Message-Id: <10604.115905@fatcity.com> From: prashasta gujrati Date: Thu, 31 Aug 2000 00:40:28 +0530 Subject: Database Crash...advice needed... --------------34283D2EBE6C59DC701AC059 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi DBAs, This is the scenario : Oracle 7.3.3 on AIX 4.3.2 Time Event 0300 Offline backup of the database done successfully. 1523 Hard disk failure on one of the disks(hdisk9) mounted on /database. This FS had 1 system DBF, 2 RBS files, 3 application DBFs. These files were taken offline automatically as reported in alert.log . 1530 Error reported of applications not working and DB having crashed. 1535 Database was shutdown with abort and tried doing a normal startup which failed. Database could only be mounted. It has been brought down since then. 1600 Another FS was created on a different disk (hdisk10) and was mounted on the same mount point (/database). 1630 We restored the DBFs from the offline backup which were residing on that mount point (/database). 1640 Database startup mount. After that we issued a recover database which prompted for the archive files which were avilable on another disk and were applied. The Redo is also intact. What we see here is that it takes us to a point between 1523 and 1535. Now when we try to open the database it says that one of the RBS files cannot be read. We tried commenting out the rollback_segments from init.ora and then did startup mount. Thereafter, we made the RBS files offline to see if the database will reopen. It fails with an error message : ORA-00704: bootstrap process failure ORA-00604: error occurred at recursive SQL level 1 ORA-00955: name is already used by an existing object After all this now we have planned to restore all the files from offline backup and use the current controlfile + redo + archives to do a PITR upto 1515 (few minutes before the disk failure) and then open the DB with resetlogs option and then take an offline backup immediately. If you all can tell me the error we made in this process, it'll help. Also your advice if the second approach we are taking will work or not ? TIA, --------------34283D2EBE6C59DC701AC059 Content-Type: text/html; charset=us-ascii Content-Transfer-Encoding: 7bit  
Hi DBAs,

This is the scenario : Oracle 7.3.3 on AIX 4.3.2

Time         Event

0300          Offline backup of the database done successfully.
1523          Hard disk failure on one of the disks(hdisk9) mounted
                 on /database. This FS had  1 system DBF, 2 RBS files,
                 3 application DBFs.

                 These files were taken offline automatically as reported
                 in alert.log .

1530          Error reported of applications not working and DB having
                 crashed.

1535          Database was shutdown with abort and tried doing a
                 normal startup which failed. Database could only be
                 mounted. It has been brought down since then.

1600         Another FS was created on a different disk (hdisk10)
                and was mounted on the same mount point (/database).

1630         We restored the DBFs from the offline backup which
                were residing on that mount point (/database).

1640         Database startup mount. After that we issued a recover
                database which prompted for the archive files which
                were avilable on another disk and were applied. The Redo
                is also intact.

                What we see here is that it takes us to a point between 1523
                and 1535.

                Now when we try to open the database it says that one of
                the RBS files cannot be read. We tried commenting out the
                rollback_segments from init.ora and then did startup mount.
                Thereafter, we made the RBS files offline to see if the
                database will reopen. It fails with an error message :
                  ORA-00704: bootstrap process failure
                  ORA-00604: error occurred at recursive SQL level 1
                  ORA-00955: name is already used by an existing object

After all this now we have planned to restore all the files from offline
backup and use the current controlfile + redo + archives to do a PITR
upto 1515 (few minutes before the disk failure) and then open the DB
with resetlogs option and then take an offline backup immediately.

If you all can tell me the error we made in this process, it'll help.
Also your advice if the second approach we are taking will work