Database Health Alerts (SAP Library

Database Health Alerts

You might see the alerts described here when monitoring your Oracle database. You can see which database objects are causing an alert by looking at the contents of the message container in the corresponding node.

Note

"[BRCONNECT]" below means that the alert is raised by the BRCONNECT database system check.

"[NON-STANDARD]" below means that the alert has an equivalent alert in another monitoring tree. If required, you can turn off the equivalent alert in the other monitoring tree and only show the alert here.

For more information, see Information on Oracle Database Alerts.

Database health check

Last successful database check

Cause: The threshold for the number of days since the last successful database check, which includes BRCONNECT return codes of 0 (successful) or 1 (successful but with warnings), has been exceeded.

Action: Run brconnect -f check.

Last brconnect -f check run

Cause: The last run of brconnect -f check has an invalid status.

Action: Correct the problem and repeat brconnect -f check.

Last evaluation of check results

Cause: Using information stored in the DBMSGORA table by the last run of brconnect -f check, the collection method of this node feeds all others marked with “[BRCONNECT].”

Action: You can start the collection method of this node manually if you want to immediately evaluate the last database check run.

Start of database check

This lets you directly start the database check with brconnect -f check.

Database administration

The alerts in this section correspond to the checks stored in the DBCHECKORA table for database administration.

Archiver destination full [BRCONNECT] [NON-STANDARD]

Cause: The used space in the archiver destination directory is above the preconfigured threshold. For more information, see ARCHIVER_STUCK in BRCONNECT Default Conditions for Database Administration.

Action: Either back up and delete offline redo logs using BRARCHIVE or add more space to the archiver destination to accommodate more logs.

The standard equivalent to this alert is Archiver destination full in Database Backup/Restore Alerts.

Control file not mirrored [BRCONNECT]

Cause: The control file is not mirrored. For more information, see CONTROL_FILE_MIRRORin BRCONNECT Default Conditions for Database Administration.

Action: Edit the init<DBSID>.oraprofile and add the additional control files under the parameter CONTROL_FILES. We recommend you have at least three copies of the file. Each copy of the file must be stored on a separate physical hard disk. After you edit the profile, restart the database to activate your changes.

Note

If you only use hardware mirroring, you can inactivate this check. For more information, see Configuring Database System Check (Oracle).

Missing control file [BRCONNECT]

Cause: The control file is missing. For more information, see CONTROL_FILE_MISSINGin BRCONNECT Default Conditions for Database Administration.

Action: Look for the parameter CONTROL_FILES, in the init<DBSID>.ora initialization file to determine which directories the control files are stored in. Make sure you can display all the files and that they are available in the file system. Restore access to any non-accessible files.

Critical autoextendable file [BRCONNECT]

Cause: There are data files with an activated auto extend feature that would consume the entire remaining free disk space if extended. A check is made to see whether the file system can be brought to overflow, due to the existing parameter setting (NEXT and MAXSIZE) during the automatic file extension. For more information, see CRITICAL_FILE in BRCONNECT Default Conditions for Database Administration.

Action: Add more space to the file system or reduce the autoextendparameters.

Critical segment [BRCONNECT] [NON-STANDARD]

Cause: There are segments with too few allocatable extents due to lack of free space in the tablespace. "‘Too few" is defined in the CHKVAL field of the DBCHECKORA table (range is 1 to 5).

For more information, see CRITICAL_SEGMENT in BRCONNECT Default Conditions for Database Administration.

Action: If the NEXT storage parameter for this segment is set too high, decrease it to allow the segment to allocate more extents. If NEXT is set as intended, add another data file to the tablespace.

The standard equivalent to this alert is Segments with too few allocatable extents in Space Management Alerts.

Data file mismatch [BRCONNECT]

Cause: There are data files that are flagged as MISSING in the Oracle dictionary.

For more information, see DATA_FILE_MISMATCH in BRCONNECT Default Conditions for Database Administration.

Action: Execute CREATE CONTROLFILE using Oracle SQLPLUS, specifying either the correct file or no file at all.

Missing data file [BRCONNECT]

Cause: There are data files that no longer exist in the file system.

For more information, see DATA_FILE_MISSING in BRCONNECT Default Conditions for Database Administration.

Action: Perform Complete Database Recovery with BR*Tools.

Database file offline [BRCONNECT]

Cause: There are data files or online redo log files that are offline.

For more information, see FILE_OFFLINE in BRCONNECT Default Conditions for Database Administration.

Action: Close all applications and enter the following commands using Oracle SQLPLUS:

...

1. connect / as sysdba

2. shutdown immediate

3. startup mount

4. alter database data file <data file> online;

5. alter database open

Alternatively, use Complete Database Recovery with BR*Tools.

File system full [BRCONNECT]

Cause: The fill level of file systems on the database host is too high. All file systems are checked by default against the same threshold value. However, you can define different threshold values for individual file systems by specifying the file system in the OBJECT field of the DBCHECKORA table (transaction DB17).

For more information, see FILE_SYSTEM_FULL in BRCONNECT Default Conditions for Database Administration.

Action: Increase the amount of space in the file system.

Harmful statistics [BRCONNECT] [NON-STANDARD]

Cause: there are tables or indexes that have statistics, although they should not have these (for example, pool and cluster tables).

For more information, see HARMFUL_STATISTICS in BRCONNECT Default Conditions for Database Administration.

Action: Remove statistics from these tables or indexes. For more information, see brconnect -f stats.

The standard equivalent to this alert is Harmful statistics in Space Management Alerts.

Invalid file type [BRCONNECT]

Cause: There are database files that have an illegal operating system type, for example, block raw files on UNIX or compressed files on Windows.

For more information, see INVALID_FILE_TYPE in BRCONNECT Default Conditions for Database Administration.

Action: Troubleshoot to determine the cause of the problem. If you are not sure how to proceed, contact SAP technical support.

Segment in wrong tablespace [BRCONNECT] [NON-STANDARD]

Cause: A segment is in the wrong tablespace, as raised by the BRCONNECT database check. For more information, see IN_WRONG_TABLESPACE in BRCONNECT Default Conditions for Database Administration.

Action: Use Reorganizing Tables with BR*Tools, which moves the tables into a tablespace of the correct type.

The standard equivalent to this alert is Segment in wrong tablespace in Space Management Alerts.

Missing index [BRCONNECT] [NON-STANDARD]

Cause: There are tables that do not have any indexes and are not specified in the DBDIFF table exception.

For more information, see MISSING_INDEX in BRCONNECT Default Conditions for Database Administration.

Action: Create the missing indexes using transaction SE11.

The standard equivalents to this alert in Database Consistency Alerts are:

● Objects missing in database: Primary indexes

● Objects missing in database: Secondary indexes

● Unknown objects in ABAP dictionary: Tables without unique index

Missing statistics [BRCONNECT] [NON-STANDARD]

Cause: There are tables or indexes that do not have any statistics, although they should have these.

For more information, see MISSING_STATISTICS in BRCONNECT Default Conditions for Database Administration.

Action: Create statistics for these tables or indexes. For more information, see brconnect -f stats.

The standard equivalent to this alert is Missing Statistics in Database Performance Alerts.

Database in NOARCHIVELOG mode [BRCONNECT] [NON-STANDARD]

Cause: The database is running, but online redo logs are not being archived.

For more information, see NOARCHIVELOG_MODE in BRCONNECT Default Conditions for Database Administration.

Action: Using BRSPACE, restart the database and change to archivelog mode.

The standard equivalent to this alert is Archiving Off in Database Backup/Restore Alerts.

Segment with non-zero PCTINCREASE [BRCONNECT] [NON-STANDARD]

Cause: Segments have a PCTINCREASE value other than zero. Since Oracle now allows a MAX_EXTENTS setting of "unlimited", setting PCTINCREASE to something other than zero no longer makes sense. Setting PCTINCREASE to non-zero can potentially introduce unnecessary fragmentation into your tablespace. Only segments in the SAP schema are considered.

For more information, see PCTINCREASE_NOT_ZERO in BRCONNECT Default Conditions for Database Administration.

Action: Evaluate whether the segment’s storage parameter can be set to zero.

The standard equivalent to this alert is Segments with non-zero PCTINCREASE in Space Management Alerts.

Redolog file not mirrored [BRCONNECT]

Cause: The online redo logs are not mirrored. Mirroring is required for safe operation. For more information, see REDOLOG_FILE_MIRROR in BRCONNECT Default Conditions for Database Administration.

Action: Using Oracle SQLPLUS, activate mirroring of the online redo logs in your database, with the statement ALTER DATABASE ADD LOGFILE MEMBER.

Missing redolog file [BRCONNECT]

Cause: An online redo log is missing, which reduces the effect of mirroring. Mirroring is required for safe operation. For more information, see REDOLOG_FILE_MISSINGin BRCONNECT Default Conditions for Database Administration.

Action: First, check the file that cannot be accessed, and check the system for problems (network problems, server down, and so on). One possible action (only for Oracle experts, as an error could result in an incomplete chain of offline redo logs, which might prevent a full recovery of the database): If a second member of the group that cannot be accessed exists, delete the non-accessible member. Using Oracle SQLPLUS, execute:

ALTER SYSTEM SWITCH LOGFILE

Then, check if the status of the online redo logs in the table V$logfile is correct, and execute:

ALTER DATABASE ADD LOGFILE MEMBER

Tablespace full [BRCONNECT] [NON-STANDARD]

Cause: A tablespace is full. All tablespaces are checked by default against the same threshold value. However, you can define different threshold values for individual tablespaces by specifying the tablespace name in the OBJECT field of the DBCHECKORA table, using transaction DB17.

For more information, see TABLESPACE_FULL in BRCONNECT Default Conditions for Database Administration.

Action: Extend the tablespace, if necessary, based on your estimate of anticipated growth.

Tablespace in backup mode [BRCONNECT] [NON-STANDARD]

Cause: A tablespace is in backup mode, which reduces the performance of the SAP System. For more information, see TABLESPACE_IN_BACKUP in BRCONNECT Default Conditions for Database Administration.

Action: Confirm that no database backup is running, either with BRBACKUP or an external tool. Then cancel the backup mode by using BRSPACE to reset the backup status.

The standard equivalent to this alert is Tablespace in backup mode in Database Backup/Restore Alerts.

Tablespace offline [BRCONNECT]

Cause: A tablespace is offline. For more information, see TABLESPACE_OFFLINEin BRCONNECT Default Conditions for Database Administration.

Action: Use BRSPACE to set the tablespace online.

Segment with too many extents [BRCONNECT] [NON-STANDARD]

Cause: There are tables or indexes, for which the number of allocated extents exceeds the specified threshold value. All tables or indices are checked by default against the same threshold value. However, you can define different threshold values for individual tablespaces by specifying the tablespace name in the OBJECT field of the DBCHECKORA table (transaction DB17).

For more information, see TOO_MANY_EXTENTS in BRCONNECT Default Conditions for Database Administration.

Action: Increase the MAXEXTENTS limit for the table or index using SQLPLUS with one of the following commands:

ALTER TABLE <name> STORAGE (MAXEXTENTS <number>);

ALTER INDEX <name> STORAGE (MAXEXTENTS <number>);

If you prefer segments with fewer allocated extents, you can use BRSPACE to reorganize the object.

Alternatively, use the –l|limit option of brconnect –f next.

The standard equivalent to this alert is Segments approaching MAX_EXTENTS in Space Management Alerts.

DBA Operations

The alerts in this section correspond to the checks stored in the DBCHECKORA table for database operations.

Last successful archive log backup too old [BRCONNECT] [NON-STANDARD]

Cause: The last successful backup of the offline redo log files with BRARCHIVE is too old. For more information, see ARCHIVE_TOO_OLD in BRCONNECT Default Conditions for Database Operations.

Action: Perform an offline redo log file backup.

The standard equivalent to this alert is Last successful archive log backup in Database Backup/Restore Alerts.

Last successful complete database backup too old [BRCONNECT] [NON-STANDARD]

Cause: The last successful complete backup of the database with BRBACKUP is too old. For more information, see BACKUP_TOO_OLD in BRCONNECT Default Conditions for Database Operations.

Action: Perform a complete database backup.

The standard equivalent to this alert is Last successful complete database backup in Database Backup/Restore Alerts.

Last archive log backup failed [BRCONNECT] [NON-STANDARD]

Cause: The last backup of the offline redo logs with BRARCHIVE has an invalid status. For more information, see LAST_ARCHIVE_FAILED in BRCONNECT Default Conditions for Database Operations.

Action: Correct the problem and repeat the BRARCHIVE run. For more information, see Command Options for BRARCHIVE.

The standard equivalent to this alert is Last brarchive run in Database Backup/Restore Alerts.

Last complete database backup failed [BRCONNECT] [NON-STANDARD]

Cause: The last complete database backup with BRBACKUP has an invalid status. For more information, see LAST_BACKUP_FAILED in BRCONNECT Default Conditions for Database Operations.

Action: Correct the problem and repeat the BRBACKUP run. For more information, see Command Options for BRBACKUP.

The standard equivalent to this alert is Last brbackup run in Database Backup/Restore Alerts.

Last operation failed [BRCONNECT]

Cause: The last DBA operation failed, which is identified by a function ID specified in the OBJECT field of the DBCHECKORA table. For more information, see LAST_OPERATION_FAILED in BRCONNECT Default Conditions for Database Operations.

Action: Correct the problem and repeat the operation. For more information, see the corresponding function in BR*Tools for Oracle Database Administration.

Last update statistics failed [BRCONNECT] [NON-STANDARD]

Cause: The last update of the optimizer statistics with BRCONNECT failed. For more information, see LAST_STATS_FAILED in BRCONNECT Default Conditions for Database Operations.

Action: Correct the problem and repeat brconnect -f stats. For more information, see Command Options for BRCONNECT.

The standard equivalent to this alert is Last brconnect -f stats run in Database Performance Alerts.

Last successful operation too old [BRCONNECT]

Cause: The last successful DBA operation, identified by a function ID specified in the OBJECT field of the DBCHECKORA table, is too old.. For more information, see OPERATION_TOO_OLD in BRCONNECT Default Conditions for Database Operations.

Action: Correct the problem and repeat the operation. For more information, see the corresponding function in the BR*Tools for Oracle Database Administration.

Last successful update statistics too old [BRCONNECT] [NON_STANDARD]

Cause: The last successful update statistics run is too old. For more information, see STATS_TOO_OLD in BRCONNECT Default Conditions for Database Operations.

Action: Correct the problem and repeat brconnect –f stats. For more information, see Command Options for BRCONNECT.

The standard equivalent to this alert is Last successful update statistics in Database Performance Alerts.

Oracle Messages

The alerts below correspond to the DBCHECKORA check conditions for database messages in the Oracle alert log. You can find the complete messages with time stamp and text in the message container of the corresponding Monitoring Tree Element (MTE) in the alert monitor.