docs/deployments/platforms/virtualization/proxmox/Fixing iSCSI Connections that Drop at Reboot.md

Purpose

You may enounter an issue where when you reboot your ProxmoxVE cluster node, it will fail to start virtual machines because of an error related to being unable to see the underlying LVM disks for the GuestVM. This is generally bandaid-fixed by running iscsiadm -m node --login on the node, which makes it reconnect to the cluster's iSCSI storage. This is not a viable long-term solution.

Configure Automatic Startup of iSCSI Services

Run these commands on every ProxmoxVE cluster node:

systemctl enable --now iscsid
systemctl enable --now open-iscsi

Discover Targets (To Create Necessary iSCSI Records)

iscsiadm -m discovery -t sendtargets -p 192.168.3.3:3260

Configure Automatic iSCSI Target Connection Behavior

Run these commands on every ProxmoxVE cluster node:

iscsiadm -m node \
  -T iqn.2026-01.io.bunny-lab:storage:iscsi-cluster-storage \
  -p 192.168.3.3 \
  --op update \
  -n node.startup \
  -v automatic

iscsiadm -m node \
  -T iqn.2026-01.io.bunny-lab:storage:iscsi-cluster-storage \
  -p 192.168.3.3 \
  --op update \
  -n node.conn[0].startup \
  -v automatic

Verify Configuration

You will want to ensure that node.startup = automatic and node.conn[0].startup = automatic when you run the following command.

iscsiadm -m node -o show | grep -E 'node.name|node.conn\[0\].address|node.startup|node.conn\[0\].startup'

Login/Mount iSCSI Targets

iscsiadm -m node \
  -T iqn.2026-01.io.bunny-lab:storage:iscsi-cluster-storage \
  -p 192.168.3.3:3260 \
  --login

!!! Success "Example Output" If everything worked correctly, you should see output like the example below: sh node.name = iqn.2026-01.io.bunny-lab:storage:iscsi-cluster-storage node.startup = automatic node.conn[0].address = 192.168.3.3 node.conn[0].startup = automatic