When the CRL value increases, so does the number of simultaneous instance failures a cluster can support. However, an increase in the CRL value also causes an increase in messaging traffic, as there are multiple copies of a lock in the cluster. An increase in overhead is required to maintain this redundancy level.
For example, other configuration parameters, such as number of locks and cache size also need more resources for CRL values greater than 1, which means you must increase max memory for the same number of locks value. CRL is resource-intensive, even though Adaptive Server performs the overhead value computations transparently.
Sybase recommends that you set a cluster redundancy level of 1 unless multiple instance failure is common.
