HotDB Server V2.5.1 version is hot Pu database on the basis V2.5.0 computing cluster node milestone on the initial results continue to polish another.
Change fragmentation information online, directly through the existing fragmentation rules to build tables, visualization online upgrade management, user roles, switching and other functions will bring more convenience for the database operation and maintenance work; large-screen home monitor, monitor logical topology, the physical topology monitoring, reporting and other pages connection will bring a different visual experience for database monitoring work; non-deterministic function agents and other features will bring greater consistency to data protection.
-
Management explain the new features
-
Online update service node compute
background note : the history of computing nodes version upgrade service manual intervention required, install and configure database update package replaces manual. Involving standby mode, step more complicated.
Features:
supports single-node management platform, standby computing node, multi-node model node online update service
support is not higher than the upgrade version of the compute node management platform services (compute nodes degradation or change the service mode is not supported)
upgrade process abnormal, automatically roll back and try to ensure support and upgrade services consistent state before
the upgrade process, allows you to view and download the detailed operation log to ensure that the upgrade process can be tracked -
Modify Online slicing scheme
Function:
support online change table fragmentation rules, fragmented field, table type, slice your data node (can not be modified to become sub-table type)
can be directly combined with intelligent slicing scheme recommended by the results of single or batch table slicing scheme change
slicing scheme service change procedure does not block
function:
initiating the need for pre-change: change the program fill, pre-screening program change, the program changes the operation confirmation
will table data consistency standby preflight detects changes proposal, if 24 hours and the available data in accordance with a detection result of standby can choose to skip the detected
changes may be provided to retain or delete the source table after the completion of treatment, deleted after 24 hours by default
long time for the big table change and causes the business performance the impact of the situation, to support replication pause period set to miss the peak of the business.
- Report shows compute node connection information
Function:
Display historical connection through the front-end computing nodes are connected to the service information reporting form
allows you to view the total duration of the connection, the number of operations, the number of connection status three dimensions. Top1 automatically display data in three dimensions and a connection information distribution accounting status in different dimensions
can view up to the computing node connected to the tip all information within one year
- Intelligent Physical Topology
Function:
server-perspective affiliations show cluster components and servers, while demonstrating server resource usage and operational status of each component services
before using the advance connection configuration of the underlying SSH "Server" menu page of information ( multi-computing node in the cluster LVS servers need to manually add)
the individual server monitoring and warning state threshold and "monitoring panel set" threshold synchronization
- smart logical topology
function:
in the original display content and function logic redesigned to show the effect of topography, replace with a new component icon and follow the layout of space layering
support for mouse scroll or zoom buttons freely topology view, large-scale cluster topology display provides better visual effect
new canvas and drag the map interactively imitation style to provide users with a new interactive experience
increases topology information panel, you can view the history of failure, warning, information topology changes. The default Save recorded seven days
- ease Topology Optimization
Function:
Logical topology map page to add the right operation buttons: a data switching node module can be directly manually from the master, dual master storage node type services; module may direct the storage node "Quick Copy connection information "copy formats: MySQL -uUSER -p -PPORT -hHOST -DDATABASE
- supports adding administrative roles user
function:
support when you add user management platform for user role settings, the default is the general role of the user, if the user is set to manage admin has administrative privileges
the user has administrative rights can switch the roles of the state in the top right of the page personal information
Support for user roles when you add management platform user settings, the default is the general role of the user, if set to manage the user has administrative privileges admin
user with administrative privileges can switch roles in the personal status information in the top right of the page
- new backup specification for the user backup function
function:
storage node configuration page to add a backup user configuration item, used exclusively as the data backup
when the backup configuration is not a user, as the user prefers connected HotDB backup backup user
after the backup configuration is not successful backup storage node user there will be warning signs, but does not affect normal use the backup feature of
- Home large screen display optimization
Function Description :
After optimized home big screen strictly relationship control color and state, blue-green normal, red for failure, the orange is warning
Optimization Home typesetting and layout, replacing the more vivid display controls
prominent clusters overview module, more concerned about the overall state of the cluster running
- Cluster deployment-related optimization
Function:
increase LVS option is installed, select the cluster for deploying
open platform IP address restrictions on the management server, namely: management platform can be installed first on a single server, and then install additional components, complex with server resources management platform
- Cluster Group Selection page optimization
Function:
Support through the cluster set of pages, direct services to see all compute nodes in the cluster, the abnormal state configuration repository services
support existing fragmentation rules by associating built directly in the form of SQL tables
non-deterministic functions proxy
optimization watchdog scope of the inspection and to improve logging
dynamically loaded storage node manual / failover Problem
specification HotDB Server-related error code
- Supports the existing fragmentation rules table built directly
support the SHOW command to view the fragment information
SHOW hotdb FUNCTIONS
SHOW hotdb FUNCTION INFOS
SHOW hotdb the RULES
SHOW hotdb DataNodes
support is built directly into the table syntax
level slice table:
the CREATE TABLE [IF the NOT EXISTS] tbl_name Shard BY {functionId | functionName} 'functionId | functionName' the USING the COLUMN 'shardcolumnname' (...
the CREATE TABLE [the IF the NOT EXISTS] tbl_name Shard BY {FunctionType} 'FunctionType' the USING the COLUMN 'shardcolumnname' ON Datanode 'datanodeid' (...
the CREATE TABLE [the IF NOT EXISTS] tbl_name SHARD BY {ruleid | rulename} 'ruleid \ rulename' [on datanode 'datanodes'] (...
global table: CREATE tABLE [IF NOT EXISTS] tbl_name SHARD BY global on datanode 'datanodeid' (...
垂直分片表：CREATE TABLE [IF NOT EXISTS] tbl_name SHARD BY vertical on datanode ‘datanodeid’(…
例一：根据functionname创建，指定datanode：
CREATE TABLE match1_tb shard by functionname ‘test_match1’ using column ‘aname’ on datanode ‘1,2’( id INT UNSIGNED NOT NULL AUTO_INCREMENT PRIMARY KEY,adnid INT default null, aname VARCHAR(32) DEFAULT ‘’, adept VARCHAR(40) , adate datetime default null)ENGINE =INNODB;
例二：根据ruleid创建：
CREATE TABLE match1_rptb1 shard by ruleid ‘1’( id INT UNSIGNED NOT NULL AUTO_INCREMENT PRIMARY KEY,adnid INT default null, aname VARCHAR(32) DEFAULT ‘’, adept VARCHAR(40) , adate datetime default null)ENGINE =INNODB;
- non-deterministic function proxy
function of time known values such as: CURDATE (), CURRENT_TIMESTAMP () and other
times like a function proxy approach:
① When the time class function appears select / insert / update / delete statement, HotDB Server will function Alternatively a specific value into the table
② when the field type is datetime (or timestamp), default is now () or precision now (6), localtime () or precision localtime (6), localtimestamp () or higher accuracy localtimestamp (6), current_timestamp () or precision current_timestamp (6) and the user does not specify a value, a range determined by parameter timestampProxy HotDB Server agent (default is automatic mode, selectable overall pattern / mode only the global table), the replace function to a specific value into the table
| Property | Value |
|---|---|
| Parameter Value | timestampProxy |
| Is visible | Yes |
| Parameter Description | Time proxy mode |
| Defaults | 0 |
| Reload whether the entry into force | Yes |
| The minimum compatible version | The minimum compatible version 2.5.1 |
Watchdog optimization
after opening enableWatchdog parameter WATCHDOG output label information into the log (added when an abnormality occurs: more than one hour long transaction, transaction, more than 24 hours backend storage nodes uncommitted connector disconnected abnormally, the rear end heartbeat failed storage node connected to the storage node handover, handover delay replication)
| Property | Value |
|---|---|
| Parameter Value | Parameter Value enableWatchdog |
| Is visible | Yes |
| Parameter Description | Whether to open Watchdog |
| Defaults | true |
| Reload whether the entry into force | Yes |
| The minimum compatible version | 2.4.5 |
- Watchdog optimization and logging perfect
logging perfect
detect more than one hour uncommitted transaction log records innodb_trx back-end connections, autocommit, lastsql other information
detect more than 24 hours uncommitted transaction logging connection IP, port, username logic library, lastsql, autocommit, innodb_trx back-end connection information and automatically roll back the transaction
when the rear end of the storage node is connected detecting abnormal disconnect, log memory information corresponding SQL
backend storage nodes detect a heartbeat connection failure when logging the corresponding heartbeat sql, id value to hotdb.log
when the storage node is detected handover, current logging operations Seconds_Behind_Master, gtid, binlog position information
when the slave device detects the presence of copy delay and greater than 10s, switch over the primary storage node handover failure information log output copy delay information
- the storage node is switched dynamically loading problems
in HotDB Server Operator reload port operation, if there is a storage node or configuration repository switching directly reload fails with an error
executed on the management platform dynamically loaded detects a storage node or configuration libraries switch the user is prompted whether to abort switch. Abort switch is switched user selects the storage node configuration database or stopping force for dynamic loading operation (new reload config force command HotDB Server management port)
-HotDB Server error coding standards
HotDB Server coding error message related abnormalities were standardized, start from 10000
