Does ip part2 and ip part3 ever get recomputed back to Node 1 when the DRAM recovers? I assume that those partitions resided in Node 1 in first place was because Node 1 was the preferred location for those partitions.
abist
I don't think it matters where the partition resides. We were using Node1 for ip part2 and ip part3 for better work distribution, but I may be wrong.
Corian
Is there a master node in this process? How are failures detected?
Does ip part2 and ip part3 ever get recomputed back to Node 1 when the DRAM recovers? I assume that those partitions resided in Node 1 in first place was because Node 1 was the preferred location for those partitions.
I don't think it matters where the partition resides. We were using Node1 for ip part2 and ip part3 for better work distribution, but I may be wrong.
Is there a master node in this process? How are failures detected?