Let me start with a simple question to business owners. “Have you ever quantified losses (financially and operationally) your business would suffer if your ERP software suddenly goes down?” Almost every time I asked this question the response is that they have never considered that their system can hit a glitch and become inaccessible. Back and restore function is usually never discussed client. However, we always bring up this topic in order to assure our clients that our ERP has a bulletproof backup and restore capability. And if ever such a problem arises our ERP backup and restore mechanism effectively brings system to live in about an hour max with the latest business data.
Our preferred ERP installation method is on a virtual private server (VPS). A VPS brings in loads of benefits versus in-house server. Each instance of client server has an image backup made automatically every week. This is fine for business with few transactions but would not work for large enterprises who are using ERP for every task.
For large enterprises we recommend having a 2nd server (preferably with same VPS company) at a different geographical location. Example: if the 1st (production) VPS is in Frankfurt then the back-up server should be anywhere else than Frankfurt. We run various scripts (cron jobs) to automatically backup all the data on an hourly basis to 2nd server during operational hours on business days. This usually creates about 9 to 10 backups per day depending on business hours of the client. Five minutes before midnight (client local time) a final backup is created. This backup ensures any work done during non-working hours is captured. As soon as a backup is created in production server it is automatically copied onto the backup server.
The backup server
Normally the backup server is of similar specification as the primary server. If it is of lower specification (due to cost savings measure) then it can be quickly expanded to be of similar specification. Usually it takes about 10 minutes to manually increase the specification.
Incoming backup files are kept in a dedicated folder of the backup server. Scripts are used to automatically copy the latest backup file onto a dedicated folder. ‘If’ the production server faces an issue and becomes inaccessible the latest backup file on 2nd server is used to restore company data. The static IP of primary server is replaced with static IP of the backup server. Once fired up the backup server is now the primary server. Users log into the system with same URL and have the exactly similar look and feel of the primary server.
Estimated loss of data is of less than an hour if issue has occurred during working hours. This surely is much better than loosing information of many days or in some cases complete historical data.
Whilst the method is not as high grade as auto-switch. But it definitely gets the job done at about 5% of the cost of auto-switch. The down time is less than 30 minutes to maximum an hour from the time we are informed about the issue by the client.
ERP data and access to the system are very critical to a business. It has become all the more important after introduction of Bahrain Valued Added Tax. Client should have extensive discussion with their service providers to understand how business operations are least affected. Paying for a robust backup and restore mechanism has high return on investment. It is like any type of insurance, you need it when you need it. You have no control on ‘when you need it’. Else you can always be without it which surely cannot be the case.