1 out of 2 controllers is faulty or 3 out of 4 controllers are faulty. What mechanisms are used to protect data when storage system works with a single controller?
When an engine works with a single controller (for example: 1 out of 2 controllers is faulty or 3 out of 4 controllers are faulty) if the faulty controller cannot return to normal within the specified time period (192 hours by default), the engine shifts to the Write Protection mode. In this mode the storage system cannot provide write services and only provides read services.
You can query the period that the system runs in a single controller environment before the write policy is switched to write protection by running following command:
admin:/>show system write_policy
You can change that period by running following command:
admin:/>change system write_policy write_protect_time=X
For more information about this command please see: https://support.huawei.com/hedex/hdx.do?docid=EDOC1100141442&id=system_change-system-write_policy&lang=en
When an engine works with a single controller and Write Protection is enabled if second controller returns to Normal state the Write Protection mode will be disabled automatically and storage continues to provide write services.
When an engine works with a single controller and Write Protection is not yet enabled if remaining controller become faulty or reboots the data in cache will not be lost. BBU will provide power to transfer data from controller’s cache to built-in M.2 SSD. When controller comes back to normal the data from M.2 SSD transfers back to controller’s cache and can be written to disks.