30 lines
1.7 KiB
INI
30 lines
1.7 KiB
INI
## Local Engineering HA Cluster
|
|
#
|
|
#KVM Cluster based on Proxmox Virtual Environment (PVE). 3 Node setup with centralized storage
|
|
#to `Truenas Scale` installation.
|
|
#
|
|
#>
|
|
#> **!!! WARNING !!! !!! WARNING !!! !!! WARNING !!! !!! WARNING !!! !!! WARNING !!!**
|
|
#>
|
|
#> You must shut down _**ALL**_ machines using shared storage before upgrading or rebooting the
|
|
#> storage server. Failure to do so will result in corruption or loss of data.
|
|
#>
|
|
#
|
|
#
|
|
## Network redundancy
|
|
#- Nics 1-4 are for client access and are `bonded` for redundancy and throughput.
|
|
#- Nic 7 is for primary cluster sync and `heartbeat` checks.
|
|
#- Nic 8 is for management access of the cluster, and secondary heartbeat link.
|
|
#- Nic 9-10 are `bonded` 10Gb links for dedicated storage access to the Truenas.
|
|
# - This is a high speed link and has redundant access. The Storage Server has dual 10Gb links but at this time it is intended to break those links and use a single pair for backup and the other for user access.
|
|
#
|
|
#
|
|
#> While the KVM Servers are part of a clustered HA environment the back end storage is not part of an HA or distributed file system.
|
|
#>
|
|
#> This means that should the storage server go down all VM's will be unavailable and potentially data corruption.
|
|
#>
|
|
#> **IF** Distributed files systems are required in order to withstand a storage server failure without service interruption is required, Proxmox natively supports working with `Ceph` distributed file systems. This would require installing at least 3 storage servers with 100GB interconnects for replication data between nodes. Proxmox would then connect to this storage server as its `storage repository`
|
|
|
|
keyboard: en-us
|
|
migration: network=10.10.22.66/24,type=secure
|