Sep 24, 2026
00:58:40
Category: Virtualization
Product: Open-E JovianDSS
Language: English
Watch a detailed solution breakdown showing where the 20% savings come from, followed by an in-depth live demo by Janusz Bak, Open-E CTO, who demonstrates the full feature set and functionality step-by-step.
During the webinar we received a number of questions. As promised here we provide all the answers from our Q&A session:
[ ? ] Can I send the ZFS snapshots to PBS, or is it Proxmox Backup Server taking the backups
Currently, you cannot send ZFS snapshots directly to Proxmox Backup Server (PBS). PBS handles backups by taking full backups that are split and stored in chunks. However, developers and the community are actively working on a plugin solution to integrate ZFS snapshot transfers with PBS in the future.
[ ? ] Is this a single plugin for both iSCSI and NFS, or are there two separate plugins?
It is a single plugin that contains both iSCSI and NFS functionality built into it. It was initially advertised for iSCSI while NFS was undergoing QA testing, but both capabilities were included in the same plugin from the beginning and are now fully supported.
[ ? ] Can Proxmox operate over iSCSI with a JovianDSS HA Metro Cluster as the target? Proxmox runs on several independent servers (hypervisors).
Yes. A JovianDSS HA Metro Cluster running on separate bare-metal servers can provide iSCSI storage to multiple Proxmox hosts or even multiple Proxmox clusters. This is supported through the cluster-prefix property in the plugin configuration.
[ ? ] Are there any performance differences of VMs when comparing iSCSI vs NFS?
The performance difference is very minimal and comparable, depending on whether you measure random or sequential reads/writes. However, iSCSI is preferred for mission-critical setups because it supports true switch-agnostic multipathing, whereas NFS requires network bonding, which depends on specific switch capabilities and can encounter issues during network disruptions.
[ ? ] Is PBS using a JovianDSS Pool as Backup Storage, or is it using disks of the 3rd PVE Node in this demo?
In the demo, PBS uses local NVMe disks on the Proxmox server (configured as a local ZFS pool passed through to the PBS instance), not a JovianDSS pool. The setup relies directly on built-in ZFS capabilities within PBS, so JovianDSS is not required for backup storage here.
[ ? ] In case we use NFS isn't better to backup snapshots to 3rd Open-E JovianDSS instance on that 3rd server?
Yes. When using NFS, JovianDSS On- and Off-site Data Protection (OODP), the built-in backup and replication functionality, can be used as a backup and disaster recovery solution.
With the NFS part of the plugin, multiple virtual disks are stored in a single ZFS dataset. A single OODP replication task can therefore back up all virtual disks stored in that dataset at the same time.
[ ? ] Does Metro Cluster need extra Network interfaces between the Open-E VM on the PVE?
Yes. The Metro Cluster requires a dedicated Mirror Path. The recommended configuration is a switchless, direct point-to-point connection using either a single network interface or bonded interfaces.
In our webinar demo, we configured one NIC as a PCI passthrough device. As a result, JovianDSS sees it as a physical NIC rather than a virtual interface. This passthrough NIC is used for the Mirror Path.
The passthrough NICs in pve1 and pve2 are connected directly to each other using a point-to-point connection, without a network switch.
A direct point-to-point connection is recommended for the Mirror Path, although a switched network connection is also supported.