All worker nodes must satisfy the following requirements:
- x86-64 CPU cores with SSE4.2 instruction support
- Linux kernel 5.13 or higher with the following modules loaded:
- ext4 and optionally xfs
Each worker node which will host an instance of a Mayastor pod must have the following resources free and available for exclusive use by that pod:
- Two CPU cores
- 4GiB RAM
- HugePage support
- A minimum of 2GiB of 2MiB-sized pages
As of version 1.0 the minimum supported worker node count is three nodes.
Note also, that when using the synchronous replication feature (N+1 mirroring), the number of worker nodes to which Mayastor pods are deployed should be no less than the desired replication factor. E.g. four-way mirroring of a volume would require Mayastor pods to be deployed to a minimum of four worker nodes within the cluster.
As of version 1.0, Mayastor supports the export and mounting of volumes over NVMe-oF TCP only. Worker nodes on which an application pod using Mayastor volume(s) may be scheduled must have the requisite NVMe initiator support installed and configured.