Although easy to implement, round robin DNS has problematic drawbacks, such as those arising from record caching in the DNS hierarchy itself, as well as client-side address caching and reuse, the combination of which can be difficult to manage. Round robin DNS should not solely be relied upon for service availability. If a service at one of the addresses in the list fails, the DNS will continue to hand out that address and clients will still attempt to reach the inoperable service.
Also, it may not be the best choice for load balancing on its own since it merely alternates the order of the address records each time a name server is queried. There is no consideration for transaction time, server load, network congestion, etc. Round robin DNS load balancing works best for services with a large number of uniformly distributed connections to servers of equivalent capacity. Otherwise it just does load distribution.
Methods exist to overcome such limitations. For example, modified DNS servers can routinely poll mirrored servers for availability and load factor. If a server does not reply as required, the server can be temporarily removed from the DNS pool, until it reports that it is once again operating within specs.