Before recovering RAC and ha, notice the difference between the two in the recovery process, but not recorded, today in the recovery is a bit forgotten, hurriedly record, used for later recovery test time to help memory.
The differences and similarities between RAC and HA can be described in many documents on the Web, which is not discussed here.
Here I only describe the differences I observed:
1: First, in RAC, the servers of different nodes have their own instances, the instance names are different, but the database names are the same, but on Ha, although there are instances, the instance names of the nodes are the same.
2: In the recovery, you need to know from which node to recover data files, in the RAC, select the node to backup data files, and ha different, HA uses only a unified name, such as Ha Two servers (client1,client2), but the external name is Bk_ Client, so when it is backed up, it uses bk_client instead of CLIENT1 or CLIENT2, and if it is the same deployment in RAC, then there will typically be one for backing up the data file, such as CLIENT1, one for backing up the archive log file, such as CLIENT2, so that when we restore the data file, the HA backup machine name is Bk_client, and the RAC backup machine name is CLIENT1.
3: When recovering archived log files, RAC needs to recover from two customer service side, and HA because the external is a server, so the backup time is also for bk_client backup, so restore time from bk_client recovery, The Oracle system automatically recovers archived log files from the nodes below ha.
这些就是我总结的大概的不同之处。
NBU when restoring Oracle, different recovery settings for RAC and HA