Re: How to protect DBF tables?
It's all a matter of design and planning. You can not just grab some backup software from the shelve and have a go with it and expect to be safe. These things need well thought out planning and choice of hard- en software if you want to do this professionally. Backups are not suited for instant recovery. Look at backups for situations of total damage. For instant recovery, use snap-shot technology, and preferably with dedicated back-end snapshot hard- and software. Bottom line is: there is a WORLD of solutions out there to protect your data, it is just a matter of budget on the one side, and the importance of your data (price of data loss) at the other side.
We used to make a snapshot from the application (whilst running live in ultra high traffic mode) every hour with a separate snapshot server, and at night we would run a daily tape backup which was then brought outside the premises every morning first thing to a safe vault in an other risk-profile geographic location to be stored. This provided enough security on top of the fail-safe 3-node cluster with heartbeat connection where it ran on.
Later, we even improved on things dramatically by running it on VmWare, on a NetApp metro cluster. That however took a huge investment. The solution of those questions always has to be tailored towards specific needs, risks, requirements and budgets. All of that information is not present.
It's all a matter of design and planning. You can not just grab some backup software from the shelve and have a go with it and expect to be safe. These things need well thought out planning and choice of hard- en software if you want to do this professionally. Backups are not suited for instant recovery. Look at backups for situations of total damage. For instant recovery, use snap-shot technology, and preferably with dedicated back-end snapshot hard- and software. Bottom line is: there is a WORLD of solutions out there to protect your data, it is just a matter of budget on the one side, and the importance of your data (price of data loss) at the other side.
We used to make a snapshot from the application (whilst running live in ultra high traffic mode) every hour with a separate snapshot server, and at night we would run a daily tape backup which was then brought outside the premises every morning first thing to a safe vault in an other risk-profile geographic location to be stored. This provided enough security on top of the fail-safe 3-node cluster with heartbeat connection where it ran on.
Later, we even improved on things dramatically by running it on VmWare, on a NetApp metro cluster. That however took a huge investment. The solution of those questions always has to be tailored towards specific needs, risks, requirements and budgets. All of that information is not present.
Comment