how Sangfor HCI behaves in the event of an unexpected power outage?

This may be a simple question, but I would like some clarification about how Sangfor HCI behaves in the event of an unexpected power outage.

In a traditional infrastructure consisting of hosts and a SAN with a disk array, the RAID controller typically includes a battery-backed cache. This allows pending writes to be safely committed and enables write-back caching even if the system loses power unexpectedly.

Of course, I understand that it is possible to deploy UPS systems that communicate with the HCI platform so that, in the event of a power failure, they can gracefully shut down the virtual machines and then the physical hosts.

However, my question is about a worst-case scenario: what happens if power is suddenly lost to all hosts, or at least to multiple parent hosts at the same time, before a graceful shutdown can take place? What happens to the data in that situation?

Throughout my career working with traditional (non-hyperconverged) infrastructures, I've experienced several unexpected power outages, and fortunately I've never encountered data corruption or startup issues after power was restored (perhaps I've just been lucky).

Unfortunately, even in well-designed environments, power failures remain a relatively common risk in small and medium-sized enterprise data centers.

I would really appreciate hearing your thoughts on this topic, as well as any real-world experiences you may have had. Most importantly, I would like to understand how Sangfor HCI is designed to protect data and prevent data loss or corruption during an unexpected power outage.

I hope this isn't too basic a question, but it's something I've been wondering about, and I would appreciate some clarification.

Thank you for your reply.

However, I don't think it addresses my question.

I'm not asking about best practices to minimize the impact of a power outage (UPS, HA, backups, replication, etc.), nor about how to recover after such an event.

My question is specifically about the storage write path and data protection mechanism.

In a traditional SAN, for example, the RAID controller typically uses battery-backed or flash-backed write cache. If power is suddenly lost, pending writes are preserved and later committed to disk, protecting data integrity.

What is the equivalent mechanism in Sangfor HCI?
What is the equivalent of the battery-backed write cache used in a traditional SAN?

If power is unexpectedly removed from all nodes at exactly the same time, what happens to writes that have already been acknowledged but have not yet reached persistent storage? How does Sangfor HCI guarantee data consistency in this scenario?

I'm trying to understand the storage architecture rather than the operational procedures for handling a power outage.

To minimize the impact, do the following:
- Install a UPS for each node and critical network equipment, and configure the UPS it to perform a gradual shutdown.
- Enable HA and configure the appropriate VM restart priority.
- Use a storage replication policy that aligns with your availability requirements.
- Perform regular backups and verify recovery procedures.
- Test power failure recovery during scheduled maintenance windows.
- Monitor cluster health after recovery to ensure replica synchronization completes successfully.