Hi,

at our biggest client, the database tends to get slower with the time
(half a day), and issuing a restart to the Firebird server fixes it.
For example, one routine takes 2 minutes before the restart. After the
restart, it takes only 10 seconds. No sweep is issued during the process.

Also, issuing a sweep with the database up doesn't change
OIT/OAT/Snapshot values.

Any ideas what may be causing it?

OS Info:
--------
Fedora Core release 3 (Heidelberg)

Machine info:
-------------
processor : 1
vendor_id : GenuineIntel
model name : Intel(R) Pentium(R) 4 CPU 3.00GHz
cpu MHz : 3002.625
cache size : 1024 KB

Memory info:
------------
MemTotal: 1553684 kB
MemFree: 7896 kB
Buffers: 66904 kB
Cached: 1082816 kB
SwapCached: 0 kB
Active: 802252 kB
Inactive: 698540 kB
HighTotal: 0 kB
HighFree: 0 kB
LowTotal: 1553684 kB
LowFree: 7896 kB
SwapTotal: 1048568 kB
SwapFree: 1048424 kB
Dirty: 1636 kB
Writeback: 0 kB
Mapped: 355992 kB
Slab: 30492 kB
Committed_AS: 876276 kB
PageTables: 2820 kB
VmallocTotal: 2564088 kB
VmallocUsed: 5588 kB
VmallocChunk: 2552028 kB
HugePages_Total: 0
HugePages_Free: 0
Hugepagesize: 2048 kB

Gstat output:
-------------

Database "/manager/gdb/EcomaxUBR2000.GDB"

Database header page information:
Flags 0
Checksum 12345
Generation 494291
Page size 4096
ODS version 10.1
Oldest transaction 339206
Oldest active 339207
Oldest snapshot 336985
Next transaction 494283
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 1
Creation date Oct 24, 2005 21:02:33
Attributes force write

Variable header data:
Sweep interval: 20000
Continuation file: /manager/gdb/EcomaxUBR20001.GDB
Last logical page: 250000
*END*


Regards
Rodrigo