT. @ sql/slave.cc rewriting failure simulations to base on DEBUG_SYNC rather than GET_LOCK()-based DBUG_SYNC_POINT. @ sql/sql_repl.cc removing an orphan failure simulation line because no counterpart in tests existing.
[19 Mar 2010 9:07]
Bugs System
A patch for this bug has been committed. After review, it may be pushed to the relevant source trees for release in the next version. You can access the patch from: http://lists.mysql.com/commits/103779 3400 Andrei Elkin 2010-03-19 Bug #51648 DBUG_SYNC_POINT is not defined on all platforms and mtr cant pre-check that DBUG_SYNC_POINT has at least one strong limitation that it's not defined on all platforms. It has issues cooperating with @@debug. All in all its functionality is superseded by DEBUG_SYNC facility and there is no reason to maintain the old less flexible one. Fixed with adding debug_sync_set_action() function as a facility to set up a sync-action in the server sources code and re-writing existing simulations (found 3) to use it. Couple of tests have been reworked as well. The patch offers a pattern for setting sync-points in replication threads where the standard DEBUG_SYNC does not suffice to reach goals. @ mysql-test/extra/rpl_tests/rpl_get_master_version_and_clock.test rewriting the test from GET_LOCK()-based to DEBUG_SYNC-based; a pattern of usage DEBUG_SYNC for replication testing is provided. @ mysql-test/suite/rpl/r/rpl_get_master_version_and_clock.result results are changed. @ mysql-test/suite/rpl/t/rpl_get_master_version_and_clock.test rewriting the test from GET_LOCK()-based to DEBUG_SYNC-based; limiting the test to run only with MIXED binlog-format as the test last some 10 secs sensitively contributing to the total of tests run. @ mysql-test/suite/rpl/t/rpl_show_slave_running.test rewriting the test from GET_LOCK()-based to DEBUG_SYNC-based. @ sql/debug_sync.cc adding debug_sync_set_action() function as a facility to set up a sync-action in the server sources code. @ sql/debug_sync.h externalizing debug_sync_set_action(). @ sql/item_func.cc purging sources from DBUG_SYNC_POINT. @ sql/mysql_priv.h purging sources from DBUG_SYNC_POINT. @ sql/slave.cc rewriting failure simulations to base on DEBUG_SYNC rather than GET_LOCK()-based DBUG_SYNC_POINT. @ sql/sql_repl.cc removing an orphan failure simulation line because no counterpart in tests existing.
[22 Mar 2010 5:50]
Bugs System
A patch for this bug has been committed. After review, it may be pushed to the relevant source trees for release in the next version. You can access the patch from: http://lists.mysql.com/commits/103922 3413 Andrei Elkin 2010-03-22 bug#51648 comments to rpl_show_slave_running test are added to ease merging
[22 Mar 2010 5:59]
Andrei Elkin
Pushed to 5.1-bt, pe trees.
[26 Mar 2010 8:24]
Bugs System
Pushed into 5.5.4-m3 (revid:alik@sun.com-20100326080914-2pz8ns984e0spu03) (version source revid:alexey.kopytov@sun.com-20100322132851-8j3m42x4ldi1kca5) (merge vers: 5.5.3-m2) (pib:16)
[26 Mar 2010 8:27]
Bugs System
Pushed into mysql-next-mr (revid:alik@sun.com-20100326081116-m3v4l34yhr43mtsv) (version source revid:alik@sun.com-20100325072612-4sds00ix8ajo1e84) (pib:16)
[26 Mar 2010 8:31]
Bugs System
Pushed into 6.0.14-alpha (revid:alik@sun.com-20100326081944-qja07qklw1p2w7jb) (version source revid:alik@sun.com-20100325073410-4t4i9gu2u1pge7xb) (merge vers: 6.0.14-alpha) (pib:16)
[6 Apr 2010 7:59]
Bugs System
Pushed into 5.1.46 (revid:sergey.glukhov@sun.com-20100405111026-7kz1p8qlzglqgfmu) (version source revid:john.embretsen@sun.com-20100322090656-w4ixy7p67fb3vr29) (merge vers: 5.1.46) (pib:16)
[13 Apr 2010 19:55]
Paul DuBois
Test suite-related changes. No changelog entry needed.
[17 Jun 2010 12:19]
Bugs System
Pushed into 5.1.47-ndb-7.0.16 (revid:martin.skold@mysql.com-20100617114014-bva0dy24yyd67697) (version source revid:vasil.dimov@oracle.com-20100331130613-8ja7n0vh36a80457) (merge vers: 5.1.46) (pib:16)
[17 Jun 2010 13:07]
Bugs System
Pushed into 5.1.47-ndb-6.2.19 (revid:martin.skold@mysql.com-20100617115448-idrbic6gbki37h1c) (version source revid:martin.skold@mysql.com-20100609211156-tsac5qhw951miwtt) (merge vers: 5.1.46-ndb-6.2.19) (pib:16)
[17 Jun 2010 13:47]
Bugs System
Pushed into 5.1.47-ndb-6.3.35 (revid:martin.skold@mysql.com-20100617114611-61aqbb52j752y116) (version source revid:vasil.dimov@oracle.com-20100331130613-8ja7n0vh36a80457) (merge vers: 5.1.46) (pib:16)
Description: DBUG_SYNC_POINT has at least one strong limitation that it's not defined on all platforms. It's declaration depends on EXTRA_DEBUG which makes the macro expanded to noop on Solaris builts in particular. A worse fact is that tests can not verify if a platform has the macro built in. That affects execution of tests exploiting the feature such as rpl_show_slave_running and may do that to rpl_get_master_version_and_clock. rpl.rpl_show_slave_running 'row' [ fail ] Test ended at 2010-03-02 16:31:36 @@ -20,7 +20,7 @@ SHOW STATUS LIKE 'Slave_running'; Variable_name Value Slave_running OFF -Slave_IO_Running= No +Slave_IO_Running= Yes The source of non-determinism of the test is undetectable by mtr null-expansion of DBUG_SYNC_POINT(). How to repeat: to execute ./mtr rpl_show_slave_running on Solaris or to look at PB2 for rpl_show_slave_running failures. Suggested fix: Refine lines of the server code and referred tests to replace DBUG_SYNC_POINT logics with combination of DBUG_EXECUTE_IF and DBUG_SYNC.