Re: Strange Oracle setup

In article <3512E126.68030FEF_at_ibm.net> semmett_at_ibm.net "Steve Emmett" writes:

>
>
> Take some time to design a more reasonable layout, a backup and recovery
> scheme,
> a fallback plan and then plan on a night (or weekend) to implement.
>
> is the instance a 24x7 or are you in luck with something less critical?
>

The system is 24x7 and is _very_ business critical. Fortunately there are actually 3 identical nodes in the system and it is possible to take one down for a couple of hours or so.

I have checked the backup scripts and they seem sensible bearing in mind that the database is not the sum total of information that has to be recovered. However I have yet to find a description of the recovery procedure, or evidence that the full recovery procedure has ever been tested. Testing it with the current hardware configuration is not practical, and my requests for a test system are meeting a 'when we get the money' response.

I am also unhappy with the system performance, DB_BLOCK_BUFFERS clearly being far too small for starters at 550 , but my few attempts at tuning have so far been met by accusations of 'fiddling' by the vendor of the package. The good point is that the system workload is quite small, and CPU utilisations of around 10% are the norm without excessive response times. This is despite what I would consider to be blatantly excessive use of full table scans (about 11,000 in 24 hours).

Keith Matthews Received on Sun Mar 22 1998 - 00:00:00 CST