Datasheets

Improve Uptime with a High Availability Database

Issue link: https://hub.liquidweb.com/i/1018553

Contents of this Issue

Navigation

Page 1 of 1

www.LiquidWeb.com 800.580.4985 Technical Overview How is data replicated? High Availability Database provides instant data replication between two redundant nodes, an active server and a hot spare. It uses proven, enterprise-grade tools including Distributed Replicated Block Device (DRBD) software and Heartbeat. With DRBD, whenever data is written to disk, the changed blocks are immediately replicated across a dedicated network link to the other node. The replication is synchronous. If your database writes to disk successfully, then the data will have been replicated successfully as well. How does it work? Only one node is active at a time. The inactive node will be waiting with a copy of your data in hot standby, ready to take over in less than a minute* in case of a catastrophic hardware failure on the active node. The cluster also can be failed over manually by our Heroic SupportĀ® technicians. This means that upgrades, hardware maintenance, or any software maintenance that requires a reboot can be performed on one server at a time and your databases will stay available throughout the entire maintenance. Compared to painful, hour-long maintenance windows for standalone dedicated servers, the benefit becomes very clear. What happens if a node fails? The two database servers are in constant communication via Heartbeat, performing health checks over redundant links to rapidly detect catastrophic hardware failure. In addition to Heartbeat, we also proactively monitor both nodes 24/7/365. Anytime our team detects an issue, they will immediately begin troubleshooting and contact you right away. How Does This Affect My Application? The short answer it doesn't. As far as your application is concerned, your High Availability Database is no different than any other standalone database server. Thanks to floating public and private IP addresses which fail over with your databases, your application will only ever need to connect to a single IP, regardless of which server is active at any given time. When uptime is your No. 1 priority, a High Availability Database is your best solution. Liquid Web offers two core classes of this product to satisfy your infrastructure needs: BUSINESS CLASS Standard Hard Drives Single Power Supply 1G Bonded Connection for DRDB Benchmarked Queries Per Second**: 8,500 BUSINESS CLASS Standard Hard Drives Single Power Supply 1G Bonded Connection for DRDB Benchmarked Queries Per Second**: 8,500 * Time estimate refers to both failure detection and failover of services. Service failover time may be affected by settings in your database's configuration such as innodb_buffer_pool_dump. ** Benchmarks were run using a read/write multi-table OLTP test with Sysbench v0.5 and represent an average of results from tests using thread counts up to 1,500. High Availability Database

Articles in this issue

Links on this page

view archives of Datasheets - Improve Uptime with a High Availability Database