Chapter 17 Actions when an Error Occurs

This chapter describes the actions to take when an error occurs in the database or an application, while Fujitsu Enterprise Postgres is operating.

Depending on the type of error, it may be necessary to recover the database cluster. The recovery process recovers the following resources:

Data storage destination
Transaction log storage destination (if the transaction log is stored in a separate disk from the data storage destination)
Backup data storage destination

Note

Even if a disk is not defective, the same input-output error messages, as those generated when the disk is defective, may be output. The recovery actions differ for these error messages.

Check the status of the disk, and select one of the following actions:

If the disk is defective
Refer to "17.1 Recovering from Disk Failure (Hardware)", and take actions accordingly.
If the disk is not defective
Refer to "17.14 I/O Errors Other than Disk Failure", and take actions accordingly.
A few examples of errors generated even if the disk is not defective include:
- Network error with an external disk
- Errors caused by power failure or mounting issues

Determining the cause of an error

If an error occurs, refer to the WebAdmin message and the server log, and determine the cause of the error.

See

Refer to "Configuring Parameters" in the Installation and Setup Guide for Server for information on server logs.

Approximate recovery time

The formulas for deriving the approximate recovery time of resources in each directory are given below.

If using the copy command with the pgx_rcvall command, the recovery time will depend on the implementation of the copy command.

Data storage destination or transaction log storage destination
```
Recovery time = (usageByTheDataStorageDestination + usageByTheTransactionLogStorageDestination) / diskWritePerformance x 1.5
```
- usageByTheDataStorageDestination: Disk space used by the database cluster
- usageByTheTransactionLogStorageDestination: Disk space used by the transaction log stored outside the database cluster
- diskWritePerformance: Measured maximum data volume (bytes/second) that can be written per second in the system environment where the operation is performed
- 1.5: Coefficient assuming the time excluding disk write, which is the most time-consuming step
Backup data storage destination
```
Recovery time = usageByTheBackupDataStorageDestination / diskWritePerformance x 1.5
```
- usageByTheBackupDataStorageDestination: Disk space used by the backup data
- diskWritePerformance: Measured maximum data volume (bytes/second) that can be written per second in the system environment where the operation is performed
- 1.5: Coefficient assuming the time excluding disk write, which is the most time-consuming step