Availability Analysis with Failover Computer Cluster Method Case Study in Academic Information System of UIN Sunan Kalijaga

Abstract

High-performance academic information systems and high availability services are requirement in every university. One of many reasons is for anticipation damage and fail server disrupting server network performance. Failover computer cluster method is applied to two servers: primary server as main server and secondary server as backup server. Four stages will be carried out: First: Installation and support software configuration. Second: installation and failover cluster configuration. Third: installation and Distributed Replicated Block Devices (DRBD) configuration. Forth: server testing with siege and nettool. This research conducted by doing server test before and after high availability. As if the main server has a system failure, it will automatically backup the main server to backup server to minimize user accessed data failures. System uses Ubuntu 16.04 LTS operating system. Based on the test data, it is acquired two data: packet data and data response time (ms). Data packets acquired from this research are: 233.3 average data sent; 228.3 average data received; 2.3 average data lost; while 59.7 average response time (ms) is showed; 2.7 minimum average; 633.8 maximum average. Data sent is 120B per data.

Analisis performansi high availability cluster server menggunakan heartbeat pada private cloud

“…Data yang dikirim adalah 120 B per data. Berdasarkan hasil pengujian, sistem dapat melayani transaksi dengan mengirimkan 1000 paket data berdasarkan lama waktu 5 menit untuk setiap 10 kali percobaan dan menghasilkan 99% paket data yang diterima dan 1% paket yang gagal [5]. Penelitian [6] menganalisis failover cluster untuk pemulihan sistem.…”