AIX storage high availability and disaster recovery solution implementation

Introduction to Basic technology

Introduction to AIX LVM Mirror Local storage high availability solutions

Logical Volume Manager (LVM) is the software for logical volume management on AIX. The LVM itself provides Logical Volume (LV) data mirroring between multiple physical Volume (PV) to achieve local high availability of storage. Write I/O in the LVM Mirror scenario and the underlying device interaction are shown in the following illustration.

Figure 1. LVM Mirror Scenario Architecture

When the server issues write I/O, the I/O is sent concurrently to two storage devices in Parallel mode. As shown in step 1, Step 2 and Step 1 ', Step 2 '. Only when step 2 and step 2 are complete will a write I/O be considered complete by the server.

In this solution, when any of the underlying storage, PV, fails, the LVM automatically stops all I/O activity on the PV and switches the corresponding reads and writes to the remaining normal PV. The entire switching process is done at the LVM level, and all LV states remain online during failover and are transparent to the application layer. The minimum switching time is about 15 seconds, or even shorter.

Introduction of the disaster-tolerant solution in the same city

The DS8000 Metro Mirror (MM) technology is a storage-level disaster-tolerant solution that meets the Recovery point Object (RPO) of 0. In this scheme, the data is replicated to the disaster-tolerant site synchronously by storing the underlying I/O to achieve the corresponding disaster-tolerant purpose. The DS8000 storage interaction between write I/O and different sites in the DS8000 MM scenario is shown in the following illustration.

Figure 2. DS8000 MM Scheme Architecture

When the server issues write I/O, first, the write I/O is first placed on the production site primary storage device, then the write I/O is sent from the production site primary storage to the disaster preparedness site for storage, the storage is sent back to write I/O completion information, and the final primary store sends write I/O completion information to the server. Throughout the process, each write I/O is guaranteed to be synchronized between production and disaster-tolerant sites.

In this solution, when a disaster occurs in a production site, a site switching scenario is triggered, that is, all equipment in the production site is down, and the disaster preparedness Site server and storage are enabled and serviced at the disaster-tolerant end. Since the data stored by the disaster-tolerant side is produced by synchronous replication, data is consistent with production, ensuring an RPO of 0. It is worth noting when DS8000 MM the scheme itself does not provide a high availability guarantee for local storage of production sites. When local storage fails and cannot be recovered, production will eventually need to be switched to a disaster-tolerant site. The whole process is usually for a few hours or higher.

AIX LVM Mirror Combined DS8000 Metro Mirror Solution Introduction

The AIX LVM Mirror combined with the DS8000 Metro Mirror solution is a storage high availability plus disaster solution that combines the benefits of both scenarios. The basic architectural diagram of the solution is shown in the following illustration.

Figure 3. AIX LVM Mirror combined with DS8000 Metro Mirror architecture