Re: Recovery Strategies

In article <331741AB.48F2@netcomuk.co.uk>, Neil Truby <ntruby@netcomuk.co.uk> wrote: >Suspending disbelief of the assumption, I believe that, in all server >failures, the combination of the disk and Informix Fast Recovery will >always guarantee the requirement without recourse to back-ups/logs (To >be precise, all users will be in their last end-transaction state prior >to the failure). Yes, Neil, provided you are using UNBUFFERED logging. With buffered logging, there's no guarantee that a committed trx. has been written to the logical log; in which case, it obviously cannot be recovered. >I further believe that, in circumstances where it's necessary to restore >and roll forward logical logs, the requirement can only be satisfied if >the log dbspace is available. In the case where the log dbspace had >been trashed, or where for some reason the assumption was false, it >would only be possible to restore the state at the time that the last >logical log back-up took place. >Do I have this right? Yes, which is why it makes sense to keep individual logs small and have a continuous backup running. ---------------------- Rudy Fernandes GIC, Kuwait ----------------------