hcq/MaxScale
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fix replication in parallel

Browse Source 
If the replication is broken between the nodes, it is now fixed in
parallel on all nodes instead of doing it one server at a time.

This reduces the time from about 120 seconds to 13 seconds. The time was
measured by running the check_backend test first with all backends broken
and then with the fixed backends subtracting time of the latter from the
former.
This commit is contained in:
Markus Mäkelä
2018-11-02 15:59:02 +02:00
parent 04e4f17618
commit ccec2a387a
2 changed files with 122 additions and 154 deletions
Download Patch File Download Diff File Expand all files Collapse all files

20
maxscale-system-test/mariadb_nodes.h
View File

@ -223,6 +223,9 @@ public:
*/
virtual int start_replication();
// Create the default users used by all tests
void create_users(int node);
/**
* @brif BlockNode setup firewall on a backend node to block MariaDB port
* @param node Index of node to block
@ -403,16 +406,20 @@ public:
void add_server_setting(int node, const char* setting);
/**
* Restore the original configuration for this server
* Get the configuration file name for a particular node
*
* @param node Node to restore
* @param node Node number for which the configuration is requested
*
* @return The name of the configuration file
*/
void reset_server_settings(int node);
virtual std::string get_config_name(int node);
/**
* Restore the original configuration for all servers
*/
void reset_server_settings();
// Same but for an individual server
void reset_server_settings(int node);
/**
* @brief revert_nodes_snapshot Execute MDBCI snapshot revert command for all nodes
@ -477,12 +484,7 @@ public:
return check_galera();
}
// int prepare_galera_server(int i);
// virtual int prepare_server(int i)
// {
// return prepare_galera_server(i);
// }
std::string get_config_name(int node) override;
};
#endif // MARIADB_NODES_H