We have carried out a hardware refresh and have consolidated to purely HP servers from our previous mixed hardware. This includes all our database and application servers and virtualisation hosts.
The previous infrastructure contained numerous brands with slightly different configurations and spare parts which would lead to difficulties in the case of a hardware failure. We could not focus on a single brand for spare parts on site, and in some cases research was required before a problem could be fixed.
The hardware refresh had numerous benefits. It allowed us to easily stock spare parts on site and concentrate our expertise on a single brand allowing us to be agile in our response to failures. It also allowed us to build a relationship with a supplier who now keeps spare parts and servers in stock so even if the onsite spares fail we can rapidly gain access to replacements.
To replace our previous database hosting server we adopted the latest HP DL360 Generation 7 servers with hard disks in RAID 10 configuration, all using redundant power supplies. These servers use dual Xeon processors and have plenty of RAM leading to a considerable performance increase when accessing your Clear Books account and database.
To ensure this hardware continues to perform at its high levels and to avoid errors occurring we carry out regular visual inspections of the servers looking for error lights, and have expanded the use of our Zabbix network monitoring tool.
Using a detailed network map with Zabbix monitoring we visualise all our network nodes to rapidly locate errors in the case of a failure. This allows us to respond to problems before they affect services for Clear Books customers. This monitoring is integrated with SMS alerts from the host-tracker.com monitoring system for all elements of our infrastructure, from coolant pressure to dropped packets.