Description:
When the partition handling thread is started there is no PSI key associated to it.
if (mysql_thread_create(0,
&partition_trx_handler_pthd,
get_connection_attrib(),
launch_handler_thread,
(void*)this))
If the user does a select in the performance schema table for threads, and this thread is running, it wont show up.
There is no issues besides the visibility one.
How to repeat:
Cause a partition situation.
And check the performance_schema.threads table.
You can use the following diff
diff --git a/rapid/plugin/group_replication/tests/mtr/t/gr_network_partition_handling_after_restore.test b/rapid/plugin/group_replication/tests/mtr/t/gr_network_partition_handling_after_restore.test
index ebb5543..648068f 100644
--- a/rapid/plugin/group_replication/tests/mtr/t/gr_network_partition_handling_after_restore.test
+++ b/rapid/plugin/group_replication/tests/mtr/t/gr_network_partition_handling_after_restore.test
@@ -99,6 +99,10 @@ CREATE TABLE t1 (c1 INT NOT NULL PRIMARY KEY) ENGINE=InnoDB;
--let $rpl_connection_name= server_1
--source include/rpl_connection.inc
+--sleep 2
+
+SELECT * FROM performance_schema.threads;
+
--disable_query_log
--eval SET GLOBAL group_replication_force_members= "$local_address_server1"
--enable_query_log
And then run
./mtr --retry=0 --timer --vardir=/dev/shm/thread_status --suite=group_replication group_replication.gr_network_partition_handling_after_restore --big-test
Suggested fix:
Create a PSI key for the thread.
Description: When the partition handling thread is started there is no PSI key associated to it. if (mysql_thread_create(0, &partition_trx_handler_pthd, get_connection_attrib(), launch_handler_thread, (void*)this)) If the user does a select in the performance schema table for threads, and this thread is running, it wont show up. There is no issues besides the visibility one. How to repeat: Cause a partition situation. And check the performance_schema.threads table. You can use the following diff diff --git a/rapid/plugin/group_replication/tests/mtr/t/gr_network_partition_handling_after_restore.test b/rapid/plugin/group_replication/tests/mtr/t/gr_network_partition_handling_after_restore.test index ebb5543..648068f 100644 --- a/rapid/plugin/group_replication/tests/mtr/t/gr_network_partition_handling_after_restore.test +++ b/rapid/plugin/group_replication/tests/mtr/t/gr_network_partition_handling_after_restore.test @@ -99,6 +99,10 @@ CREATE TABLE t1 (c1 INT NOT NULL PRIMARY KEY) ENGINE=InnoDB; --let $rpl_connection_name= server_1 --source include/rpl_connection.inc +--sleep 2 + +SELECT * FROM performance_schema.threads; + --disable_query_log --eval SET GLOBAL group_replication_force_members= "$local_address_server1" --enable_query_log And then run ./mtr --retry=0 --timer --vardir=/dev/shm/thread_status --suite=group_replication group_replication.gr_network_partition_handling_after_restore --big-test Suggested fix: Create a PSI key for the thread.