Shulou(Shulou.com)06/01 Report--

In this issue, the editor will bring you about how to achieve real linux non-downtime data migration. The article is rich in content and analyzes and describes for you from a professional point of view. I hope you can get something after reading this article.

Interpretation of the three stages of the Development of Storage Virtualization engine

Nowadays, the systems of more and more information centers require continuous operation 24 hours a day, so as to ensure business continuity, especially in the financial and medical industries. Therefore, many information centers have deployed server dual cluster, storage cluster, and through the deployment of storage virtualization engine to achieve heterogeneous storage integration, data migration, high storage availability, dual live data center and other functions. Through this series of functions to achieve data security protection and rapid fault switching, so as to reduce business interruption time and ensure business continuity.

However, users are most concerned about how the storage virtualization engine moves into the customer's original environment, whether it needs downtime, and what kind of impact it will have on the user environment.

Let's talk about the deployment of the storage virtualization engine from the three stages of the development of the virtualization engine.

Stage 1: move into the user environment, and the original data needs to be exported and re-imported

In stage 1, the storage virtualization engine needs to initialize the array before it can see and manage the storage space, which cannot be directly taken over by the array where the user already has data. The following data migration operations need to be done:

1. Take an additional array B, which is taken over and initialized by the virtual engine, and then copy the data from array A to array B through the server, as shown in figure ① above.

two。 Copy complete, shut down the server and disconnect Array A from the server. Then connect array A to the storage virtualization engine, initialize array A (the original data will be erased), repartition the LUN, and map it to the server.

3. Then copy the data from array B back to array A. Thus, the storage virtualization engine takes over array A.

But in the whole process, we can see that it takes a long time for data to be copied back and forth, which is unbearable for ordinary users.

Phase 2: downtime and move into the user environment without data migration

Data does not need to be exported and imported, and the storage virtualization engine can directly take over the array with data. You only need the following steps:

1. The server needs to be down or the IO hangs, and then disconnect the connection between Array An and the server.

two。 The array connects to the storage virtualization engine, which inherits the data in the original array intact without making any changes.

3. The original storage LUN is mapped to the corresponding server through the storage virtualization engine.

4. Start the server and mount the corresponding LUN

Phase 2 solves the process of data export and re-import, but it also requires downtime for deployment. For some system environments that need 7 to 24 continuous operation, it still can not meet the requirements. At present, the Encapsulation technology of IBM SVC's Imagemode,EMC Vplex carries out the migration of the storage virtualization engine through similar principles.

Phase 3: non-downtime deployment, which can move in and out online

Phase 2 still needs downtime because, as shown in the state of phase 3, the same LUN An is mapped directly to the server and to the server through the storage virtualization engine. The MPIO multipathing software on the server side will think that it is two different disks, but it actually points to the same disk, that is, LUN A, so the data will not be written to the back-end array through the virtualization engine. This is caused by the VPD inconsistency of the storage device hardware information. Although the storage virtualization engine will not make changes to the data in the LUN taken over, the VPD (vital productdata) hardware information of the storage device, such as the GUID number of the disk, will be changed after the takeover, resulting in the same disk being considered as two different disks by the MPIO software.

Therefore, InfoCore overcomes this technical problem through transparent path migration technology, and realizes the function that the storage virtualization engine can move in and out of the user environment online.

InfoCore's transparent path migration technology can inherit the hardware information of the storage device while inheriting the data in the disk intact, so as shown in the figure ② above, the GUID number mapped by the same LUN through path 1 and path 2 is the same, and MPIO multipath software will think that this is the same disk mapped directly by the array through another path. In this way, the two paths can be aggregated together without downtime, and only one disk can be recognized by the system, so that the data can be written to the back-end storage through path 1 and path 2, and the storage gateway can move in and out online. The specific implementation process is as follows:

1. Disconnect path 2, and the server can still continue to read and write data to array A through path 1.

two。 The array is connected to the virtualization engine through path 2, and the data in the array is encapsulated and inherited by the virtualization engine, so that the server can still read and write data through path 1.

3. Map LUN to the front-end server through path 2, using transparent path migration technology, the server-side MPIO will think that path 2 is also directly mapped through the array, is the same disk, so that you can read and write data through path 2. Therefore, the failover or load balancing functions of path 1 and path 2 can be realized.

4. You can also disconnect path 1, automatically switch the IO path to path 2, and connect path 1 to the server through the virtualization engine. As shown in the above figure ③.

5. Through transparent migration technology, the storage virtualization engine can be moved into and out of the user's original environment online without downtime. As shown in the figure above ①-> ②-> ③, ③-> ②-> ①, you can also maintain the status ②.

Through InfoCore's transparent path migration technology, the storage virtualization engine can be deployed to the user's original environment without downtime, avoiding the impact on the user environment, meeting users' requirements for no-downtime deployment, and realizing functions such as failover and load balancing through MPIO software. it can also realize automatic tiering, cache acceleration, high storage availability, CDP, disaster recovery backup and other functions through the storage virtualization engine. Improve the security and storage performance of the user environment.

The above is the editor for you to share how to achieve the real linux non-downtime data migration, if you happen to have similar doubts, you might as well refer to the above analysis to understand. If you want to know more about it, you are welcome to follow the industry information channel.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.