WL#7817: RPL Monitoring: Move status variables to replication P_S tables
Affects: Server-5.7
—
Status: Complete
This work is a requirement for Multi-source replication. To be able to monitor the following variables on each channel, we cannot have them as global variables but variables per channel. Show status like 'Slave_running'; Show status like 'Slave_retried_transactions'; Show status like 'Slave_last_heartbeat'; Show status like 'Slave_received_heartbeats'; show status like 'Slave_heartbeat_period'; This task aims at moving these variables to replication performance schema tables established in WL#3656 so that these can be reported per channel in these performance schema tables.
Functional requirements: ======================= F1: P_S tables must be read-only, available for SQL queries (SELECT ...) as for regular tables. F2: Before CHANGE MASTER statement P_S table should be empty. F3: After CHANGE MASTER statement P_S table should be filled by a data, as earlier. F4: START/STOP SLAVE statements and running slave should update P_S "status" tables. F5: Concurent access to P_S replication tables must not affect to replication (no locks, no errors, etc) Non-functional requirements: =========================== NF1: Adding these variables to P_S tables should not affect server performance. NF2: Disabed P_S for the server shouldn't affect to the behavior of replication.
- In this section we list all the status variables mentioned in HLD and discuss
them one-by-one.
1) SLAVE_RUNNING: The documentation says-
=============
This is ON if this server is a replication slave that is connected to a
replication master, and both the I/O and SQL threads are running;
otherwise, it is OFF.
Decision: This will NOT be added to performance_schema replication tables.
Justification: table performance_schema.replication_connection_status and
table performance_schema.replication_execute_status each
have a field that corresponds to the IO and SQL threads
respectively. The variable Slave_running is same as
(SELECT SERVICE_STATE FROM performance_schema.replication_connection_status AND
SELECT SERVICE_STATE FROM performance_schema.replication_execute_status).
Hence we drop this request and don't add this field to the P_S
tables.
2) SLAVE_RETRIED_TRANSACTIONS: The documentation says-
==========================
If a replication slave SQL thread fails to execute a transaction because of
an InnoDB deadlock or because the transaction's execution time exceeded
InnoDB's innodb_lock_wait_timeout or NDBCLUSTER's
TransactionDeadlockDetectionTimeout or TransactionInactiveTimeout, it
automatically retries slave_transaction_retries times before stopping with
an error.
Decision: This will be added to the
performance_schema.replication_execute_status.
3) SLAVE_LAST_HEARTBEAT: The documentation says-
====================
Shows when the most recent heartbeat signal was received by a replication
slave, as a TIMESTAMP value.
Decision: Added to performance_schema.replication_connection_status
4) SLAVE_RECEIVED_HEARTBEATS: The documentation says:
=========================
This counter increments with each replication heartbeat received by a
replication slave since the last time that the slave was restarted or reset,
or a CHANGE MASTER TO statement was issued.
Decision: Added to performance_schema.replication_connection_status
5) SLAVE_HEARTBEAT_PERIOD: The documentation says:
======================
Shows the replication heartbeat interval (in seconds) on a replication slave.
Decision: Added to performance_schema.replication_connection_configuration.
- Here is how the names of global variables map to the names in the P_S tables:
+----------------------------+-----------------------------+
| Variable_name | P_S name |
+----------------------------+-----------------------------+
| SLAVE_HEARTBEAT_PERIOD | HEARTBEAT_INTERVAL |
+----------------------------+-----------------------------+
| SLAVE_RECEIVED_HEARTBEATS | COUNT_RECEIVED_HEARTBEATS |
+----------------------------+-----------------------------+
| SLAVE_LAST_HEARTBEAT | LAST_HEARTBEAT_TIMESTAMP |
+----------------------------+-----------------------------+
| SLAVE_RETRIED_TRANSACTIONS | COUNT_TRANSACTIONS_RETRIES |
+----------------------------+-----------------------------+
1. System tables script changes
============================
The script mysql_system_tables.sql is updated to include the new fields in the
replication P_S tables.
== modified file 'scripts/mysql_system_tables.sql'
--- scripts/mysql_system_tables.sql revid:jorgen.loland@oracle.com-
20140321120934-464u92uocbh7pcvq
+++ scripts/mysql_system_tables.sql 2014-04-15 09:35:11 +0000
@@ -2127,7 +2127,8 @@
"SSL_CRL_FILE VARCHAR(255) not null,"
"SSL_CRL_PATH VARCHAR(255) not null,"
"CONNECTION_RETRY_INTERVAL INTEGER not null,"
- "CONNECTION_RETRY_COUNT BIGINT unsigned NOT NULL"
+ "CONNECTION_RETRY_COUNT BIGINT unsigned not null,"
+ "HEARTBEAT_INTERVAL FLOAT not null"
") ENGINE=PERFORMANCE_SCHEMA;";
SET @str = IF(@have_pfs = 1, @cmd, 'SET @dummy = 0');
@@ -2143,6 +2144,8 @@
"SOURCE_UUID CHAR(36) collate utf8_bin not null,"
"THREAD_ID BIGINT unsigned,"
"SERVICE_STATE ENUM('ON','OFF','CONNECTING') not null,"
+ "COUNT_RECEIVED_HEARTBEATS BIGINT unsigned,"
+ "LAST_HEARTBEAT_TIMESTAMP TIMESTAMP(0) not null,"
"RECEIVED_TRANSACTION_SET TEXT not null,"
"LAST_ERROR_NUMBER INTEGER not null,"
"LAST_ERROR_MESSAGE VARCHAR(1024) not null,"
2. Add functions to set float data type in P_S
==========================================
P_S doesnt have support for float fields. Add infrastructure for the same.
=== modified file 'storage/perfschema/pfs_engine_table.cc'
--- storage/perfschema/pfs_engine_table.cc revid:jorgen.loland@oracle.com-
20140321120934-464u92uocbh7pcvq
+++ storage/perfschema/pfs_engine_table.cc 2014-04-15 09:33:53 +0000
@@ -590,6 +590,13 @@
f2->store_timestamp(& tm);
}
+void PFS_engine_table::set_field_float(Field *f, float value)
+{
+ DBUG_ASSERT(f->real_type() == MYSQL_TYPE_FLOAT);
+ Field_float *f2= (Field_float*) f;
+ f2->store(value);
+}
+
ulonglong PFS_engine_table::get_field_enum(Field *f)
{
DBUG_ASSERT(f->real_type() == MYSQL_TYPE_ENUM);
=== modified file 'storage/perfschema/pfs_engine_table.h'
--- storage/perfschema/pfs_engine_table.h revid:jorgen.loland@oracle.com-
20140321120934-464u92uocbh7pcvq
+++ storage/perfschema/pfs_engine_table.h 2014-04-15 09:32:52 +0000
@@ -138,6 +138,12 @@
*/
static void set_field_timestamp(Field *f, ulonglong value);
/**
+ Helper, assign a value to a float field.
+ @param f the field to set
+ @param value the value to assign
+ */
+ static void set_field_float(Field *f, float value);
+ /**
Helper, read a value from an enum field.
@param f the field to read
@return the field value
3. Add fields to the tables.
========================
3.1 Add HEARTBEAT_INTERVAL in the table replication_connection_configuration.
=== modified file
'storage/perfschema/table_replication_connection_configuration.cc'
--- storage/perfschema/table_replication_connection_configuration.cc
revid:jorgen.loland@oracle.com-20140321120934-464u92uocbh7pcvq
+++ storage/perfschema/table_replication_connection_configuration.cc 2014-04-
15 09:30:49 +0000
@@ -116,12 +116,17 @@
{C_STRING_WITH_LEN("CONNECTION_RETRY_COUNT")},
{C_STRING_WITH_LEN("bigint")},
{NULL, 0}
- }
+ },
+ {
+ {C_STRING_WITH_LEN("HEARTBEAT_INTERVAL")},
+ {C_STRING_WITH_LEN("float")},
+ {NULL, 0}
+ }
};
TABLE_FIELD_DEF
table_replication_connection_configuration::m_field_def=
-{ 16, field_types };
+{ 17, field_types };
PFS_engine_table_share
table_replication_connection_configuration::m_share=
@@ -279,6 +284,8 @@
m_row.connection_retry_count= (ulong) active_mi->retry_count;
+ m_row.heartbeat_interval= active_mi->heartbeat_period;
+
mysql_mutex_unlock(&active_mi->rli->data_lock);
mysql_mutex_unlock(&active_mi->data_lock);
mysql_mutex_unlock(&LOCK_active_mi);
@@ -358,6 +365,9 @@
case 15: /** connect_retry_count */
set_field_ulonglong(f, m_row.connection_retry_count);
break;
+ case 16:/** number of seconds after which heartbeat will be sent */
+ set_field_float(f, m_row.heartbeat_interval);
+ break;
default:
DBUG_ASSERT(false);
}
=== modified file
'storage/perfschema/table_replication_connection_configuration.h'
--- storage/perfschema/table_replication_connection_configuration.h
revid:jorgen.loland@oracle.com-20140321120934-464u92uocbh7pcvq
+++ storage/perfschema/table_replication_connection_configuration.h 2014-04-
15 09:39:25 +0000
@@ -79,6 +79,7 @@
uint ssl_crl_path_length;
uint connection_retry_interval;
ulong connection_retry_count;
+ float heartbeat_interval;
};
3.2 Add COUNT_RECEIVED_HEARTBEATS and LAST_HEARTBEAT_TIMESTAMP
and in the table replication_connection_status.
=== modified file 'storage/perfschema/table_replication_connection_status.cc'
--- storage/perfschema/table_replication_connection_status.cc
revid:jorgen.loland@oracle.com-20140321120934-464u92uocbh7pcvq
+++ storage/perfschema/table_replication_connection_status.cc 2014-04-15
10:13:08 +0000
@@ -54,6 +54,16 @@
{NULL, 0}
},
{
+ {C_STRING_WITH_LEN("COUNT_RECEIVED_HEARTBEATS")},
+ {C_STRING_WITH_LEN("bigint(20)")},
+ {NULL, 0}
+ },
+ {
+ {C_STRING_WITH_LEN("LAST_HEARTBEAT_TIMESTAMP")},
+ {C_STRING_WITH_LEN("timestamp")},
+ {NULL, 0}
+ },
+ {
{C_STRING_WITH_LEN("RECEIVED_TRANSACTION_SET")},
{C_STRING_WITH_LEN("text")},
{NULL, 0}
@@ -77,7 +87,7 @@
TABLE_FIELD_DEF
table_replication_connection_status::m_field_def=
-{ 7, field_types };
+{ 9, field_types };
PFS_engine_table_share
table_replication_connection_status::m_share=
@@ -217,6 +227,13 @@
m_row.service_state= PS_RPL_CONNECT_SERVICE_STATE_NO;
}
+ m_row.count_received_heartbeats= active_mi->received_heartbeats;
+ /*
+ Time in Milliseconds since epoch. active_mi->last_heartbeat contains
+ number of seconds so we multiply by 1000000.
+ */
+ m_row.last_heartbeat_timestamp= (ulonglong)active_mi->last_heartbeat*1000000;
+
mysql_mutex_lock(&active_mi->err_lock);
mysql_mutex_lock(&active_mi->rli->err_lock);
@@ -249,7 +266,9 @@
memcpy(m_row.last_error_message, temp_store,
m_row.last_error_message_length);
- /** time in millisecond since epoch */
+ /*
+ Time in millisecond since epoch. active_mi->last_error().skr contains
+ number of seconds so we multiply by 1000000. */
m_row.last_error_timestamp= (ulonglong)active_mi->last_error().skr*1000000;
}
@@ -294,18 +313,24 @@
case 2: /** service_state */
set_field_enum(f, m_row.service_state);
break;
- case 3: /** received_transaction_set */
+ case 3: /** number of heartbeat events received **/
+ set_field_ulonglong(f, m_row.count_received_heartbeats);
+ break;
+ case 4: /** time of receipt of last heartbeat event **/
+ set_field_timestamp(f, m_row.last_heartbeat_timestamp);
+ break;
+ case 5: /** received_transaction_set */
set_field_longtext_utf8(f, m_row.received_transaction_set,
m_row.received_transaction_set_length);
break;
- case 4: /*last_error_number*/
+ case 6: /*last_error_number*/
set_field_ulong(f, m_row.last_error_number);
break;
- case 5: /*last_error_message*/
+ case 7: /*last_error_message*/
set_field_varchar_utf8(f, m_row.last_error_message,
m_row.last_error_message_length);
break;
- case 6: /*last_error_timestamp*/
+ case 8: /*last_error_timestamp*/
set_field_timestamp(f, m_row.last_error_timestamp);
break;
default:
=== modified file 'storage/perfschema/table_replication_connection_status.h'
--- storage/perfschema/table_replication_connection_status.h
revid:jorgen.loland@oracle.com-20140321120934-464u92uocbh7pcvq
+++ storage/perfschema/table_replication_connection_status.h 2014-04-13
18:38:53 +0000
@@ -57,6 +57,8 @@
ulonglong thread_id;
bool thread_id_is_null;
enum_rpl_connect_status_service_state service_state;
+ ulonglong count_received_heartbeats;
+ ulonglong last_heartbeat_timestamp;
char* received_transaction_set;
int received_transaction_set_length;
uint last_error_number;
Copyright (c) 2000, 2024, Oracle Corporation and/or its affiliates. All rights reserved.