A high-availability method and device based on multi-master dns architecture

Abstract

An embodiment of the present invention provides a high-availability method and device based on a multi-master DNS architecture, wherein the method includes: each check node detects each master DNS at a set frequency, and collects the detection results of the check node and the check The heartbeat information of nodes and each decision node stores its heartbeat information in a distributed database; each decision node determines that the number of check nodes and decision nodes is not less than the set number threshold based on the heartbeat information of check nodes and decision nodes , according to each of the detection results, it is judged whether there is a primary DNS entering the switching process; if there is a decision node that determines that the primary DNS is entering the switching process, then all decision nodes that determine that the primary DNS is entering the switching process pass the exclusive lock lock. Perform primary DNS switchover. According to this solution, under the multi-master DNS architecture, based on distributed multi-point survival detection, data storage, distributed multi-point decision-making, and then quickly switch according to the rules to achieve high availability of multi-master DNS.

Description

technical field [0001] The present invention relates to the field of data processing, in particular to a high-availability method and device based on a multi-master DNS architecture. Background technique [0002] DNS (Domain Name System) is a domain name resolution system that provides a method for mutual conversion between domain names and IPs. It not only provides services for Internet users externally, but also provides traffic switching portals for administrators internally. Now users respond to The time and availability requirements are getting higher and higher, and higher requirements are put forward as the underlying basic components. [0003] DNS is a scenario that reads more and writes less. In most cases, it is used for domain name resolution. However, when system administrators switch traffic or change domain names, the traditional methods are as follows: [0004] 1. Master-slave architecture: One master node provides the write function, multiple slave nodes pro...

AI Technical Summary

Problems solved by technology

[0005] 2. Same-level structure: all DNS are at the same level, without distinguishing between master and slave roles. Once there is a domain name change record, rely on SSH or other piping services to make all DNS services use the new local configuration file and roll restart to take effect

[0008] 1) If the master node fails, the DNS resolution records cannot be changed, but can only be read, which will affect the overall situation

[0009] 2) The change of the master node requires configuration changes and restarts of all nodes, which is a high risk

[0011] 1) It is unavailable during the short rolling restart process. Even if the upper layer architecture has the function of detecting survival and removing it, there is a time delay, so that reading is also affected due to a few important change requirements

[0012] 2), because all DNS needs to be restarted to make the configuration take effect, so the domain name takes a long time to take effect

Images