So we have 2 PI Data Archive servers in different networks, connected via interface. Both networks have had time issues, where the source archive server has an advanced timing of more than 10 mins (Around 15-20mins). We are now planning to resolve the issue and sync the times. Now we are considering what the possible consequences of this to our data.
Should the source server's time be adjusted back in time, then there is a possibility that it then assigns timestamps that already have values in the destination servers(possibly to the millisecond). My question is, does this get stored as a duplicate timestamp (with a different event) or does the source server overwrite the data or does the destination retain the already stored event and neglect the source server's data?
Edit: Also, what other concerns do we have to look for on this? We are aware that out-of-order events will be stored and the trend sequence would be off since it would store current data in previous timestamps. Other than that, what should we look out for?