Continuing the discussion from SingleStore High Availability and Storage Distribution on Node Fail over:
thanks mmanthena for your detailed explanation .
So my another observation is , one availability group is completely down / failure and after that one of the node leaf is came up in the group , in that situation that node should have an entire storage capacity of other failed nodes in that availability group till other remaining nodes has to come online . Is that understanding is correct ?
Hey Anish,
When HA is enabled (recommended) two copies of the data are stored across the two availability groups. In your example of 4 leaf nodes, there would be 2 leaves in AG1 and two leaves in AG2. If you store something within the database, one copy will be on a leaf in AG1 and the other copy on a leaf in AG2.
When a leaf (or underlying host) fails, any running workload will be processed by pulling data from the replica data on the leaf in the other AG. Because of that, there is no provisioning or increase in storage needed on the leaves during failover; they are already provisioned with multiple copies of the data allowing continuous operation through failure.
I hope this is helpful, please let me know if you have additional questions.
Yes Micah , I am clear about the HA part . I am just sharing my observations in a screenshot .
Case 4 : please check
Is any update on this …please correct me if i am wrong .
