I had another interesting recovery operation a few weeks ago. We had a customer that was getting an error message when trying to open the database. The error was ‘file needs more recovery’. I think we have all seen that one before.
However, this was a slightly different type of problem. The customers database was not in archive log mode, and they were not and never had taken hot backups. They were taking only weekly cold backups.
The customer had very little Oracle experience, but they were paying for Oracle support. So they opened a Service Request with Oracle. In the SR they clearly stated that their production database was down. And they stated it more than once.
Of course, they also stated that the database was not in archive log mode. Perhaps Oracle didnt think they were serious? Even though production was down, Oracle did not make this a severity 1 issue, which I cannot understand. It met every requirement for a Severity 1 issue. In addition, Oracle support hemmed and hawed, first asking for substantial amounts of irrelevant information, including an RDA, etc. Over the course of a week, with the production database down, Oracle support neither raised this to a severity 1 issue, nor offered a single suggestion on how to get the database working. Evidently as far as Oracle was concerned, if the database was not in archive log mode, they did not have to find a solution.
After trying to work with Oracle support for a week, the customer called us in. I had an idea of a solution to the customer’s problem almost immediately of course, but to be safe, I first had the customer copy all of the Oracle database files in place, just in case I needed to take it back to the starting point.
Then I ran the following commands:
Sqlplus /nolog
Connect / as sysdba
Startup mount
Alter database open
(at this point I received an error that the undo tablespace needed more recovery)
So I ran this command:
Recover database;
Then I did:
Alter database open;
And the problem was resolved.
Oracle support evidently did not know that a recover command could be run on an instance that was not in archive log mode, or Oracle did not consider this customer a priority. In either case, this reflects extremely poorly on Oracle support.
December 16, 2008 at 2:37 pm |
That is horrible, Oracle tech support ignoring a production down system.
They were lucky to have you.