On 23/02/11 00:26, Greg Stark wrote:
>
> It's also possible there's a bug of course. If someone was using that
> buffer and somehow failed to notify the vacuum that they were done it
> would wait for a very long time (forever?). However if vacuum
> eventually continued when the query was canceled then it seems likely
> it was working as intended.
>
Greg, thanks for clarifying this.
Unfortunately this time around I canceled the vacuum and then the query.
However *next* time I'll get rid of the query 1st and see what happens.
Cheers
Mark