Recovery question?!

All:

Let me describe my dilemma. An Oracle 7.1.16 DB began getting ORA-00376 (file can't be read at this time [/u05/oradata/dataJIT_03]) errors two days ago. When I found out about it, I looked at Oracle and in the v$datafiles table it showed the file status as AVAILABLE. I tried to alter the associated tablespace OFFLINE NORMAL to bring the file off-line and then on-line. I got an ORA-01191 (file already off-line - can't do a normal off-line). Then we altered the tablespace OFFLINE IMMEDIATE. The tablespace wouldn't come back on-line, and we got an ORA-01113 error (file needs media recovery [/u04/oradata/dataJIT_01.dbf]).

Instead of applying the archivelog files, I shut down the db using SHUTDOWN NORMAL, and did an image backup of the DB. Just for laughs, we tried to start the DB again, and we got another ORA-00376 (file can't be read at this time [/u05/oradata/dataJIT_02.dbf]).

Here's where we screwed up. We didn't shut down the DB while we were restoring from Monday night's image BU. That totally stepped on the DB files - including the control files. Someone realized the DB was up (duh), and tried to shut it down. It wouldn't shutdown until I executed a shutdown abort. When, we finished restoring the Monday night's BU, we started having control file version differences (ORA-00376 (how did that happen?)). When we finally got that ironed out by copying increasingly newer versions of one control file on top of the others, we got an ORA-01113 (system tablespace needs media recovery [/u01/oradata/sysJIT_01.dbf]).

So, in the last 2 days, we've gotten messages that 3 different devices need media recovery. Did Oracle "blow its cookies," or do we have an obscure hardware (controller?) problem.

Also, should we just restore the Wednesday morning BU, and apply the archive logs after having HP run diagnostics on the machine? Received on Thu Apr 01 1999 - 08:54:05 CST