Our great sponsors
-
WorkOS
The modern identity platform for B2B SaaS. The APIs are flexible and easy-to-use, supporting authentication, user identity, and complex enterprise features like SSO and SCIM provisioning.
For block, I normally go down two roads. The first is Rancher's local path which is basically bind-mounts but presented as PVCs. link This is great for things like DB clusters where you don't have node-level redundancy of the data because the application is taking care of that for you. The second one is Rancher's Longhorn Link which provides iSCSI and NFS storage but from inside the k8s cluster. I personally find the NFS performance is pretty good when it comes to Longhorn because it has been tuned, it's designed to scale out IE each share has its own NFS server and it's written in GO.
For block, I normally go down two roads. The first is Rancher's local path which is basically bind-mounts but presented as PVCs. link This is great for things like DB clusters where you don't have node-level redundancy of the data because the application is taking care of that for you. The second one is Rancher's Longhorn Link which provides iSCSI and NFS storage but from inside the k8s cluster. I personally find the NFS performance is pretty good when it comes to Longhorn because it has been tuned, it's designed to scale out IE each share has its own NFS server and it's written in GO.
Related posts
- Diskomator – NVMe-TCP at your fingertips
- Difference between snapshot-cleanup and snapshot-delete in Longhorn recurring job?
- Ask HN: Any of you run Kubernetes clusters in-house?
- Container redundancy with multiple Unraid servers?
- How do I clean up a Longhorn volume? Trimming the volume doesn't work, "cannot find a valid mountpoint for volume"