RE: Recovery with logs, then incremental, then more logs?

Brandon,  

What you say makes sense to me, logically. But, I've never been through that sequence of events w/ RMAN, and I'm not sure if/when RMAN "evaluates" or "re-evaluates" what incrementals/archive logs are available to it.  

Now, what are you doing reading and writing email?? Go home! Get some sleep! ;-)  

-Mark

-- 
Mark J. Bobak 
Senior Oracle Architect 
ProQuest Information & Learning 

For a successful technology, reality must take precedence over public
relations, for Nature cannot be fooled.  --Richard P. Feynman, 1918-1988

 

________________________________

From: oracle-l-bounce_at_freelists.org
[mailto:oracle-l-bounce_at_freelists.org] On Behalf Of Allen, Brandon
Sent: Monday, April 17, 2006 12:21 PM
To: oracle-l_at_freelists.org
Subject: Recovery with logs, then incremental, then more logs?


Hello all, me again -
 
Hopefully you all had a better weekend then I did - I worked about 20
hours recovering a crashed production database due to running out of
disk space.  Anyway, my question is regarding the unusual recovery
process that Oracle performed - I'm curious if any of you have ever seen
a recovery like this before:
 
1) Oracle 9.2.0.6 on AIX 5.2
2) L0 incremental backup taken on 4/7
3) L0 incremental backup taken on 4/8, but tape was pulled and shipped
to DR site
4) L1 incrementals taken every night 4/9 - 4/14
5) Database crashes on 4/15 due to full filesystem, database opens, but
returns ORA-607 and ORA-600 [4193] on any DML or DDL statements
6) Restore started, but fails because it can't access L0 backup from 4/8
since the tape has been sent off site (no copy retained - duh!)
7) L0 backup from 4/8 changed to unavailable in rman
8) Restore restarted and all datafiles from 4/7 L0 backup are restored
successfully
9) Recovery begins restoring & applying 750 logs from all week long
10) I abort the recovery because the restored logs are going to fill up
the filesystem again (poor planning on my fault) and I don't feel like
manually moving them off every 2 hours since I am dreadfully in need of
some sleep by this point
11) I restart the recovery after changing the archive log desination to
a filesystem with plenty of space
12) At this point, rman skips 400 archive logs and instead applies the
incremental backup from 4/14, then continues to apply the remaining logs
up until the morning of 6:30 (which is where I "set until time" to),
then it says recovery is complete
13) I open the database with resetlogs and everything seems peachy so
far
 
It seems to me that if I hadn't aborted the recovery at step 10, rman
would have continued applying all ~750 logs, but since I aborted it and
then restarted, when it reevaluated the position of the database (SCN in
datafiles vs. backups, etc.) it saw that it could now apply the
incremental L1 even though it's base L0 backup (from 4/8) was never
restored, due to the application of redo past the time that it's base L0
backup was taken. Does this sound correct?  I'm just trying to make sure
I understand exactly what happened.  Anyone else ever seen a recovery
like this?
 
Thanks in advance for any feedback and taking the time to read all this!
 
Regards,
Brandon
 
Privileged/Confidential Information may be contained in this message or
attachments hereto. Please advise immediately if you or your employer do
not consent to Internet email for messages of this kind. Opinions,
conclusions and other information in this message that do not relate to
the official business of this company shall be understood as neither
given nor endorsed by it.


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