- File deletion/corruption disk failures
- User errors
- Network problem
- Instance error
File Deletion/Corruption & Disk Failures
Installation of the operating system and postgresql at the same patch level on the new
server,
Copying terabytes of data from the backup unit to the target server,
Restoring data can take hours or even days.
With hot standby, it is possible to avoid all these time & work losses. The new system
can be reactivated without losing data in just one restart.
Damage to the actively written WAL file may stop the system and cause data loss.
Therefore, RAID cards with security measures such as backup batteries should be
preferred for WAL files.
User Errors
If there is a standby server available, it can be accessed. It can be restored from a
backup to a specific point in time (point-in recovery). (To avoid affecting the live
environment, it should be done on a different server).
Network Errors
If 4 network cards are bonded in pairs and if Postgresql listens to these two IPs,
redundancy is ensured. In virtualization environments, the system administrator should
be consulted.
Instance Errors
Finding parts can take days or weeks if annual support of the hardware has not been
purchased or has expired. Again, one of the most effective solutions here is hot
standby.