Sample Outsourced Managed Cluster (24/7)
A single technical contact person is designated to your account from the initial solution set up to solution operation. The solution itself is 100% redundant (i.e., failure proof). We also take care of system maintenance and surveillance (24/7) and deliver an SLA of 99.9% (guaranteed) over the entire chain (system, network, IT management) with a Guaranteed Response of 20 minutes to 4 hours (depending on the contract you opt for).
Description of servers/services comprising the infrastructure
The LVS servers are used to distribute the load over the 2 Web servers.
The servers are configured for failing-over; therefore, even if one of the two server fails, this will not lead to any disruptions.
The WEB servers will host your Internet sites, applications, extranet, databases, etc..
The load is distributed between these two front-end servers.
Moreover, a query-prioritization rule is defined to distribute the load between the Web servers.
The HTTP service can tolerate the failure of one or several servers subject to the remaining capacity being able to handle all the queries (on a related note, this is why we recommend installing at least n+1 servers, where ‘n’ is the minimum number of servers required).
The system used to ensure a tolerance to failures is based on server redundancy (rsync reputation) as well as on RAID 1+ back-ups on one of the servers.
With this configuration, the MySQL service is hosted on the Web servers. The MySQL servers replicate each other, and are configured as Master/Slave in order to avoid any problems related to query indexing. When using the MySQL Master/Slave, no particular restrictions apply, and the Slave can automatically take on the role of Master.
Magic Online recommends sharing the use of the front-end server with the data base in order to have a more stable and, especially, more upgradeable, solution.
The servers are used to update the data on the Web 1 and 2 servers. Files are uploaded via the FTP service.
Several servers (or indeed all servers) have a second hard disk; this can be used, by each machine, to save its own data, and to have its data saved by another server.