Synchronizing parallel servers
A prerequisite for the continuous synchronization of parallel servers is that they are paired together.
Parallel servers are automatically synchronized. Manual synchronization is only performed when a single server is converted to a parallel server. Manual synchronization can be performed via the application VDogParallelServerUtil.exe.
When you run the VDogParallelServerUtil.exe on server A (connected to server B), this tool will cause server B to be overwritten. It will delete the server archive and remote share (if present) from server B. The data is not restorable in the event of errors or non-completion. Please make a backup of your data!
When changes are detected in the following data, parallel servers are automatically synchronized:
- Scheduler Job results
- Global server settings
- Job configuration
- Components
- Component types
- Scripts
- User management
- Agents
- User data
- Upload types
- Compare types
In normal operation, synchronization takes place directly. If the server load is very high (for example, many large check-in and check-out operations at the same time and many scheduler jobs running in parallel), a delay may occur.
If the synchronization fails, no rollback will take place on the server that has already been changed.
|
Why was the last checked-in version of the component on a server overwritten during the parallel server operation? If the connection between the two servers in parallel server mode fails and this goes unnoticed for a period of time, it is possible that different versions of the component may have been checked in on the two servers. Once the connection has been restored, a check of the serial number will take place. The last checked-in version on the server with the smaller serial number will be retained and copied to the other server. The version on the server with the larger serial number will therefore be overwritten and will after no longer available. |
| Related topics: |
|---|