witness sangfor 6.10.0R2 5

What does witness sangfor 6.10.0R2 do? What are the advantages if there is witness?

11111111111111111

111111111111

11111111111111111

11111111111111

Hello, in a circumstance when there are just two nodes in a hci cluster, a witness is required. I advise using three nodes when building an HCI cluster rather than a server that serves as a witness.

Hello Dear

A high-availability component called Sangfor Witness 6.10.0R2 is utilized in Sangfor HCI systems to guarantee cluster availability and data integrity in the event of node failures. Having a Witness has several benefits, such as increased fault tolerance, cluster stability, and the capacity to continue providing services even in the case of hardware failure or network partitioning, guaranteeing low downtime and data consistency throughout the cluster.

Sangfor Witness 6.10.0R2 is a high-availability component used in Sangfor HCI environments to ensure data integrity and cluster availability during node failures. The advantages of having a Witness include enhanced fault tolerance, improved cluster stability and the ability to maintain continuous service even in the event of network partitioning/hardware failure ensuring minimal downtime and data consistency across the cluster.

HiHi,

To ensure data stability and reduce the risk of data split-brain scenarios, please add a witness node to a 2-host cluster to create a 2+1 configuration. Data split-brain can lead to virtual machines failing to start properly or data loss.

Please find the attahced file for your Ref.

2 1_Scenario_Witness_Deployment_Guide.pdf (752.2 KB, Downloads: 38)

2024-Dec-05 11:32 Upload

Click file name to download attachment

Hi, a witness node is used when Sangfor HCI is deployed as 2 nodes only. As you know, by having 2 nodes deployment you can benefit the 2 copies features. Where if 1 node is offline, the data is still available on node 2. The storage of sangfor using almost the same concept of RAID 10.

So when the data is being copied to the 2 nodes, a data sync is done. This is an automatic thing from HCI when new data is input or the nodes suddenly offline due to blackout and recovers back again, This is to ensure the data integrity and to fix if data is anomaly

However, when using 2 nodes has a flaw. When one of the nodes or both of them frequently offline or suddenly powered due to circumstances, sometimes the data sync is unable to fix data, causing the 2 copies to be an anomaly thus affects the storage and productions due to "brainsplit or splitbrain".

So here where the witness node come, where as it acts something like a standby node where 1 of the nodes is suddenly offline. I've not fully read and remember what are the specifics it will do but in a nutshell, it will work by to avoid the brainsplit or splitbrain issue from happenning. The version of the witness node is just because it's like a normal HCI node, just used for something else. So it will have to be the same version of the existing 2 nodes.

Hi, a witness is necessary on a scenario with an hci cluster formed by only two nodes. If you consider to build an hci cluster, i suggest to do it with three nodes (avoiding use of a server that act as witness).