On some occasions, the CanRisk Tool will generate an error message when you attempt to process a pedigree that contains a loop (for example where two brothers in one nuclear family marry two sisters in another nuclear family). Loops make the risk calculations more complex which increases memory usage and runtimes, and so on some occasions these calculations fail.
Unfortunately, there is no direct workaround for this. You can try to build the pedigree without the loop and run the risk calculation. However, this may result in some inaccuracy in the predicted risks. As a result, the risks given when you ignore the loop should be treated as indicative risks.