@olivierlambert Yes, when the host reboots, we're seeing 10+ VMs all configured with the same IP address power up. Where can I confirm that start/start_on is indeed blocked for the DR VMs?

The DR VM has auto_poweron set to true like the source VM:

[root@xcp01 ~]# xe vm-param-get param-name=other-config uuid=41d5dc37-0512-6810-185b-ae7da536cb3b
xo:backup:sr: ac43e953-1c29-f425-5c14-74e5fdb209f6; import_task: OpaqueRef:a7cc8a3f-70b0-42dd-ada5-53d1f1573a1e; mac_seed: 237107d5-1903-5701-11b3-6a6fe9c18dd2; xo:backup:schedule: 0cf32de6-1a4a-49b7-844f-5dea7bb71fd0; xo:backup:job: 7a8da676-e336-49d8-a1ba-02db6990f5a3; xo:backup:vm: e7b0b978-aaa9-a9b3-8079-d13964ab06fd; auto_poweron: true; base_template_name: Other install media; install-methods: cdrom

@julien-f I'm experiencing this as well. I've a remote site running an older version of XCP-ng (7.6). Two hosts without shared storage, Xen Orchestra is configured in Disaster Recovery mode, copying the VMs from one host to the other. It works well, except when the site has a power issue and both the Live and DR VMs power up at the same time.

HA is not configured (no shared storage), to avoid this I need to manually edit the backup VMs and un-tick "Autostart on server boot", but this doesn't work for unexpected outages.

Any suggestions as to what I should check from here?