I have just finished up a full 3 node cluster/HA deployment in the lab, and I am finding it to be a time bomb for small/medium businesses due to too many single points of failure for the cluster.
My concerns revolve around cluster corruption due to power failures outside of a data center environment, 3 hosts off a single UPS is an accident waiting to happen, a UPS per host with graceful shutdown is complex. . So rather than a 3 node cluster I am wondering if a 2 node set up without HA and just using replication between a primary host and a secondary host with a "Cold spare" VM is a better fit.
Just a single Managed switch reboot runs the risk of corrupting a cluster, so clustering proxmox has got me focused on additional support systems, redundant switches, multiple redundant UPS. etc...
If I had 2 nodes one hot one cold and a proxmox back up server, I think this would provide a more resilient set up where entire system can weather a full hard power down scenario. I am in South Florida where Hurricanes have become annual events.
Any forum guidance would be greatly appreciated.
My concerns revolve around cluster corruption due to power failures outside of a data center environment, 3 hosts off a single UPS is an accident waiting to happen, a UPS per host with graceful shutdown is complex. . So rather than a 3 node cluster I am wondering if a 2 node set up without HA and just using replication between a primary host and a secondary host with a "Cold spare" VM is a better fit.
Just a single Managed switch reboot runs the risk of corrupting a cluster, so clustering proxmox has got me focused on additional support systems, redundant switches, multiple redundant UPS. etc...
If I had 2 nodes one hot one cold and a proxmox back up server, I think this would provide a more resilient set up where entire system can weather a full hard power down scenario. I am in South Florida where Hurricanes have become annual events.
Any forum guidance would be greatly appreciated.
Last edited: