CN1275476C - Clustering system for utilizing sharing internal memory in mobile communiation system and realizing method thereof - Google Patents
Clustering system for utilizing sharing internal memory in mobile communiation system and realizing method thereof Download PDFInfo
- Publication number
- CN1275476C CN1275476C CN 03131803 CN03131803A CN1275476C CN 1275476 C CN1275476 C CN 1275476C CN 03131803 CN03131803 CN 03131803 CN 03131803 A CN03131803 A CN 03131803A CN 1275476 C CN1275476 C CN 1275476C
- Authority
- CN
- China
- Prior art keywords
- machine
- state
- standby
- server
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Images
Abstract
The present invention relates to a cluster system using a shared memory in a mobile communication system, which comprises a plurality of virtual communication nodes. The cluster system is characterized in that each virtual communication node is composed of communication nodes where two servers are, the two servers are jointly connected to one shared memory device, and each server at least comprises the following functional modules: a master control module, a plurality of monitored object module and a communication module. Compared with the prior art, a public magnetic matrix does not need to be used in the present invention owing to the adoption of a database locally established, and the present invention has the advantages of cluster server cost saving, data security improvement, data usability improvement and system reliability improvement.
Description
Technical field
The present invention relates to the close coupling multimachine system of field of mobile communication, particularly relate to a kind of database of setting up in cluster system server this locality to improve device and its implementation of attaching position register (Home Location Register is called for short HLR) user data availability.
Background technology
In the mobile communcations system, very high to the availability requirement of user data, need to realize 24 hours addressable, if user data is unavailable, will cause a large number of users can't use mobile phone, bring loss difficult to the appraisal.
Improve the availability of user data, mainly be to adopt the system of Cluster Server (Cluster) to realize at present, Cluster Server is a kind of close coupling multimachine system, it by one group independently computer constitute, these computers are cooperated each other, just as one-of-a-kind system, can allow the client use all the time in order to guarantee mission critical program and resource.
The concrete scheme of prior art is: the shared magnetic battle array of two-server, database is built on the magnetic battle array.Two-server is shared an external IP, and by active/standby competition, wherein a station server obtains external IP becomes main frame during startup, and other one then becomes standby host; When system moves, whether the database service of Cluster software supervision is normal, if it is unusual to find that the database service of main frame occurs, then carry out active/standby switching, allow standby host take over external IP and treatment S QL (Structured Query Language, SQL) request continues to provide service.Because entire database is based upon on the public magnetic battle array, so when Cluster Server was done active/standby switching, the data in the database can not lost.
Though adopt the mode of the public magnetic battle array of Cluster Server can solve active and standby machine data consistency problem, also brought huge hidden danger thus:
1. in the running, depositing of data relies on public magnetic battle array fully, in case after public magnetic battle array is damaged, can cause loss of data, and can't recover;
2. use the cost of public magnetic battle array too high;
3. the Cluster software on the market is generic management software, can't combine closely with existing mobile communcations system, and when database broke down, Cluster software can not Real-time Alarm.
Summary of the invention
Purpose of the present invention overcomes the defective for above-mentioned prior art just, and propose in a kind of mobile communcations system to use cluster system and its implementation of shared drive, database is set up in server this locality at cluster system, and utilizes the shared drive device to replace public magnetic battle array to improve the availability of data in the HLR register customer data base.
In order to achieve the above object, the invention provides the cluster system that uses shared drive in a kind of mobile communcations system, comprise several virtual communication nodes, it is characterized in that each virtual communication node is made of the communication node at two server places, and these two servers are connected on the shared drive device jointly, and each server comprises following functional module at least:
One main control module, be used to realize between virtual communication node server two-shipper active/standby state competition, active/standby machine is switched and monitoring mutually, is used to monitor each monitored object on this machine simultaneously;
A plurality of monitored object modules are used for all application programs relevant with business of supervisory control system, and the submonitor of other application and resource;
One communication module is used for being in mainly when use state when this machine, binds external IP and sets up communication link with other nodes and communication interface with other nodes is provided.
The shared drive device, comprise two-shipper control board and shared drive plate, the two-shipper control board provides bus interface, the active/standby State Control service that connects the shared drive plate for server, the two-shipper control board has expansion paged memory management function, make server can directly visit memory resource on the shared drive plate, the shared drive plate is preserved the data of monitored object visit and the data that server can directly be visited.
The present invention also provides the cluster system implementation method of using shared drive in a kind of mobile communcations system, and this method may further comprise the steps:
When two servers of composition dummy node power on, finish active/standby state competition;
The database manipulation agent process is collected the database manipulation request of all clients, and these requests are stored into the shared drive district;
The master carries out practical operation with the data in server agent process to database,
Main carry out Log backup, and described backup log is transferred to standby server, data are applied in the local data base of standby server by the data synchronizing process of standby server with the data in server synchronized process.
The present invention compared with prior art owing to taked database to build this locality in, does not need to use public magnetic battle array, has saved the Cluster Server cost, has improved safety of data and availability; And Active/Standby Changeover of the present invention all realizes by the hardware register of operation two-shipper control board, thoroughly solves the situation that commercial Cluster software is switched unusually, failed.And database Real-time Alarm function is provided, and combine closely with existing mobile switch system, improve the operational reliability of whole system greatly.
Below in conjunction with embodiment, and with reference to accompanying drawing the present invention is described in detail so that purpose of the present invention, feature and advantage are had more deep understanding.
Description of drawings
Fig. 1 is hardware system structure figure provided by the invention;
Fig. 2 is a software system function module map provided by the invention;
Fig. 3 is an active/standby machine competition flow chart provided by the invention;
Fig. 4 is that standby host provided by the invention is switched the operational flowchart into main frame;
Fig. 5 a is that main frame provided by the invention is switched the operational flowchart into standby host;
Fig. 5 b is that main frame provided by the invention is switched the operational flowchart into standby host;
Fig. 6 is a database request process chart provided by the invention;
Fig. 7 is the active/standby database synchronization operational flowchart of using provided by the invention;
Fig. 8 is an alarming processing functional block diagram provided by the invention.
Embodiment
As shown in Figure 1, be cluster system hardware structure diagram provided by the present invention, a virtual communication node in the cluster system is made of jointly the communication node of Server A and Server B, these two server S erver A (server A) 101 and Server B (server B) 102 connect a shared drive device 103 jointly.Wherein two servers can carry out the active/standby state exchange of using, and concrete configurations such as its CPU (central processing unit CentralProcessing Unit), internal memory, hard disk depend on concrete application and decide.
Described shared drive device is made up of two parts: a part is the two-shipper control board, and another part is the shared drive plate.Wherein, the two-shipper control board as one independently machine element be inserted on the PCI slot of each server master board, bus interface, active/standby State Control and other service that connects the shared drive plate is provided for server, has expansion paged memory management function, server logical address can be mapped as physical address, so that can directly prevent asking the memory resource on the shared drive plate; The shared drive plate is a physical storage device, logically be divided into 2 parts: a part is the data field of monitored object visit, message sink, the transmission exchange area that another part is main control module, it is used to preserve the data that Server A and Server B can directly visit, but the extendible parallel work-flow of the capacity of shared drive plate, simultaneously, the Single Point of Faliure that causes for fear of server outage and the consequence of the shared drive plate loss of data that brings, the shared drive plate adopts dual power supply.
As shown in Figure 2, system function module of the present invention mainly comprises: main control module 201, communication module 202 and monitored object module 203.Wherein, main control module be used for realizing between two-shipper (server A as shown in Figure 2 and server B) active/standby state competition, active/standby machine is switched and function such as monitoring mutually, is used to monitor each monitored object on this machine simultaneously; Communication module 202 is main to be responsible for and communication interface realizes, mainly bind external IP during with state and set up communication link with other nodes and communication interface with other nodes is provided when being in; Monitored object module 203 comprises all monitored objects, and they can be the application programs relevant with business, also can be the submonitor (monitored object is exactly DbAgent process and DbSync process in the current version) of other application of monitoring or resource.
Server in the native system adopts the active/standby working method of using.During operate as normal, only allow a machine works under master mode.Because two servers were not determined active/standby state when system powered on, therefore need to start earlier main control module and carry out active/standby state competition, two resource distribution registers are all arranged: partner's machine status register and local state register on the two-shipper control board of active/standby machine, some circuit is controlled the state variation of above-mentioned two registers of active and standby machine on two-shipper control version, and the principle of competition is to determine by the status register of the partner's machine on the inquiry two-shipper control board.
As shown in Figure 3, be the active/standby state competition of server in the cluster system of the present invention flow chart, in order to the active/standby state competition of server after finishing system and powering on.Its operating procedure is as follows: after active/standby machine powers on, at first the base register on the initialization two-shipper control board (be included in the line index setting, main with allow, permissions that reset certainly, external bus state wrongly interrupt allowing, post that even parity check is interrupted allowing, external clock interrupts allowing, hardware watchdog is opened, the bus error isolation features open, 5 milliseconds of regularly interrupt clear, startup, the removing of fault interrupt request register, communication disruption removings), begin this machine master and use state application, step 301; The time-delay circulation is set, the current state of from the status register of partner's machine, reading partner's machine, step 302; Judge whether the application of this machine is main successful with state, step 303; If failure, it is main usefulness that partner's machine is set, and then the local state register is standby, step 304; Otherwise, the current state of from the status register of partner's machine, reading partner's machine, step 305; If partner's machine has been set to main usefulness, it is standby that the local state register then is set, step 308; If partner's machine is not set to main usefulness, it is main usefulness that the local state register then is set, step 306; It is standby that partner's machine is set, step 307; After main and standby competition is finished, detect the whether existing external IP of this machine,,,,, otherwise add external IP if the existing external IP of this machine is not then added for main frame if having then delete this external IP for standby host.Start communication module then and set up active and standby internal communication link and external communication link.Under the normal condition, communication link can be set up successfully about 2 seconds.After main and standby competition is finished, start monitoring module according to local state.
As shown in Figure 4, switch into the operating process of main frame for standby host application in the cluster system of the present invention, detailed process is as follows: judge at first whether this machine is in stand-by state, step 401; If not, this EP (end of program) then; If this machine is in stand-by state, then obtain the state of partner's machine by hardware, promptly utilize 10 milliseconds of timed tasks in the standby host main control module to monitor state as partner's machine two-shipper state of a control register of main computer, step 402; Judge whether current partner's machine as main computer is in the master and uses state, if not, shows that current main computer is non-service attitude, then this machine is led with application, step 404; By this machine main control module as standby host this machine two-shipper state of a control register being set is that the master uses state, and step 405, this step comprise that also it is that the master is the main state of using with state, standby host monitoring module that the standby host main control module is set.
As shown in Figure 5, switch into the operating process of standby host for main frame application in the cluster system of the present invention, its concrete steps are as follows: at first, judge whether this machine is in main use state, step 501; As denying then current EO; Use state if this machine is the master, then judge whether carrying out data sync operation, step 502 between the active/standby machine; Judge whether active/standby machine is in warning information synchronous regime, step 505; If carrying out data sync, or active/standby machine is in the warning information synchronous regime, and whether then need further judge currently needs to carry out active/standby machine period of state and switch step 503; In this way, then postpone to carry out after five minutes active/standby machine state and switch step 504; As not, then operating process finishes, and does not carry out state and switches; If foregoing active/standby machine is not in the warning information synchronous regime, then obtain the current state of partner's machine, step 506 by this machine two-shipper state of a control register; Whether current partner's machine is in main use state, step 507; Use state if partner's machine is the master, this machine stand-by state then is set, step 508; Do not use state if partner's machine is not the master, judge then whether it is in stand-by state, step 509; If this moment, the partner machine neither stand-by state, the hardware that this machine then is set is in main uses state, step 510; And this machine software of setting is in main use state, step 511; If this moment, partner's machine was a stand-by state, then this machine hardware is stand-by state, step 512; Obtain the state of partner's machine in the system, step 513 by this machine two-shipper control board; Whether current partner's machine is in is main use state, step 515; If the master uses state, it is stand-by state that this machine hardware then is set, step 516; And this machine software of setting is in stand-by state; Step 517; Otherwise, if partner's machine is not in stand-by state at present, whether then judge the set time-delay cycle-index of system, if this machine hardware then is set is in main use state, step 519 greater than 10; And this machine software of setting is in main use state, step 510; Otherwise, then postpone 100ms after, return step 512; Continue follow-up flow process.
In the above-mentioned flow process, its switch flow process and standby host/main frame to switch flow process the same, it all is the status register parameter that this machine two-shipper control board at first is set, it is stand-by state that the main frame main control module is set then, the host monitor module is a stand-by state, below several situations can realize that the main frame auto switching becomes standby host:
1. Manual Switch is pressed and is led with the button of switching on the panel, if satisfy condition: a. partner machine is in stand-by state; B. the non-data sync state that is in of active and standby machine; C. the non-warning information synchronous regime that is in of active and standby machine; Main frame will be switched and be that standby host, standby host will be switched and be main frame.
(2) backstage is man-machine switches, and selects the master with switching order from the alarm management of backstage, if satisfy the condition of switching of (1), main with switching to standby, standby will switching is main using.
(3) cycle switches, and the permission cycle is set switch on alarm device, and switching time is set, and switching time arrives, if satisfy the condition of switching of (1), switches immediately and realizes.Do not switch condition if do not satisfy (1), will postpone to switch again in 5 minutes.
(4) main with detecting the database manipulation failure, if the condition of switching of satisfied (1) will be switched to standby, it is main using that partner's machine is switched.
(5) main with detecting with other mobile switch system node communication failures, the master will switch into standby request command with sending to monitoring module, if satisfy the condition of switching of (1), partner's machine and background communication are normal simultaneously, main computer will be switched to standby, and standby will switching is main using.
(6) main little with detecting the disk remaining space, if satisfy the condition of switching of (1), will switch into standbyly, partner's machine will switch main usefulness.
As shown in Figure 6, for database request provided by the invention is handled block diagram, it may further comprise the steps: client-side program 601 is communication interface with the request communication node 602 of client, and the communication node 604 that sees through network 603 and corresponding server section is forwarded to the master and acts on behalf of DBAgent process 605 with data in server; DbAgent process by this main usefulness stores the up-to-date log serial number that obtains in operation requests and the database in the shared data 606 in shared drive district again; The DbAgent process of main usefulness is applied to operation requests main with database 606, and possible operating result is returned to client; The data sync DbSync process 607 of main usefulness backs up the daily record of the up-to-date variation of database, and backup file is transferred to standby host, is responsible for data are returned in the local data base by the DbSync process of standby host; After standby DbSync process is finished database recovery, the DbAgent process of notice standby host, the record ignored in search and deletion shared drive district.As what mention among Fig. 4 and Fig. 5, cluster system provided by the present invention, its active/standby state can be according to circumstances artificial or auto switching, so, the active/standby state of server A and server B is relative, supposition server B is that the master uses state among Fig. 6, and then relative server A just is called stand-by state.
As shown in Figure 7, being the standby host state transition diagram of database synchronization module provided by the invention, is example with standby host state transition, arrow is depicted as the migratory direction of state among the figure, and the literal on arrow next door is for triggering the condition of state transition, at first, when synchronized process started, standby host was in starting state 701; Wait for the notice that powers on of main control module, receive standby host power on notify after, switch to no initializtion state 702; Carry out initialization operation, this content of operation comprises initialization SQL service, FTP service, carry out that database recovers fully, acquisition and the same time point of host data base etc., switches to idle condition 703 after the initialization success; Under the idle condition, standby host sends the daily record synchronization request to main frame, and enters log request state 704; After receiving Log backup response (promptly finishing notice), enter daily record and return to form 707, finish data and recover; After the success, return idle condition 703, wait for sending of request next time; If in above-mentioned daily record synchronously, mistake occurs, then enter variance data solicited status 705 from the current idle condition of standby host, the request variance data is synchronous, wait to receive that notice is finished in the difference request after, enter difference and return to form 708; If difference repeatedly mistake occurs synchronously, the complete data request is then proposed, standby host enters complete data solicited status 706 from idle condition, the requested database complete data; Receive finish notice after, enter complete data and return to form 709;
In the above-mentioned state transition process,, to the unified alarm of user, and change wrong attitude 710 in any state switches by current state by alarm server no matter the mistake that takes place all can send the alarm server of the system of alarm; Analyze the Fail Type 711 of this alarm, generally comprise daily record and recover failure, difference recovery failure or recover failure fully, alarm if daily record recovers Fail Type, then change difference solicited status 708 over to, it is synchronous to finish difference; If difference is recovered failure, the complete data request is then proposed, migrate to the complete data solicited status, the requested database complete data; If recover failure fully, then standby host state transition is to no initializtion state 702; Again carry out initialization and, carry out new state transition according to new request.For guaranteeing the unanimity of master/slave data database data, regularly execute all data backup and restore funcitons, in various request attitudes, constantly the transmit status check request is given main frame, judge by main frame whether active/standby machine state is consistent, if state is inconsistent, main frame sends the solicited status of notice compulsory commutation standby host, and standby host will be initiated request again.
As shown in Figure 8, be alarming processing functional block diagram of the present invention, the alarm of broad sense comprises general notice and high severity alarm, the former only shows the not reproducible or transient fault of this system's emerged in operation, reset as standby server, its processing is very simple, and main is to give the backstage this forwards with server.And the high severity alarm meeting continues for some time, and until the disappearance of fault, it handles more complicated, and the desired solution of the present invention is exactly this high severity alarm (being designated hereinafter simply as alarm).
When the monitoring process on the server detects a fault, send an alarm information to main control module, comprise alarm head and alarm body in this message.Include alarm code, alarm cause in the alarm head, wherein alarm code and alarm cause are Unified coding.It in the alarm body content that each alarm must further describe this alarm.Main control module is kept in the internal memory after receiving an alarm, simultaneously it is passed to the backstage alarm module.
After an alarm produces, solid line among the figure represents that partly monitoring program passes to alarm the flow process of backstage alarm module, in this flow process, in order to prevent to repeat alarm, each alarm source in certain part after main control module alarm, should put one " alarming " sign, main control module also can be changed to the alarm attitude to corresponding part after receiving an alarm.
Equally also send to recover message during alarm clearing, recover to comprise alarm code in the message, and the content of the sign trouble location relevant, as three grades of faults of database etc. with alarm code to main control module.
The alarm that in the master control warning processing module, has a Buffer Pool that can hold hundreds of alarms to be used to preserve this machine, when a new alarm produces, to in Buffer Pool, occupy a position, also distribute simultaneously an Aid to give it, Aid makes the serial number form at present, nybble is long, can accomplish that the Aid of each bar alarm does not repeat.
In most cases, the alarm and the alarm recovery message that mail to the backstage alarm module can not lost, but in order to prevent contingency, in program, added synchronisation measures, specifically: the backstage alarm module is regularly receiving that all alarm Aid of a module issue main frame master control alarm module, by master control alarm module retrieval alarm Buffer Pool, a certain Aid is arranged in the synchronization message that the backstage alarm module transmits, and do not have in the Buffer Pool, then send out this Aid and recover message, otherwise send out the alarm information of this Aid to the backstage.Dotted portion is represented the message flow of backstage to main frame master control alarm module Synchronize Alarm among the last figure.
For further specifying technical scheme of the present invention, be example with WINNT operating system+SQLSERVER2K customer data base below, the cluster system implementation method of using shared drive in the mobile communcations system provided by the invention is described, it may further comprise the steps:
The first step: on two-server Server A and the Server B in the cluster system of supposition the identical SQL Server DBMS of version (Database Management System data base management system) is installed respectively, at the local hard drive of Server A and the Server B identical customer data base of construction respectively.And cluster system software provided by the invention is installed respectively;
Second step: when server S erver A and Server B power on, finish active/standby competition, idiographic flow is referring to Fig. 4, and active/standby machine enters the Active/standby state respectively; Main control module in the software of trooping provided by the invention is active/standby machine resource of monitoring and external communication state in real time, monitors mutually by the status register information that reads on the two-shipper control board between the active/standby machine simultaneously.Virtual communication node of the common composition of the communication node of Server A and Server B;
The 3rd step: client-side program is realized by the inside story interaction mechanism the visit of database server, database manipulation agent process (DbAgent) is collected the database manipulation request of all clients, and these requests are stored into the shared drive district, and memory contents comprises operation requests statement, up-to-date log serial number and completion status sign.The master carries out practical operation with the DbAgent process of server to database, upgrades " completion status " sign in shared drive district after operation is finished, and this sign is changed to " finishing ";
The 4th step: the data synchronizing process DbSync of main frame backs up the daily record of the up-to-date variation of database of record, and this backup file is transferred to standby host, by the DbSync process of standby host data is applied in the local data base;
After the DbSync process of the 5th step standby host was finished database recovery, the DbAgent process of notice standby host searched the operation requests record of having finished in the shared drive district according to log serial number, is deleted.
In the above-mentioned flow process, if break down, then by data read operation to shared drive, the application again of request msg, can guarantee main standby machine switching like this after, database data can not lost.The concrete steps of troubleshooting operation are:
1) standby host monitors hostdown (stipulating some monitoring standards) or main frame initiatively is initiated to the request of changing;
2) communication node of standby host is taken over dummy node, accepts the request that client is initiated;
3) record in the standby host DbAgent process retrieval sharing data area, the lsn of not deleted record and the lsn in the current database are compared,, then re-execute it if there is record not to be applied in the current database, subsequently, empty completed operation requests in the shared drive;
4) all processes of standby host are transformed into and main use state, and all processes of main frame change into stand-by state.
Embodiment described above is illustrative and not restrictive, and under the situation that does not break away from the spirit and scope of the present invention, all variations and modification are all within protection scope of the present invention.
Claims (12)
1. use the cluster system of shared drive in the mobile communcations system, comprise several virtual communication nodes, it is characterized in that each virtual communication node is made of the communication node at two server places, and these two servers are connected on the shared drive device jointly, and each server comprises following functional module at least:
One main control module, be used to realize virtual communication node server active/standby state competition, active/standby machine is switched and monitoring mutually, is used to monitor each monitored object on this machine simultaneously;
A plurality of monitored object modules are used for all application programs relevant with business of supervisory control system, and the submonitor of other application and resource;
One communication module is used for being in mainly when use state when this machine, binds external IP and sets up communication link with other nodes and communication interface with other nodes is provided;
Described shared drive device, comprise two-shipper control board and shared drive plate, the two-shipper control board provides bus interface, the active/standby State Control service that connects the shared drive plate for server, the two-shipper control board has expansion paged memory management function, make server can directly visit memory resource on the shared drive plate, the shared drive plate is preserved the data of monitored object visit and the data that server can directly be visited.
2. use the cluster system of shared drive in the mobile communcations system as claimed in claim 1, it is characterized in that described two-shipper control board is one to be plugged on the individual components on the server master board PCI slot.
3. use the cluster system of shared drive in the mobile communcations system as claimed in claim 1, it is characterized in that described shared drive plate is a physical storage device.
4. use the cluster system of shared drive in the mobile communcations system as claimed in claim 1, it is characterized in that described monitored object comprises database broker process and database synchronization process.
5. use the cluster system of shared drive in the mobile communcations system as claimed in claim 1, it is characterized in that described functional module comprises a warning processing module, warning processing module is made of master control alarm module and backstage alarm module, the master control alarm module is positioned at main control module inside, the backstage alarm module is positioned at the main control module outside, and be connected with the communication of master control alarm module, the master control alarm module produces alarm, and report the backstage alarm module, if alarm clearing, then report and alarm recovers message to the backstage alarm module, and the backstage alarm module provides the storage of warning information, filter, Presentation Function.
6. method that realizes using in the mobile communcations system as claimed in claim 1 the cluster system of shared drive, this method may further comprise the steps:
When two servers of composition dummy node power on, finish active/standby state competition;
The database manipulation agent process is collected the database manipulation request of all clients, and these requests are stored into the shared drive district;
The master carries out practical operation with the data in server agent process to database,
Main carry out Log backup, and described backup log is transferred to standby server, data are applied in the local data base of standby server by the data synchronizing process of standby server with the data in server synchronized process.
7. use the method for the cluster system of shared drive in the realization mobile communcations system as claimed in claim 6, the flow process of wherein said active/standby state competition may further comprise the steps:
After active/standby server powered on, the base register on the described two-shipper control board of initialization began this machine master and uses the state application;
The time-delay circulation is set, the current state of from the status register of partner's machine, reading partner's machine;
Judge whether the application of this machine is main successful with state;
If failure, it is main usefulness that partner's machine is set, and the local state register is standby;
Otherwise, the current state of from the status register of partner's machine, reading partner's machine;
If partner's machine has been set to main usefulness, it is standby that the local state register then is set;
If partner's machine is not set to main usefulness, it is main using that the local state register then is set;
It is standby that partner's machine is set;
The external IP of deletion standby server;
Add the main server external IP of using;
Start communication module and set up active and standby internal communication link and external communication link.
8. use the method for the cluster system of shared drive in the realization mobile communcations system as claimed in claim 7, the operating process of main frame is switched in wherein said standby host application, may further comprise the steps:
Judge whether this machine is in stand-by state;
If this machine is in stand-by state, then obtain the state of partner's machine by hardware;
Judge whether current partner's machine as main computer is in the main state of using,
If not, show that current main computer is non-service attitude, then this machine is led with application;
By this machine main control module as standby host this machine two-shipper state of a control register being set is the main state of using.
9. use the method for the cluster system of shared drive in the realization mobile communcations system as claimed in claim 7, the operating process of standby host is switched in wherein said main frame application, may further comprise the steps:
Judge whether this machine is in the main state of using;
Use state if this machine is the master, then judge whether carrying out the data sync operation between the active/standby machine;
Judge whether active/standby machine is in the warning information synchronous regime;
If carrying out data sync, or active/standby machine is in the warning information synchronous regime, and whether then need further judge currently needs to carry out active/standby machine period of state and switch; When described warning information promptly monitored fault synchronously, main frame can send a warning message to the backstage, regular synchronous altering message between main frame and the backstage;
In this way, then postponing to carry out after five minutes active/standby machine state switches;
As not, then operating process finishes, and does not carry out state and switches;
If active/standby machine is not in the warning information synchronous regime, then obtain the current state of partner's machine;
Judge whether current partner's machine is in the main state of using;
Use state if partner's machine is the master, this machine stand-by state then is set;
If the non-master of partner's machine uses state, judge then whether it is in stand-by state;
If this moment, the partner machine neither stand-by state, the hardware that this machine then is set is in mainly use state, and this machine software of setting is in the state of using of leading;
If this moment, partner's machine was a stand-by state, then this machine hardware is stand-by state;
Obtain the state of partner's machine in the system;
Judge that whether partner's machine is in is the main state of using;
If the master uses state, it is stand-by state that this machine hardware then is set, and this machine software of setting is in stand-by state;
Otherwise,, judge that then whether the set time-delay cycle-index of system is greater than 10 if partner's machine is not in stand-by state at present;
If this machine hardware then is set is in and main use state, this machine software of setting to be in to lead the state of using;
Otherwise, then postpone 100ms after, continue follow-up operation.
10. use the method for the cluster system of shared drive in the realization mobile communcations system as claimed in claim 6, the flow process that wherein said database request is handled comprises that also step is forwarded to the main data in server agent process of using with the request of client;
Store the up-to-date log serial number that obtains in operation requests and the database into the shared drive district by main with the data in server agent process;
The master is applied to the master with the data in server agent process with operation requests and uses server database, and the result is returned client;
The master backs up the daily record of the up-to-date variation of database with the data in server synchronized process, and backup file is transferred to standby server;
The data synchronizing process of standby server returns to data in the local data base;
The record ignored in search and deletion shared drive district.
11. use the method for the cluster system of shared drive in the realization mobile communcations system as claimed in claim 6, the operation of the data synchronizing process of wherein said standby server may further comprise the steps:
Synchronized process starts, and standby host is in starting state;
After receiving the notice that powers on, standby host switches to the no initializtion state;
Carry out initialization operation, and standby server is switched to idle condition;
Under the idle condition, standby host sends the daily record synchronization request to main frame, and enters the log request state;
After receiving the Log backup response, enter daily record and recover turntable, finish data and recover;
If in described daily record synchronously, mistake occurs, then enter the variance data solicited status, the request variance data is synchronous;
If difference repeatedly mistake occurs synchronously, then enter the complete data solicited status, the requested database complete data;
The mistake that takes place during above-mentioned state switched sends the described warning processing module of the system of alarming, and the user is unified alarm.
12. use the method for the cluster system of shared drive in the realization mobile communcations system as claimed in claim 11, the operating process of wherein said warning processing module may further comprise the steps:
Analyze the Fail Type of this alarm;
If daily record recovers the Fail Type alarm, then standby server is changed over to the difference solicited status, it is synchronous to finish difference;
If difference is recovered failure, the complete data request is then proposed, change standby server over to the complete data solicited status, the requested database complete data;
If recover failure fully, then standby server is changed over to the no initializtion state, accept new request again, newly operate.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 03131803 CN1275476C (en) | 2003-06-04 | 2003-06-04 | Clustering system for utilizing sharing internal memory in mobile communiation system and realizing method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 03131803 CN1275476C (en) | 2003-06-04 | 2003-06-04 | Clustering system for utilizing sharing internal memory in mobile communiation system and realizing method thereof |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1553716A CN1553716A (en) | 2004-12-08 |
| CN1275476C true CN1275476C (en) | 2006-09-13 |
Family
ID=34322947
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 03131803 Expired - Fee Related CN1275476C (en) | 2003-06-04 | 2003-06-04 | Clustering system for utilizing sharing internal memory in mobile communiation system and realizing method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN1275476C (en) |
Families Citing this family (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100431371C (en) * | 2005-04-04 | 2008-11-05 | 中兴通讯股份有限公司 | System for automatic switching database in imputation position register and method thereof |
| CN1979447B (en) * | 2005-12-01 | 2010-05-05 | 孙贤和 | Inmemory data server |
| CN100486178C (en) * | 2006-12-06 | 2009-05-06 | 中国科学院计算技术研究所 | A remote internal memory sharing system and its realization method |
| CN101471810B (en) * | 2007-12-28 | 2011-09-14 | 华为技术有限公司 | Method, device and system for implementing task in cluster circumstance |
| CN103164384B (en) * | 2011-12-15 | 2016-05-18 | 中国银联股份有限公司 | The synchronization realizing method of multi-computer system shared drive and system thereof |
| CN103812674A (en) * | 2012-11-07 | 2014-05-21 | 北京信威通信技术股份有限公司 | Method for main and standby server replacement |
| CN103532753B (en) * | 2013-10-11 | 2016-08-17 | 中国电子科技集团公司第二十八研究所 | A kind of double hot standby method of synchronization of skipping based on internal memory |
| CN105204977A (en) * | 2014-06-30 | 2015-12-30 | 中兴通讯股份有限公司 | System exception capturing method, main system, shadow system and intelligent equipment |
| CN104468221B (en) * | 2014-12-12 | 2018-09-18 | 北京国双科技有限公司 | Server failure hot change-over method, device and system |
| CN105824571A (en) * | 2015-01-05 | 2016-08-03 | 中国移动通信集团四川有限公司 | Data seamless migration method and device |
| CN108819884B (en) * | 2018-05-30 | 2022-01-25 | 江铃汽车股份有限公司 | Terminal power supply control method for Internet of vehicles |
| CN114189429A (en) * | 2021-11-25 | 2022-03-15 | 山东云海国创云计算装备产业创新中心有限公司 | System, method, device and medium for monitoring server cluster faults |
-
2003
- 2003-06-04 CN CN 03131803 patent/CN1275476C/en not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| CN1553716A (en) | 2004-12-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US8161321B2 (en) | Virtual machine-based on-demand parallel disaster recovery system and the method thereof | |
| CN102394774B (en) | Service state monitoring and failure recovery method for controllers of cloud computing operating system | |
| CN1275476C (en) | Clustering system for utilizing sharing internal memory in mobile communiation system and realizing method thereof | |
| US8032786B2 (en) | Information-processing equipment and system therefor with switching control for switchover operation | |
| CN111078667B (en) | Data migration method and related device | |
| WO2007028248A1 (en) | Method and apparatus for sequencing transactions globally in a distributed database cluster | |
| CN103345470A (en) | Database disaster tolerance method, database disaster tolerance system and server | |
| CN105302661A (en) | System and method for implementing virtualization management platform high availability | |
| CN102955720A (en) | Method for improving stability of EXT (extended) file system | |
| CN1801107A (en) | Data recovery method | |
| CN102394914A (en) | Cluster brain-split processing method and device | |
| CN102937955A (en) | Main memory database achieving method based on My structured query language (SQL) double storage engines | |
| CN103220183A (en) | Implement method of Hadoop high-availability system based on double-main-engine warm backup | |
| CN103226483A (en) | DHBS (dual hot-backup system) and method thereof based on SOA (service-oriented architecture) and cloud storage | |
| Barry et al. | Implementing journaling in a linux shared disk file system | |
| CN102045187B (en) | Method and equipment for realizing HA (high-availability) system with checkpoints | |
| CN113515316A (en) | Novel edge cloud operating system | |
| CN101527656B (en) | Emergency switched system of NGN service platform database and realizing method thereof | |
| CN108259569A (en) | It is a kind of based on IPSAN share storage without acting on behalf of continuous data protection method | |
| CN107357800A (en) | A kind of database High Availabitity zero loses solution method | |
| CN101686261A (en) | RAC-based redundant server system | |
| CN105007172A (en) | Method for realizing HDFS high-availability scheme | |
| CN111752758B (en) | Bifocal-architecture InfluxDB high-availability system | |
| CN111601299B (en) | Information association backfill system under 5G framework | |
| CN100499387C (en) | A method of singleboard N+1 backup in communication system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20060913 Termination date: 20140604 |