https://rancher.com/ logo
Title
h

handsome-autumn-77266

04/21/2023, 2:07 PM
It seems way easier to just use longhorn then manually setting up a db to provide persistent storage across node failures. Is Longhorn recommended for k3s?
c

creamy-pencil-82913

04/21/2023, 3:37 PM
databases and replicated block devices are not really the same thing. What problem are you trying to solve?
h

handsome-autumn-77266

04/24/2023, 1:40 PM
@creamy-pencil-82913 If I have a 5 node cluster would it make more sense making that 5th node a server node or having 2 agent nodes running workloads?
@creamy-pencil-82913 For maintaining all your application state across node failures, you would generally use a replicated database? Replicated databases provide a way to store data across multiple nodes so that even if one node fails, data can still be accessed from another node. This is typically achieved using replication protocols such as master-slave or master-master replication. Replicated block devices, on the other hand, provide a way to store data on multiple disks so that even if one disk fails, the data can still be accessed from another disk. While replicated block devices can be useful for maintaining data availability, they are not typically used for maintaining application state across node failures. This is what I found on internet so I guess I was wrong.