Details
-
Bug
-
Status: Closed (View Workflow)
-
Major
-
Resolution: Fixed
-
3.4
-
None
Description
|
https://buildbot.mariadb.org/#/builders/369/builds/26206/steps/7/logs/stdio |
unit.conc_connection w15 [ pass ] 2
|
MTR's internal check of the test case 'unit.conc_connection' failed.
|
This means that the test case does not preserve the state that existed
|
before the test case was executed. Most likely the test case did not
|
do a proper clean-up. It could also be caused by the previous test run
|
by this thread, if the server wasn't restarted.
|
This is the diff of the states of the servers before and after the
|
test case was executed:
|
mysqltest: Logging to '/home/buildbot/amd64-ubuntu-2004-debug/build/mysql-test/var/15/tmp/check-mysqld_1.log'.
|
mysqltest: Results saved in '/home/buildbot/amd64-ubuntu-2004-debug/build/mysql-test/var/15/tmp/check-mysqld_1.result'.
|
mysqltest: Connecting to server localhost:19660 (socket /home/buildbot/amd64-ubuntu-2004-debug/build/mysql-test/var/tmp/15/mysqld.1.sock) as 'root', connection 'default', attempt 0 ...
|
mysqltest: ... Connected.
|
mysqltest: Start processing test commands from './include/check-testcase.inc' ...
|
mysqltest: ... Done processing test commands.
|
--- /home/buildbot/amd64-ubuntu-2004-debug/build/mysql-test/var/15/tmp/check-mysqld_1.result 2024-11-21 07:27:00.936421798 +0000
|
+++ /home/buildbot/amd64-ubuntu-2004-debug/build/mysql-test/var/15/tmp/check-mysqld_1.reject 2024-11-21 07:27:49.359574547 +0000
|
@@ -900,7 +900,7 @@
|
mysql.help_category 920235239
|
mysql.help_keyword 1315967545
|
mysql.help_relation 3123006815
|
-mysql.plugin 0
|
+mysql.plugin 768818758
|
mysql.procs_priv 0
|
mysql.roles_mapping 0
|
mysql.tables_priv 454043219
|
@@ -909,7 +909,7 @@
|
mysql.time_zone_name 2502923589
|
mysql.time_zone_transition 3719776009
|
mysql.time_zone_transition_type 267780600
|
-mysql.global_priv 3162002917
|
+mysql.global_priv 2548605602
|
PLUGIN_NAME PLUGIN_VERSION PLUGIN_STATUS PLUGIN_TYPE PLUGIN_TYPE_VERSION PLUGIN_LIBRARY PLUGIN_LIBRARY_VERSION PLUGIN_AUTHOR PLUGIN_DESCRIPTION PLUGIN_LICENSE LOAD_OPTION PLUGIN_MATURITY PLUGIN_AUTH_VERSION
|
binlog 2.0 ACTIVE DAEMON 110701.0 NULL NULL MySQL AB This is a plugin to represent the binlog in a transaction GPL FORCE Stable 2.0
|
mysql_native_password 1.0 ACTIVE AUTHENTICATION 2.2 NULL NULL R.J.Silk, Sergei Golubchik Native MySQL authentication GPL FORCE Stable 1.0
|
@@ -985,5 +985,6 @@
|
THREAD_POOL_STATS 1.0 DISABLED INFORMATION SCHEMA 110701.0 NULL NULL Vladislav Vaintroub Provides performance counter information for threadpool. GPL OFF Stable 1.0
|
THREAD_POOL_WAITS 1.0 DISABLED INFORMATION SCHEMA 110701.0 NULL NULL Vladislav Vaintroub Provides wait counters for threadpool. GPL OFF Stable 1.0
|
partition 1.0 DISABLED STORAGE ENGINE 110701.0 NULL NULL Mikael Ronstrom, MySQL AB Partition Storage Engine Helper GPL OFF Stable 1.0
|
+parsec 1.0 ACTIVE AUTHENTICATION 2.2 auth_parsec.so 1.15 Nikita Maliavin Password Authentication using Response Signed with Elliptic Curve GPL ON Beta 1.0
|
VARIABLE_NAME VARIABLE_VALUE
|
DEBUG_SYNC ON - current signals: ''
|
Result length mismatch
|
not ok
|
Attachments
Activity
This seems to be broken after updating the libmariadb submodule to include the supposed fix.
https://buildbot.mariadb.org/#/builders/710/builds/3762/steps/7/logs/stdio
https://buildbot.mariadb.org/#/builders/710/builds/3847/steps/7/logs/stdio
This is a flawed test case and the fix was insufficient.
Here's the scenario that is running now both in new and old buildbot:
- The test is run as part of unit.conc_connection
- The test case for test_parsec is skipped, at this line (connection.c:2367)
This is because the test can't find the client parsec.so library.
As a principle, I believe we should make two changes to this test case:
Change 1
The test case should either report as FAILED, or SKIP before running any actual SQL queries against the server.
This line (connection.c:2358) should not run before any of the SKIP tests.
Or, this line (connection.c:2367) should be marked as FAIL.
Change 2
It seems that the test looks for client parsec.so library in the default path, which doesn't work if the server is tested from the source directory, instead of from an installed system path.
To see this, change line 2367 to FAIL and you will see this line in the logs:
# parsec plugin not available
|
# error: Plugin parsec could not be loaded: /usr/local/mysql/lib/plugin/parsec.so: cannot open shared object file: No such file or directory
|
not ok 3 - test_parsec
|
In old buildbot, the test_parsec is always skipped, leading to check_testcase failure, but the actual unit.conc_connection is not marked as failed. Hence this has gone unnoticed for a long time.
I am not sure how this issue should be fixed, likely the connector library in the unit test must have some paths set correctly.
Thanks georg!