Re: Two instances with same SID in a same machine

From: Hans Forbrich <forbrich_at_yahoo.net>
Date: Sun, 29 Aug 2004 02:12:01 GMT
Message-ID: <RXaYc.71105$X12.53526@edtnps84>

Igor Laletin wrote:

> Hans Forbrich <forbrich_at_yahoo.net> wrote in message
> news:<5GFXc.19626$A8.2621_at_edtnps89>...

>> Igor Laletin wrote:
>> 
>> > What if two people (listeners) called two different Marcelo's
>> > (instances with the same SID) in two different rooms (oracle homes)?
>> > 
>> 
>> Your attempt to twist this {home=room} analogy is not quite appropriate
>> to the situation ... in Oracle's case {server=room} and {home=chair}

>
> Not a twist. The analogy is quite appropriate to the listeners. For
> them oracle home is a room. Everything listener needs is (or could be)
> in OH.
>
> It's actually quite appropriate to the database engine itself. You can
> keep almost everything in separate rooms for them. Some things like
> /tmp/.oracle seems to be hardcoded but again I never tried to change
> TMP/TMP_DIR. There is a chance even this could be kept separate.
>
> In short, on Unix (server != room && OH == room:)
>
> Regards,
> Igor

Now you have me wondering ... especially since one LISTENER, regardless of which home it comes from, can service all the instances on the machine, regardless of which homes they belong to.

So perhaps you could help with this test I've got here .... one userid, two oracle homes, two listeners.

Both instances are up, both listeners are up (one on port 1521 the other on 1531)

Oracle9i R2, 9.0.4
SuSE Linux Enterprise 8
user=oracle
ORACLE_HOME=/opt/oracle/product/ora904
ORACLE_HOME=/opt/oracle/product/ora904b
both SID=PTS1

From my remote Oracle Networking client's TNSNAMES.ora,

PTS1A looks at port 1521
PTS1B looks for the service on port 1531

Both apparently ONLY talk to the second instance - any changes I make in the instance are identical regardless of whether I connect to PTS1A or PTS1B.

What tests, or configuration files, would you suggest? Since you sound so positive, I'm sure I just misconfigured something ...

/Hans

oracle_at_fcc02:~> ps -ef | grep pts

oracle   26267     1  0 13:48 ?        00:00:00 ora_pmon_pts1
oracle   26269     1  0 13:48 ?        00:00:00 ora_dbw0_pts1
oracle   26271     1  0 13:48 ?        00:00:00 ora_lgwr_pts1
oracle   26273     1  0 13:48 ?        00:00:05 ora_ckpt_pts1
oracle   26275     1  0 13:48 ?        00:00:00 ora_smon_pts1
oracle   26277     1  0 13:48 ?        00:00:00 ora_reco_pts1
oracle   26279     1  0 13:48 ?        00:00:00 ora_cjq0_pts1
oracle   26281     1  0 13:48 ?        00:00:06 ora_qmn0_pts1
oracle   26283     1  0 13:48 ?        00:00:00 ora_s000_pts1
oracle   26285     1  0 13:48 ?        00:00:00 ora_d000_pts1
oracle    8999     1  0 14:43 ?        00:00:00 ora_pmon_pts1
oracle    9004     1  0 14:43 ?        00:00:00 ora_dbw0_pts1
oracle    9006     1  0 14:43 ?        00:00:00 ora_lgwr_pts1
oracle    9008     1  0 14:43 ?        00:00:03 ora_ckpt_pts1
oracle    9010     1  0 14:43 ?        00:00:00 ora_smon_pts1
oracle    9012     1  0 14:43 ?        00:00:00 ora_reco_pts1
oracle    9014     1  0 14:43 ?        00:00:00 ora_cjq0_pts1
oracle    9016     1  0 14:43 ?        00:00:05 ora_qmn0_pts1
oracle    9018     1  0 14:43 ?        00:00:00 ora_s000_pts1
oracle    9020     1  0 14:43 ?        00:00:00 ora_d000_pts1
oracle   13580     1  0 19:44 pts/6    00:00:00
/opt/oracle/product/ora904/bin/tnslsnr LISTENER -inherit
oracle   14135     1  0 19:48 pts/7    00:00:00

/opt/oracle/product/ora904b/bin/tnslsnr LISTENER -inherit oracle 14730 12881 0 19:53 pts/7 00:00:00 grep IS oracle_at_fcc02:~> Received on Sat Aug 28 2004 - 21:12:01 CDT