Which of the following statements about data routing and cluster management in Huawei's Fusionstorage architecture is correct? (Multiple choice)

Which of the following statements about data routing and cluster management in Huawei’s Fusionstorage architecture is correct? (Multiple choice)

Which of the following statements about data routing and cluster management in Huawei’s Fusionstorage architecture is correct? (Multiple choice)
A . When the system starts, the MDC interacts with ZK to determine the MC, and the main MDC and other MDCs monitor each other’s heartbeats. The main MDC decides who will take over after a certain MDC fails. Other MDCs find that the main MDC fails and interact with ZK to take over the main MDC.
B . When VBS starts, it queries the primary MDC and registers with it (the primary MDC maintains an active VBS list, which it synchronizes to other MDCs, notifies VBS of OSD status changes, and asks MDC to confirm whether it is the leader)
C . Distributed storage system writes data. VBS generates multiple keys in units of 1MB based on the LBA and block size from the application layer, hashes the keys into partition IDs, and then follows up the partition to find the primary OSD. The primary OSD notifies the OSD corresponding to the partition to start writing data.
D . When creating a distributed storage system, N partitions are generated in the space of 2 to the power of 32, and the backup OSD corresponding to the partition is generated according to the distribution of the disk to start writing data
E . When OSD starts, it queries MDC and reports the status to the affiliated MDC, which sends the status change to VBS. When the affiliated MDC fails, the active MDC designates an MDC to take over, and a maximum of two pools belong to one MDC.

Answer: ABCE