Jan, you should consider setting the sweep interval to 0 (no auto
sweeps) and set up a routine manual sweep for a time when the db is
not being used (usually overnight or first thing in the morning).
This will ensure users do not suffer a performance loss when a sweep
kicks in after the default 20,000 old transactions.

Regards,
Andrew Ferguson

--- In ib-support@y..., Jan Agermose <post@a...> wrote:
> sometimes my database workload goes through the roof. I think it
could
> be that the server is doing a sweep. No log - no knowledge, or is
this
> actually logged somewhere? Any ways, I tried looking at:
>
> Database header page information:
> Flags 0
> Checksum 12345
> Generation 19793574
> Page size 8192
> ODS version 10.0
> Oldest transaction 19783145
> Oldest active 19783146
> Oldest snapshot 19783144
> Next transaction 19793565
> Bumped transaction 1
> Sequence number 0
> Next attachment ID 0
> Implementation ID 19
> Shadow count 0
> Page buffers 0
> Next header page 0
> Database dialect 3
> Creation date Dec 9, 2001 10:51:57
> Attributes
>
> Variable header data:
> Sweep interval: 20000
> *END*
>
>
> ''''
> now my database pool size is 30 connections and connections are set
to
> autoCommit(true) (JDBC). What I do not understand is:
>
> Oldest transaction 19783145
> Oldest active 19783146
> Oldest snapshot 19783144
> Next transaction 19793565
>
> "Next transaction" keep getting bigger and "Oldest XX" stays the
same.
> Shouldn't the "oldest" values increase? And shouldn't the
difference be
> something in the neighborhood of 30 (the pool size) again since the
> queries should be auto committed?
>
> Might be that my understanding of the above figures should be
corrected,
> please do.
>
> Jan Agermose
>
>
>
> [Non-text portions of this message have been removed]