Hi Stefan,
That makes perfect sense. The SID should have been capital, but when I first
started playing with my script I was passing it a lowercase SID. I wasn't
aware Oracle would just go ahead and start a new instance because of that.
Thank you for the clarification.
Stephen
From: Stefan Knecht <knecht.stefan@xxxxxxxxx>
To: dmarc-noreply@xxxxxxxxxxxxx
Cc: Oracle-L Freelists <oracle-l@xxxxxxxxxxxxx>
Sent: Tuesday, May 19, 2015 9:19 AM
Subject: Re: Archivelog mode and receiving ORA-19602
There's not too many options here. I'm assuming you're on Linux, where Oracle
identifies a database instance with two things: the ORACLE_HOME and the
ORACLE_SID.
The ORACLE_HOME being where the database software is to be found, and the
ORACLE_SID being the name you want to give your instance.
If they both match (and it's a 100% case sensitive text match that's required)
then you'll connect to your running instance. If they don't Oracle will assume
you're aiming to start a new instance.
For example:
ORACLE_HOME=/u00/home_1ORACLE_SID=d1
Is NOT the same like
ORACLE_HOME=/u00/home_1/ORACLE_SID=D1
Even though, from an OS perspective, both ORACLE_HOMEs point to the same
directory, and they're both valid ways to reference that directory /u00/home_1,
for Oracle it's two different ORACLE_HOMEs. d1 and D1 equally are two different
ORACLE_SID values. Therefore technically you could start many different
instances by mixing these two variables in their various forms.
Hope that helps to clarify what may have happened
Stefan
On Tue, May 19, 2015 at 8:08 PM, stephen van linge
<dmarc-noreply@xxxxxxxxxxxxx> wrote:
Hi Pavan, You are correct, it turns out I have 2 db instances running with the
same sid hence the confusion. My restore script was passing the service name
to connect to the database where I was using OS authentication through
sqlplus/rman, so my restore script got the instance that the listener wasn't
blocking, and I got the instance that the listener was blocking. The one I was
connecting to was in archivelog mode and the one the script was connecting to
was in noarchivelog mode. I'm just going to nuke the VM and start over,
something funky must have happened during dbca (I swear I only ran it once...).
Might have been some leftover files after deleting the last instance (with the
same name). Thank you for your time, Stephen
From: pavan kumar <pavan_843@xxxxxxxxxxx>
To: "dmarc-noreply@xxxxxxxxxxxxx" <dmarc-noreply@xxxxxxxxxxxxx>; Oracle-L
Freelists <oracle-l@xxxxxxxxxxxxx>
Sent: Monday, May 18, 2015 9:08 PM
Subject: Re: Archivelog mode and receiving ORA-19602
Hi Stephen.
Are you sure that your database is in archive log mode.Can you re-confirm once
? to be on safeReason for asking : Since rman internal gets information from
the database dictionary only, whether it's archive is enabled or not, it won't
pickup from any where else.
- ThanksPavan Kumar N
From: stephen van linge <dmarc-noreply@xxxxxxxxxxxxx>
To: Oracle-L Freelists <oracle-l@xxxxxxxxxxxxx>
Sent: Tuesday, 19 May 2015 5:59 AM
Subject: Archivelog mode and receiving ORA-19602
Hi Oracle-L, I'm working in a playground database at the moment. It's a single
instance 11gR2 (11.2.0.3) with no replication and currently running in
archivelog mode, which I can verify with archive log list and by switching
logfiles and having an archivelog created. The database was newly created.
Now, when I attempt to run a simple rman copy incremental: backup incremental
level 1 for recover of copy with tag 'daily_bkp' database plus archivelog; I
receive the errors: channel ORA_DISK_1: starting datafile copy
RMAN-03009: failure of backup command on ORA_DISK_1 channel at 05/19/2015
00:17:22
ORA-19602: cannot backup or copy active file in NOARCHIVELOG mode
continuing other job steps, job failed will not be re-run Which is clearly
incorrect... There are so many... repetitive... questions on Google asking why
when they're in noarchivelog mode they get this error, that I can't find
anything relevant to my case. Has anyone experienced this before? Thanks,
Stephen
