A while ago, I wrote a note about disaster recovery i.e. how to quickly restore the services after a failure in the primary server. However, as the old adage goes Prevention is better than cure, here are some tips on how you can prevent system failures. This applies to all our customers including those who may have deployed an email disaster recovery system.
The two key characteristics of the email system that make it vulnerable to failure are:
- A multilayer system – failure can occur at any layer of the system leading to a disruption of the mail services. Therefore, ensuring the health of each layer of the system would help keep the failure rates to the minimum.
- The Email system is exposed to the external world on the Internet. As a result intrusions and security breaches are an ever present danger. Securing the system adequately is one of the most important steps in ensuring availability.
We recommend the following to prevent failures at the different layers of the system.
- Install equipment in recommended temperature and dust control environments, with limited and authorized access for more reliability.
- Ensure uninterrupted power supply of the required rating for the equipment’s.
- Use of standard/branded servers, routers, switches, designed for continuous operations with sufficient built in redundancy like – dual power supply, dual CPU, dual NIC cards, RAID for the storage, SCSI or better disks, etc. Resist from using PCs, and assembled hardware which offer poor capacity and reliability.
- Provision for hardware with sufficient headroom – Ensure that all the equipment is not used beyond 60-70% of its capacity. This is required partly to accommodate the organic growth in load and partly to offset the load from a failed component during a switchover.
- Use security appliances for Anti-SPAM, Anti-Virus, IDS(intrusion detection) for additional protection if permissible within the budget.
- Deploy the recommended Operating system: Subscribe to the vender support network to ensure automatic download of patches and hot fixes.
- To prevent unexpected bursts in load due to virus infestations of the Outlook clients on the PCs, ensure every client PC/laptop in the network is secured by a good Internet protection tool.
- Use backup software for large mail stores.
Application Software Layer
- Deploy and secure Mithi Connect Xf in one of the recommended standard configurations.
- To prevent unexpected software conflicts, consult Mithi Customer Care before installing any new software or performing any major reconfiguration on the server.
Process Layer – Follow preventive maintenance procedures
- Ensure your administrators are well trained to handle the complete mailing system. We recommend periodic refresher training for all the administrators. Mithi has regular training programs at its office in Pune. (Please check with Mithi Customer care to set-up a training date)
- Ensure your servers are covered under warranty/maintenance. In addition, if necessary maintain some spares for a reduced downtime. Subscribe to the same business day on site support services of the vendor.
- Document and follow strict preventive maintenance procedures for the entire setup. Contact your hardware vendor to keep the platform infrastructure running smoothly. Consult Mithi for a document on the standard maintenance procedures for the software.
- Use a change management process to initiate configuration modifications on the server.
- Ensure regular periodic back-up and archival.
Above all, have a bias towards Maintaining rather than Repairing.
Quiet but persistent and firm, she assists Sunil to manage the Services division.