Quick question, I have a number of servers that replicate to another server, which has extended replication to a third server offsite.
The first group of servers (original and first replica) are onsite and replicate over internal network and all is fine, but the ones to the offsite server travel over the internet, it is also hampered by a very slow RAID array.
Sometimes I have experienced that the replication between the second and third offsite server will fail and cannot recover either with 'Resume replication' or 'resynchronization' and the only option is to remove the replicated VM and start again from scratch.
My question is, is there an amount of time that determines whether or not the replication process can be 'resumed', 'resynced' or is just too long to recover and therefore the only option is to delete and replicate the whole thing again?
Is this "time-based", or "volume of data that has changed-based"? If time-based, what is the time period?
I mainly ask because I am taking the extended replication server offline to replace the RAID card with a better (faster) one and want to try and establish if I have a time limit to have the thing offline before I know I'll need to delete all the replicas and start from fresh.
Thank you.
