Hi Marcos,

Sweep interval = 20000 (default) Sweeping the db does not seem to make
any difference. Memory is not getting freed.

Helen, Martin ... The indexes do not seem to be the problem ( at least
that is my opinion ) coz the degrade is gradual over a period of say 5
hours or so.

Am posting the top d 10 statistics on the linux box below.

2:01pm up 0:31, 2 users, load average: 0.10, 0.18, 0.09
63 processes: 60 sleeping, 3 running, 0 zombie, 0 stopped
CPU states: 0.0% user, 0.2% system, 0.0% nice, 99.8% idle
*Mem: 1031084K av, 664624K used, 366460K free*, 0K shrd,
13036K buff
Swap: 1044216K av, 0K used, 1044216K free 564820K
cached


6:09pm up 4:39, 2 users, load average: 0.10, 0.18, 0.09
63 processes: 58 sleeping, 5 running, 0 zombie, 0 stopped
CPU states: 0.0% user, 0.3% system, 0.0% nice, 99.6% idle
*Mem: 1031084K av, 894624K used, 136460K free*, 0K shrd,
13036K buff
Swap: 1044216K av, 0K used, 1044216K free 564820K
cached

You can see what has happenned to the memory in this time frame of
roughly 4 hours. I have shut down almost every other task on the server.
Only FB is running and each query (Close ... Open) is taking up memory
on the server. Why does FB take up the physical memory ?

The problem has started as the db size crossed some sort of barrier.
Sweeping does not release the memory. If all users log out ... memory
does not get deallocated. In fact if I restart the fb server it will
come back to the earlier level of 664 MB. I have only one db running on
this machine.

Is there some sort of optimal limit to the db size in fb. Without
touching the indexes a query takes 1 second at startup and moves to 17
secs in a few hours. Should the query not run slow even at startup if
the index is indeed a problem.

The db stats that I posted were taken at a slow moment ! Meaning the
query was coming down to 5 secs when I posted them. Its now at 14 secs.

My isc_config file has all parameters commented except for the temp
folder location and temp folder size set to 2GB.

Now all users have logged out and fb is still running and the top d 10
out is :

6:26pm up 4:57, 2 users, load average: 0.31, 0.34, 0.23
56 processes: 54 sleeping, 2 running, 0 zombie, 0 stopped
CPU states: 0.0% user, 0.2% system, 0.0% nice, 99.7% idle
*Mem: 1031084K av, 910176K used, 120908K free*, 0K shrd,
16388K buff
Swap: 1044216K av, 0K used, 1044216K free 561136K
cached

If I restart firebird on server now they become :

6:28pm up 4:59, 2 users, load average: 0.04, 0.22, 0.19
55 processes: 53 sleeping, 2 running, 0 zombie, 0 stopped
CPU states: 8.6% user, 0.8% system, 0.0% nice, 90.5% idle
*Mem: 1031084K av, 637968K used, 393116K free*, 0K shrd,
16764K buff
Swap: 1044216K av, 0K used, 1044216K free 560804K
cached

Clearly firebird is unable to free the memory for some reason ? What can
I do to improve this ?

Warm regards,


RP

Marcos Vinicius Dufloth wrote:

>And what about sweep interval?
>
>