Hi,

At June-07-11, 4:35 AM, Maya Opperman wrote:

> OK, GSTAT results (while running slowly this morning)

> Database header page information:
> Flags 0
> Checksum 12345
> Generation 839074
> Page size 8192
> ODS version 11.1
> Oldest transaction 838791
> Oldest active 838792
> Oldest snapshot 838792
> Next transaction 838793
> Bumped transaction 1
> Sequence number 0
> Next attachment ID 273
> Implementation ID 16
> Shadow count 0
> Page buffers 3000
> Next header page 0
> Database dialect 3
> Creation date Jun 2, 2011 17:23:14
> Attributes force write, no reserve
>
> Variable header data:
> Sweep interval: 20000
> *END*

> The OAT and OIT seem to be fine. I have asked them to change sweep interval from 20 000 to 120 000.

> Any idea for what to look at in the monitoring tables exactly?

I don't think that there's anything to look at in the monitoring
tables.

The problem is the size of the page buffers. The page buffers is set
too high, for classic. Change the page buffers, ideally for classic,
anywhere between 75 and 150 pages.

What happens is that every connection(for your scenario), which is a
separate instance of the Firebird engine, can allocate up to 3,000
pages for each connection. So, you can imagine that the available
memory can be consumed with just a few connections, and then you would
be talking about memory disk paging with Windows.

You can change the page buffers size with the Firebird's command line
GFIX.

--
Best regards,
Daniel Rail
Senior Software Developer
ACCRA Solutions Inc. (www.accra.ca)
ACCRA Med Software Inc. (www.filopto.com)