Optimizing Git for handling database backups: findings and insights from another 6,000 hours of experimentation

International Conference on Information Technology and Digital Applications 2018

计算机科学;无线电电子学

Roig, J.V.^1

Advanced Research and Consulting, Asia Pacific College, Philippines^1

关键词: Back-up systems;    Disaster recovery;    Distributed sources;    Space efficient;    Storage spaces;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/482/1/012037/pdf
DOI  :  10.1088/1757-899X/482/1/012037

学科分类：计算机科学（综合）

来源: IOP

【 摘 要 】

This is the second in a series of papers detailing results from on-going investigations of using Git- A distributed source control system-as a database backup and disaster recovery tool. The first paper, published early 2017 after 2,000 hours of data processing and simulations, warns that the default configuration of Git is unsuitable for database backups, but also shares configuration tweaks that make it safe and more space-efficient than zipped.sql archives-as low as only 1% of the storage space needed by.sql archives. This second paper focuses on optimization, looking specifically to identify which configuration tweaks and factors affect the reliability, scalability and efficiency of a Git-based database backup and disaster recovery system. Critical insights regarding the reliability and scalability of a Git-based database backup system were found after 6,000 hours of experimentation, centered around the problem of configuring Git to be able to handle.sql files measured in gigabytes.