The Geo SCADA server configuration tool contains a field in the Historic Configuration | Historic Data dialog - Historic Data section titled ‘Archive after’. This article will describe what this field is used for and the effect it has on server synchronization when multiple Geo SCADA servers are configured in a redundant configuration (I.E. Hot/Standby, Triple Standby).
To locate the ‘Archive after’ field, open the Geo SCADA server configuration dialog and navigate to:
Historic Configuration | Historic Data
On this dialog, locate the ‘Historic Data’ section and note the ‘Archive after’ field:
When Geo SCADA is first installed, the default setting for this field is set to ‘not enabled’ and ‘0 weeks’.
The purpose of the ‘Archive after’ field is to determine when historic data is changed from ‘read/write data’ to ‘read only data’. When the ‘Archive after’ field is not enabled, all historic data is set to ‘read/write data’. In the example from the screen shot above, historic data is being stored for 52 weeks. This means that all 52 weeks’ worth of historic data may be modified or deleted (or new data added).
When it comes to redundancy, the ‘Archive after’ field has a significant impact on server synchronization. When a Standby server is first started up and synchronized to a MAIN server, it copies over all objects and history from the MAIN server’s database. Once synchronized, the Standby server’s database is a mirror image of the MAIN server’s database. That way, should the MAIN server go down, the Standby server will take over and all configuration/data will remain consistent.
If both MAIN and Standby servers continue to run, but they lose connection to each other (I.E. a network communication outage), when the connection is re-established, the Standby server will resynchronize to the MAIN server. If ‘Archive after’ is disabled, the Standby server will synchronize all historic data from the MAIN server (in this case, all 52 weeks’ worth of historic data).
Let’s examine what occurs when we enable the ‘Archive after’ field. In this example, ‘Archive after’ has been enabled and set to a value of 1 week (on both MAIN and Standby servers):
For the database, this means that all historic data that is up to 1 week old will be ‘read/write data’. Any historic data within this time window may be modified or deleted; and new data may be added. Any historic data that is older than 1 week will be set as ‘read only data’ and my not be modified, deleted or added to in any way.
With regards to redundancy, this setting affects the time it takes a Standby server to synchronize to a MAIN server as follows:
Once this change is made, the first time a Standby server re-synchronizes to a MAIN server, no noticeable difference will be seen. The synchronization will occur the same as when ‘Archive after’ was disabled.
Any subsequent re-synchronizations will be noticeably faster. Here’s why:
Of the 52 weeks’ worth of historic data that is being stored, the oldest 51 weeks of that data is now ‘read only’ data and the newest 1 week of data is ‘read/write data’. If the Standby server loses synch with the MAIN server and then reconnects, the first thing it will do is synchronize its database to the MAIN server’s database.
The standby server will first synchronize all database objects. It will then move on to synchronizing historic data. The Standby server knows that any historic data older than 1 week is ‘read only data’ and it will not have been modified. As such, the Standby server will not synchronize that historic data (as it knows that its historic data matches the MAIN servers historic data for those 51 weeks) and will only synchronize historic data that is up to 1 week old. This equates to far less data (I.E. gigabytes – potentially) that needs to be transferred between the MAIN server and the Standby server, resulting in a (potentially) far shorter synchronization time between servers.
As the ‘keep online for’ field may be set to a longer period (I.E. 300 weeks), the affect of setting this field becomes substantially more noticeable with ‘keep online for’ set to larger values; compared to synchronization with ‘Archive after’ disabled.
Summary
With ‘Archive after’ disabled, Geo SCADA users trade longer synchronization times between redundant servers with the ability to add/remove/modify all historic data in the database.
With ‘Archive after’ enabled, Geo SCADA users trade not having the ability to add/remove/modify older historic data with shorter synchronization times between redundant servers.
