An overview

According to the bidding requirements of this period, the center shall realize the integration of migrated applications and newly built business platforms after completion. Considering that the existing applications of the command center to be relocated include the analysis and management platform and the command platform, which are the core and important applications of the center, our company believes that the relocation of the original system will be the focus and difficulty of the project construction. This project design is based on the preliminary technical exchange between our company and the user’s current system contractor company and the analysis of the user’s current situation, and gives the relocation project design.

System migration requirement analysis

According to the bidding requirements of users, the specific demand analysis of system migration in this period is as follows. All the original application systems in the center will be migrated to the virtualization service platform. During the migration, the work shall not be interrupted and the historical data shall not be lost. After migration, the system is integrated and connected with the multimedia communication command platform. The difficulty of system migration is the reasonable planning of system cutting time nodes and the transformation of telephone access routes. The nondestructive migration of historical data is also the difficulty and focus of system migration.

Overall conclusion of central system migration demand analysis

Based on the demand analysis of the center’s existing application migration, considering that the original system construction company is not our company, there are still engineering risks to coordinate with the original construction company in the process of migration. Our company believes that the key contents of system migration include: access cutting involving operators, migration of original data, and reasonable node planning of cutting time.

General idea of the migration solution

Central system migration is a whole system engineering. Migration must ensure the relevant requirements of user system construction, in the design of migration scheme, we focus on several issues.

Minimize service downtime

Service interruption poses great recovery risks to both the running environment and the test environment. Such risks are unacceptable especially for time-sensitive data and data integrity services. We give such a requirement, considering how to minimize the downtime, can achieve the construction goal of zero downtime? 1. For the server operating system, we can adopt the P2V method and use the Volume Shadow Copy Copy service of the operating system as the basis to realize the system in the old system environment without modification and without shutdown. Migrate data, application software, operating system environment, and system environment variables to the new server in the form of snapshots. Thus the whole migration of the server environment is realized. 2, for other application servers and middleware application, we can expand dynamically based on the application server cluster, to implement the server does not stop node operating environment to increase business, it can realize the application server in the “hot” is added to the new environment failover/load balance cluster system, In some application services, we can use session session replication to realize the global environment variables and session request state of the old system to migrate to the new environment. Considering the fast real-time of session replication and state, we can adopt session memory replication, and considering the security of session replication and state, we can adopt session database replication management. 3, for the database, we can based on the database itself own database image technology, the database log transmission technology to realize the respective depots, migrate, library building, database mirroring technology allows us not only to ensure that the database migration non-stop, but also can guarantee one thousand migration in outage also does not affect the source database, The log transfer technology to build the migration can ensure that the system database migration in an asynchronous way, so that our system environment in the case of network failure can still carry out the normal work of the migration task window.

Service cutting time node optimization

For applications such as ××× system that need to ensure uninterrupted external services, it is necessary to select the optimal cutting time node by analyzing user history applications, and propose backup links and manual processing methods during cutting.

Post-migration integrity testing

Migration involves not only the operation of applications, instances, and databases, but also the integrity testing of pre-migration planning and post-migration testing. These tests include but are not limited to data consistency tests, data integrity tests, application session state integrity tests, connection interruption tests, and data recovery tests. Only in this way can the security and effectiveness of migration be guaranteed.

Server hardware environment migration scheme

According to the bidding requirements of users, the service hardware environment of this project is mainly the migration from the original blade server to the cloud service platform of the new procurement in this bidding. The cloud service platform can virtualize the original server hardware environment and operating system environment, reducing migration difficulty.

The migration assessment

Before the migration, our company will evaluate the migration plan to ensure the success of the migration. Firstly, our company will send staff to investigate the architecture and resource usage of the existing system. The evaluation process must include the following information and contents:

  • The number of services supported by the existing system and their distribution in the server;
  • The resource usage of existing physical servers, including CPUS, memory, disks, and network connections, must be the same as that of the original physical server. To ensure successful migration, ensure that the specifications of the target VM are not lower than those of the original physical server.
  • Whether the current physical environment supports virtualization and resource expansion because virtualization must be performed on physical servers before migration.
  • To evaluate the current storage capacity and resource utilization, plan the storage space required for migration in the target system. For example, on some servers, system disks and user disks are created on local disks, and on some servers, system disks are created on local disks and user disks are created on SAN/NAS.

Migration plan

Through the evaluation of the existing network environment, we have a clear understanding of the existing resource utilization, service and system requirements, and then we can start to plan the migration. The steps are as follows:

  1. Determine the migration steps, including the migration sequence of all servers, in descending order of risk.
  2. Determine the backup plan. Because the existing system will be hardened and some servers can be reused through virtualization, all data must be cleared before virtualization. Therefore, back up these servers to ensure service continuity.
  3. Determine and prepare the tools required for migration, including the necessary functions of the tools in migration and the network environment required for using the tools.
  4. Identify an additional test environment before the actual migration begins that can boot the tests to ensure a successful migration. Therefore, the test environment needs to specify the number of servers and storage to be designed.
  5. Plan the network environment. Because servers on the network are in different locations, consider the network connection, data backup mode, and network traffic source during migration to determine whether network traffic causes network congestion
  6. Determine the migration cycle and the people involved, including migration start and end times, team capacity building, and the roles of team members.

The test plan

After the migration plan, perform a small batch test migration solution, which involves testing and auditing the first migration, as follows:

  1. Prepare the test system environment to which the first batch of servers will be migrated for the test migration.
  2. Install and verify the migration tool to perform the P2V migration for the first batch of servers.
  3. For the first batch of servers, you need to analyze the storage system, regardless of whether the server uses local disk storage or remote SAN/NAS storage systems during storage migration.

The migration test

After the small batch test migration of the first servers and services, test the migrated servers, including unit and performance tests.

The migration implementation

During the migration, all servers will be migrated to the virtualization system. Perform the following steps:

  1. Ensure that the entire network environment for batch migration is ready, and complete the connection between the source system and the target system through the migration tool. The target system here is a transit system.
  2. Perform performance reviews and health checks on the migrated system, and if system health monitoring is performed, deactivate the old system and temporarily move its services to the new virtualization system.
  3. Some of the available old hardware can be reused in server virtualization, and some software resources need to be expanded, such as memory and hard disk. These servers make up the final virtualization infrastructure, the final system.
  4. Finally, V2V migration is performed between the target system and the final system.

Back up data before virtualization

To virtualize physical servers in the old system, consider the impact of server virtualization. For example, existing servers are reused. CPU, memory, and hard disk resources of these servers are reused during server virtualization. However, some services on these servers are still running. Therefore, before performing migration and virtualization, you must first back up the required servers. A physical backup server has been provided and virtualized. Data and servers have been backed up to the virtualization system.

  1. First, there are typically multiple services running on the servers to be migrated, and these servers are considered reusable in virtualization scenarios after the migration evaluation. However, a long downtime is not allowed during migration. Therefore, you need to prepare a backup VM that uses a virtualization platform and use P2V to back up the server to the VM.
  2. After backing up all servers to be virtualized, install virtualization software on these servers to virtualize them. Based on the capacity plan determined in the evaluation phase, create VMS of corresponding specifications on the virtualization platform and use their computing resources to take over services in the old system.
  3. After all VMS are ready, plan and install migration tools to migrate services from a backup system to VMS in a virtualization system. Vm migration refers to migrating application services from a backup virtualization system to the final virtualization system.
  4. After the VM migration is complete, test these services, disable the original system, and switch all services to the virtualization system.

Detailed migration procedures

The detailed steps and description of the migration are as follows:

  1. During the evaluation phase, the following information needs to be collected before virtualization and migration:
    • Performance statistics include CPU usage, memory usage, disk IOPS, and disk usage.
    • Configure physical servers, including CPU specifications, memory capacity, and disk capacity. Locate physical servers, determine whether virtualization is supported, count the number of servers that support virtualization, and plan hardware to be added in virtualization.
    • Collect information about the usage and configuration of the current system in the non-agent collection and agent collection scenarios. You can use an information collection tool.
  2. Analyze the dependency conditions of existing services and back up the current system.
    • Determining the dependency of application systems on servers can be used as a reference for determining the migration priorities of all servers.
    • After determining the dependency conditions of each service, back up the server to be virtualized.
  3. Capacity planning and virtualization execution
    • Calculate the capacity required by virtualization based on the current resource usage and demand.
  4. Planning Application Services
    • In the planned solution, VMS of the same type are deployed in the same computing resource pool and can share storage or computing resources with each other in the same pool. The failure of one cluster does not affect other resource pools.
  5. Virtualization planning and VM allocation
    • After establishing the virtualization platform, prepare the final migration resources. Before the migration, if server A has dual-core cpus and 2 GB memory, create a 2-core /2 GB VM on the virtualization platform and allocate disks to the VM.
  6. Planning migration tools
    • Use a migration tool to copy disks from a physical or virtual server to the final virtual system.
  7. Perform online migration using tools
    • After preparing the source system, target vM, and target system, determine the migration tool and migration policy.
  8. The migration test
    • After the migration, perform tests to verify whether the migration is successful. The test scenarios are as follows: Monitoring basic virtualization functions after the migration of application services.
    • Whether the features and functions of application services are almost the same before and after migration;
    • Performance monitoring of virtualization systems;
    • … .
  9. Decommissioning the old system
    • Existing servers have been virtualized and reused. Services on other servers that do not support virtualization have been migrated to the virtualization platform. You can switch application services to the virtual system and disable the old system.

Carrier access link (route) migration

Carrier access link (route) migration mainly involves the application of physical link required by the new center, the making of phone number access route, the opening of test number before application formal cutting, and emergency measures on the cutting day. For the content of the first four parts, a full backup can be made according to the needs of the center, and simulation operation test can be carried out before the system is formally cut. Emergency safeguard measures should be taken on the cutting day. If the cutting fails, the operation of the original routing guarantee system should be switched back quickly. At the same time, back up enough backup links in the new command to support manual processing. The specific configuration scheme of the above link will be further confirmed after the bid is awarded.

Application system and database migration solution

For the construction of this project, before the migration of the application system and database, we will deploy the same operating system, middleware and server management platform software environment as the original application in the cloud platform newly purchased by users through bidding to ensure the lowest risk of environmental changes in the migration.

Application Server Migration

For the application system migration of this project, the original system is based on Tomcat application environment and Java application framework. This plan proposes to build NLB cluster of Tomcat environment for application environment such as Tomcat and Java application framework, add the current system to NLB cluster without stopping, make it become one node in the cluster, and the new environment is another node. Exit the migration cluster after implementation and add the new environment to the newly built NLB cluster. NLB not only achieves load balancing, but also many forms of redundancy. NLB is mainly used in environments where files do not change much and are not resident in memory, such as WEB services, FTP services, and VPN services. When users access the cluster, the cluster can allocate the access request to each server in the cluster to achieve load balancing. These servers are called cluster nodes. In load balancing, the file requirements for each node are generally the same. This way, each node returns a consistent result to the customer. Generally speaking, an NLB requires at least two nodes, one of which cannot be used, and all the load falls on the remaining node, i.e., full load. The NLB provides three redundancy functions: software redundancy, hardware redundancy, and site redundancy.

Database migration implementation

For the database migration of this project, it is necessary to move the historical data files accumulated in the center to the new server of the center, and it requires the minimum downtime. Meanwhile, the difficulties also include that the servers are not in the same machine room.

  1. Analysis and design ideas
  • In view of the characteristics of the database relocation environment of this project: the first is that the database file is relatively large; The second is that the speed of file transfer may be slow (wan transfer). The initial solution is as follows.
  • In order to minimize the downtime, we use full backup and differential backup to migrate the database. We make a full backup (xxx_full.bak) of the database to be migrated during the day, and copy the backup file (FTP software can be used for resumable) to the target server for restoration. Make a differential backup (xxx_diff.bak) after off-hours, copy the differential backup to the target server, and then make a differential restore on the basis of the full restore.
  • The downtime here is equal to differential backup time + differential backup file transfer time + differential backup file restoration time. There is no downtime.
  1. Ensure data security and auditability during data migration
  • The security of data migration cannot be ignored. The design of this scheme realizes migration security and operation auditing based on multiple data auditing functions.

Specific organization and implementation plan of system migration

For the construction of this project, the relocation of the production system of the center is involved. The system mentioned above has the characteristics of short downtime requirement, complex system structure, long test time, various equipment, multiple users and complex levels. The time for the relocation of this project is very tight, and the stability between the equipment is also a test. Therefore, it is necessary to coordinate the relationship between all units and make concerted efforts to complete the relocation project within the scheduled time. The relocation organization of this project shall be established on the premise of not affecting daily work as much as possible or reducing the impact to the minimum, that is, the relocation shall start at the minimum time node of the minimum guarantee content day, and the relocation, installation and testing of the servers and network equipment that must be relocated shall be completed as soon as possible. And after the boot, continue to track the operation of the system, at any time to deal with the abnormal situation of system operation. Relocation requires full coordination and cooperation of the original system construction company personnel to complete the relocation task.

The relocation plan

Implementation process:

  • The process is mainly formulated according to the needs before relocation, mainly to understand the current system equipment and system operation in detail. Make detailed relocation plan and emergency plan according to the knowledge.
  • Professional engineers understand the equipment room status and requirements of the customers after relocation. In order to minimize the impact of computer room migration on users, we should make a detailed and feasible migration plan by fully considering all possible situations in the implementation process.
  • Prepare physical layout table, connection table, and cable number table before and after relocation. It can be classified into multiple systems according to user conditions.
  • XXX technical personnel are required to cooperate closely in the process of relocation.
  • In order to ensure the smooth, orderly and safe operation of the relocation work, detailed relocation process will be formulated, detailed division of labor will be carried out, and the specific work will be arranged to the person and the responsibility to the person.
  • At least 2 people should be assigned to each job to ensure the accuracy of the work.

Detailed implementation Plan

In order to move smoothly on time and ensure the normal operation of the equipment after the move, we have developed a series of simple and clear worksheets to help the engineering implementation personnel determine whether the work to be performed in the various relocation work is completed. Avoid work mistakes, avoid delays caused by relocation work.

Implementation process:Requirements for the equipment room:

Before moving the equipment room, check whether necessary equipment and facilities meet the requirements. This worksheet is the prerequisite to ensure the stable and normal operation of the equipment room after the move. The person in charge of the move and related personnel should fill in this worksheet before the move.

Emergency treatment

If any exception occurs after the device is relocated, the on-site technical personnel should immediately check the device to determine the fault location. If the hardware fault is within the scope of spare parts, replace it immediately. If the spare parts are not within the scope, use the spare device immediately and enable the spare device as soon as possible. If the configuration data or the system fails to start, use backup software tools such as the system CD-ROM to install or restore the system software.