RE: v$session with same serial#

(disclaimer: I fully reserve the right to have read this completely wrong and am now going down a rabbit trail...)

Doesn't your example actually show the opposite? Sure, the background procs are 1 (but I think we can agree that those are "special" processes), but the others seem to be fairly different.

Here's mine:

set lines 180
set pages 24
set head on
set verify on
set feed on

1.first check how many sessions in each node

select inst_id, count(*) from gv$session where type != 'BACKGROUND'
group by inst_id
/

   INST_ID COUNT(*)
---------- ----------

         1         87
         2         67
         3         63

3 rows selected.

2. Now check how many repeated serial#s

select inst_id, serial#, count(*)
from gv$session
where type != 'BACKGROUND'
group by inst_id, serial#
having count(*) > 1
order by count(*) desc
/

   INST_ID SERIAL# COUNT(*)
---------- ---------- ----------

         3 10 2

Regards,

Chris

-----Original Message-----
From: oracle-l-bounce_at_freelists.org [mailto:oracle-l-bounce_at_freelists.org] On Behalf Of Mark W. Farnham Sent: Thursday, January 10, 2013 4:03 AM To: ralberto.hernandez.ext_at_juntadeandalucia.es Cc: oracle-l_at_freelists.org
Subject: RE: v$session with same serial#

As Iggy and I have both told you, this is quite normal. For example:

SQL> select count(*),serial# from v$session group by serial#;

  COUNT(*) SERIAL#
---------- ----------

        15          1
         1         28
         1        394
         1        133
         4          3

In fact all the oracle background processes usually have the same serial# apart from ones that get started and stopped without the instance being restarted. I'm actually wondering a bit about your other instances and what would make them NOT have duplicate serial# values consistently.

mwf

--
http://www.freelists.org/webpage/oracle-l