"Checked for relevance on 29-Jun-2007"
Problem Description:
====================
Your database performance is degrading and eventually you crash.
You look in the PMON trace file and find the following errors.
error 470 detected in background process
00447, 00000, "fatal error in background process"
// *Cause: One of the background processes died unexpectedly.
// *Action: Warm start the system.
ora-470
00470, 00000, "LGWR process terminated with error"
// *Cause: The log writer process died
// *Action: Warm start instance
DBWR and SMON indicate the same thing.
Example:
=========
The above has been an on going problem for about two weeks.
The standard procedure for this particular company is to shutdown the system
every Monday and do a full backup every Sunday with incrementals everynight.
Customer running Oracle applic. with about 5 users on when DB dies.
You are receiving the following errors in PMON, LGWR and SMON trace files:
error 470 detected in background process
OPIRIP: Uncaught error 447. Error stack:
ORA-00447: fatal error in background process
ORA-00470: LGWR process terminated with error
No errors are being reported in the alert.log. You did receive the following
alert.log entry: "ORACLE Instance - Can not allocate log, archival
required"
You are forced to do the following to get the database back up:
1. SHUTDOWN ABORT
2. STARTUP NORMAL
3. SHUTDOWN IMMEDIATE
4. STARTUP NORMAL
Solution Description:
=====================
You need to verify and possibly increase your MAXFILES setting.
Solution Explanation:
=====================
In example explored in the problem section, the user had recently added new
datafiles to his instance.
On checking the MAXFILES kernal parameter, he discovered that it was set to
60. Checking V$datafiles showed 47 data files. Computing the total of data
files, control files, mirror redo logs, init.ora and alert.log files, total
exceeded the 60 file limit.