Chapter 4. Understanding Archives

One of the main duties of the database administrator is ensuring the safety of the data. Although each of the Informix engines is a very robust product, a number of factors can cause the system to go down. In OnLine systems, these problems often arise from user errors such as killing engine processes with a UNIX kill -9 command. Other things that can crash the system are hardware problems and UNIX problems that crash the entire computer. IDS's multithreaded architecture and its strict prohibition against killing oninit jobs make it less likely that a UNIX kill command will crash your database, but there are many other things that can crash your system.

In most instances, the database will recover from the error during the fast recovery process that occurs automatically when the shared memory is first started. In most cases, user errors and software problems cause no physical damage to the data, and this fast recovery is adequate to bring the system back up. It is the hardware crashes that most often cause the database administrator to have to recover the database from the archives. When this happens, the DBA quickly realizes that the only thing between him and the deep blue sea is a flimsy length of recording tape. It behooves the DBA to be absolutely certain of the Informix archive and recovery system. This chapter will explain the archive and recovery system and point out some of the ways that the DBA can assure the integrity of this system.