GlassFish Server Open Source Edition 3.1 (build 43)
Oracle 11g
I've recently started looking at Glassfish connection pooling.
Our Oracle installation has a limit of 900 connections and was getting filled during usage spikes causing it to be unavailable to users.
My understand of glassfish connection pooling was that a maximum pool size could be set, which would limit the number of Oracle connections being used.
Under testing, I cannot understand the link between the connection pool settings in Glassfish, and the number of connection processes exhibited at the Oracle end. With a connection pool max of 16, under testing, the number of processes seems to vary between 18 and 31.
I am looking up Oracle processes with the following SQL:
select a.username, a.status, a.schemaname, a.osuser, a.machine, a.program, b.SPID AS os_processid
from v$session A
JOIN V$PROCESS B
ON B.ADDR = A.PADDR
where A.username = '[DB USERNAME]'
ORDER BY SPID;
Attached are Glassfish pool settings for this test.
If someone could reassure me that I'm going about this right, and maybe explain why the Oracle processes would be higher than the pool size (and suggest how much higher it will get) that would be great.
Attachment | Size |
---|---|
gfish_settings.jpg | 522.77 KB |