Does your online business depend on cloud providers, databases, distant servers, or saved information of some type?
Of course it does.
Are you in fixed concern of downtime?
Of course you’re.
In our 24-hour, always-on, all the time related, someone-is-always-awake market(s), uptime is important. Simply put, going darkish is dangerous for enterprise… and it’s unacceptable. That’s why excessive availability engineering options—like ones supplied by Liquid Web, and different trendy webhosting corporations—are so necessary. From excessive availability infrastructure and SQL databases, to redundant replication, transaction logs, and the elimination of single factors of failure, an internet host is simply nearly as good as its excessive availability (HA) engineering providers.
Let’s assume your utility can be hosted on a conventional managed infrastructure.
Now, let’s have a look at why a excessive availability server is a greater answer.
High Availability Summarized
When it involves HA, the three ideas of reliability engineering have to be thought of:
- Reduce or get rid of single factors of failure.
- In redundant methods, be certain that crossover factors are dependable.
- The system in place should detect and react to failures in real-time.
When these three ideas are reliably applied, a big discount in downtime is achieved. A top quality net host can have these ideas in thoughts when designing their providers.
Reducing single factors of failure in an HA system means redundancy in information—digital, bodily, or a mix of the 2. An HA construction can have a major quantity, and no less than one bodily backup quantity. A regular configuration is comprised of two, similar, major volumes backed up by two, similar, Distributed Replicated Block Device (DR:BD) bodily volumes, backed up by two DR:BD digital volumes. DR:BD volumes carry out selective, synchronous information replication, which implies that solely blocks of modified information (not the whole quantity) are rewritten and backed up in real-time.
DR:BD volumes in the end cut back backup instances as they require much less computing assets at one time. Each backup tier (two similar major volumes, two similar DR:BD digital volumes, and many others.) is saved on separate bodily servers—some hosts will even again as much as a server at a distant location. A configuration with a distant location eliminates one other single level of failure by defending your information from pure disasters and different location-based points like energy outages and community failures.
What To Do With the Database
In an HA system, it’s advisable that your SQL database is saved on a separate, redundant server surroundings because it improves efficiency and reduces overhead in your major server. A devoted SQL server additionally works with/towards the ideas of reliability engineering as it’s particularly designed for prime availability, together with automated and dependable crossovers and real-time failure detection.
SQL databases additionally create incremental transaction logs; one other guard towards single factors of failure. Transaction logs report each change within the database at set intervals as ceaselessly as one minute—the SQL database can use the transaction logs as a dataset, writing to the backup servers in your HA configuration.
Liquid Web’s commonplace configuration for SQL database internet hosting features a day by day backup of the whole database and 24 rollings hours of hourly transaction logs.
Monitoring for Failover
At the very core of the HA configuration must be a monitoring system that’s continuously and persistently maintaining a tally of the well being of the clustered servers and routinely performing failovers when vital. The monitoring subsystem hottest throughout the trade is Heartbeat. Heartbeat is a Linux-based monitor that may reliably assist a number of nodes. Heartbeat can rapidly and precisely establish important failures and routinely transition the system to a redundant server.
As you possibly can see, every half of the HA system works with/towards multiple of the three ideas of excessive availability engineering. The redundant information nodes (bodily and digital) cut back single factors of failure and create dependable crossover factors.
A devoted SQL server creates one other layer of redundancy, one other guard towards single factors of failure, and has built-in, automated crossover factors.
Finally, Heartbeat sits on the heart of the whole configuration, monitoring the system in real-time and automating crossovers when vital.
With a high quality HA system in place, downtime is lowered or just about eradicated, retaining your online business on and operational all day, day by day.
