Data cannot be lost. You need to have a reliable backup system. And this does not mean that it is enough to choose a good program, it is necessary to have the correct copy policy for each company and to verify that the data is recovered in a time that is considered adequate for each of the possible scenarios of data loss.
Client: a bank
Need: Design and operate a backup system for the cluster that was going to carry out the bank's operations.
Previous situation: the previous system depended on an IBM mainframe in the data center of the parent bank, which carried its own system of copies.
Project conditions: should interfere as little as possible with the normal operation of the bank.
System description: A Sun cluster with an Oracle database in Archivelog mode.
Solution chosen: Veritas Netbackup, now from Symantec, with a tape robot and an agent for Oracle.
Implementation: The copy policy was designed in collaboration with bank managers to define copy frequency, retention period, and recovery requirements. This established a full copy on weekends and incremental copies every night in duplicate, one of which was sent to the central bank facility every day. Data recovery tests were performed with time measurement, and of the entire system. During the project the cluster had to be migrated to another location and the recovery system was tested again. Throughout the project I worked in collaboration with a bank staff member so that he could continue to operate the system after it was finished.
Cost of the project: 100 man days + Netbackup license with Oracle agent + tape robot.
No comments:
Post a Comment