[SOLVED] Proxmox seems to ignore delayed start on VMs

[user51269]

Hi -

Running Proxmox VE 8.1.0, all patches applied.

I have a bunch of VMs but two of them need to start in order. There's a TrueNAS SCALE VM, running 24.04.0. It is set to start at boot, order=1, delay=0. I also have an app VM running a bunch of containers, some of which require access to TrueNAS shares. If they can't access the shares, the containers can't start properly. That VM is is set to start at boot, order=2, up=300. If I'm reading the docs correctly that should mean the Docker VM starts 300s after the TrueNAS VM.

Is that understanding correct?

I'm asking because my findings are very different and my containers all need to be started manually after failing to start when they can't access the storage. `uptime` on both the TrueNAS VM and the Docker VM shows differences of under a minute so the Docker VM is definitely not waiting 300s before starting.

Why is this?

I'm using the docs here: https://pve.proxmox.com/pve-docs/pve-admin-guide.html#qm_virtual_machines_settings

Thanks

 

[nosoop4u]

The delay setting is how to long to wait AFTER that VM is started before starting the next/dependent VMs. So you put it into the parent (TrueNAS) VM in your case, not the dependent VMs.
Think about it this way, if you had multiple dependent VMs, you'd have to put that delay in each, and if your parent VM started taking longer, you'd have to modify multiple dependent VMs. The way it is now, you only set the delay in the VM that causes the issue, not the ones dependent on it. And if the startup time changes, you only change the delay in one place.
Note the wording in the documentation:

• Startup delay: Defines the interval between this VM start and subsequent VMs starts. For example, set it to 240 if you want to wait 240 seconds before starting other VMs.

 

[user51269]

The delay setting is how to long to wait AFTER that VM is started before starting the next/dependent VMs. So you put it into the parent (TrueNAS) VM in your case, not the dependent VMs.
Think about it this way, if you had multiple dependent VMs, you'd have to put that delay in each, and if your parent VM started taking longer, you'd have to modify multiple dependent VMs. The way it is now, you only set the delay in the VM that causes the issue, not the ones dependent on it. And if the startup time changes, you only change the delay in one place.
Note the wording in the documentation:

• Startup delay: Defines the interval between this VM start and subsequent VMs starts. For example, set it to 240 if you want to wait 240 seconds before starting other VMs.

Thanks @nosoop4u that works. Uptime deltas are what I would expect now.