problem with checkpoint of databases

[Lawrence Greenfield]

--On Monday, December 02, 2002 10:11 PM +0100 Eric.Doutreleau at int-evry.fr 
wrote:
[...]
> it takes around one hour per day of failure of checkpointing to restart.
>
> -Is there a way to speedup that process?

Recovery speed is relative to how often checkpoints are done. Since your 
checkpoints aren't finishing, it isn't working...

> -will the changing the format of the database  help to solve that
> problem?

Well, non-Berkeley db backends don't have these sort of checkpoints.

> -what could cause the checkpoint to fail?

Some internal Berkeley db being held. This might be something to 
investigate with Sleepycat or at least search their website.

Larry