Hi,
I would like to build a 2-node Proxmox Cluster with a third node being a Raspberry PI using LINSTOR and DRBD.
Next, please describe how to configure LINSTOR on PVE nodes and RasPI to achieve this setup. What I understand at the moment: one doesn’t configure DRBD in conf files or other way, for all will be done through LINSTOR. Correct?
Finally please help with Proxmox integration. From my point I get confused, because it seems that cluster setup in LINSTOR docs is not what should be done / is necessary doing LINSTOR on PVE (e.g. I understand that I don’t do LINSTOR resource definition on PVS, only controller, nodes, storage-pool, resource-group and volume-group). Correct? Please give a working example.
You also can point me to an existing documentation for such a setup and I will be happy. Please not to a full-fledged 3 node setup, because my scenario with diskless RasPi is different and I need examples where resources from a resource-group will only have two replicas on the PVE nodes. A resource should and can never be placed on the RasPi, but the RasPi should automatically be the tie breaker for every LINSTOR / DRBD resource and the cluster should be as resilient and self-healing as possible.
Thank you very much.
I would like to build a 2-node Proxmox Cluster with a third node being a Raspberry PI using LINSTOR and DRBD.
- PVE, current Proxmox software, nodes have disks
- RasPi, current Raspberry PI OS, is diskless (only SD card) and should only be the quorum / tie breaker node
Next, please describe how to configure LINSTOR on PVE nodes and RasPI to achieve this setup. What I understand at the moment: one doesn’t configure DRBD in conf files or other way, for all will be done through LINSTOR. Correct?
Finally please help with Proxmox integration. From my point I get confused, because it seems that cluster setup in LINSTOR docs is not what should be done / is necessary doing LINSTOR on PVE (e.g. I understand that I don’t do LINSTOR resource definition on PVS, only controller, nodes, storage-pool, resource-group and volume-group). Correct? Please give a working example.
You also can point me to an existing documentation for such a setup and I will be happy. Please not to a full-fledged 3 node setup, because my scenario with diskless RasPi is different and I need examples where resources from a resource-group will only have two replicas on the PVE nodes. A resource should and can never be placed on the RasPi, but the RasPi should automatically be the tie breaker for every LINSTOR / DRBD resource and the cluster should be as resilient and self-healing as possible.
Thank you very much.