Re: Database writing architecture

In article <4kkgis$1o7_at_news.computek.net>, scottv_at_computek.net (Van Scott) says:
>
>In article <316D6BD4.41C67EA6_at_exu.ericsson.se>, Gerald Ward <exuward_at_exu.ericsson.se> says:
>>
>When a Unix box has crashed with a SQL Server on it,
>typically there are problems with the databases. Relevent
>factors here include the amount and timing of physical writes.
>For example, if a checkpoint of a large database happened
>right before the power failure and the log information
>corresponding to that activity is not written to the log
>device, database recovery will probably be problematic.

Ammend: the above applies to a database on a Unix file system. There should be less problems, if any, in this case with a database on a raw partition since there is no buffering.

Van Received on Fri Apr 12 1996 - 00:00:00 CEST