gc cr failure

 Hi,

We have a RAC with 4 nodes. 

We have had a problem with a SELECT because it hasn´t terminated. Normally this SELECT runs about 5 minutes, but in this case it runs about 5 hours (since 17:00H) and then we have had to kill it.

In the statistics and AWR we see:

- First (17:00H to 18:00H):

Instance 1: Event "gc cr multi block request" is 60% DB Time

        gc cr blocks received  (2.712 blocks per second)

Instance 3: gc cr blocks served (1.353 blocks per second)

Instance 4: gc cr blocks served (1.347 blocks per second

- Since 18:00H:

Instance 1: Event "gc cr failure" is 100% DB Time

What is "gc cr failure" meaning? Can be a network problem? How can I check the interconnect health?