We got a call yesterday from a customer who was using our xSQL Data Compare to compare a 1TB database. He was concerned that the comparison was taking a very long time – when he called it had been running for about 10 hours and was still going! While we don’t think there are many users out there comparing 1TB or larger databases we thought it might be helpful to those few out there if we explained why such compare may take 10 hours or more depending on the environment. Here are the factors to consider:
- Connection speed - in a typical scenario xSQL Data Compare is running on a client machine let’s call that Client1 and the databases being compared reside on let's say Server1 and Server2. In order to compare those databases the whole 1TB worth of data from Server1 and another 1TB worth of data from Server2 will have to be "brought" over to Client1. So you are gradually transferring 2TB of data over the wire (or over the air) and depending on how fast the connections Client1 – Server1 and Client1-Server2 are this process alone may take not just 10 but 30 or 40 hours.
- Processing power – xSQL Data Compare running on Client1 needs to pair those millions of rows and compare them one by one. If you have a slow machine, even if the data is readily available it will take a long time to process all that data.
- I/O and local disk speed – 2TB worth of data is being temporarily stored on the Client1 hard drive and that process alone may take a long time.
In short, when you are comparing large and very large databases don't be shocked if the process takes many hours to complete. The important thing is that, provided you have sufficient disk space on the machine where xSQL Data Compare is running, the databases will be compared successfully and you will be able to generate the synchronization script.