Error: Unable to find free port (61000-61099)

[holr]

Hello,
I have one proxmox server in a cluster (proxmox 6.1-7) that is running many virtual machines. After a recent cloning process to create another batch, I am finding several cannot be started, instead exhibiting the following error message:

Error: Unable to find free port (61000-61099)

We are running spice and seem to be running into the issue described in step 3 in:

https://forum.proxmox.com/threads/capacity-testing-with-proxmox-4-4.32513/

Unfortunately, there wasn't a solution in that thread, or in others we've been looking into. Is there a way, other than changing the SPICE back to NoVNC, that could help alleviate this cap please?
Thanks!

 

[dcsapak]

yeah, it seems we did not increase the limit ... can you please open a bug/feature request on https://bugzilla.proxmox.com so that we can track this

 

[matrix]

Hi,

Can you see if the port 61000 in this file? cat /var/tmp/pve-reserved-ports if the port there already, see which program listen to port netstat -tlp, also see journalctl -f as well.

 

[holr]

yeah, it seems we did not increase the limit ... can you please open a bug/feature request on https://bugzilla.proxmox.com so that we can track this

Thank you for the recommendation, I filed a bug request at https://bugzilla.proxmox.com/show_bug.cgi?id=2618

 

[JosephKav]

Did you have any luck sorting this? I'm getting that exact error on 6.4.8. I'm running 25-30 LXC containers and only 1 VM this error happens when I try and start it with SPICE as the Display (I haven't added a new LXC/VM in many months and have never seen this issue). Changing the Display to default gets it booting again. I tried it with a new VM and that errored with SPICE. Also, not sure if this could be related (they're both network related though, so maybe?), but NUT doesn't seem to work anymore. upsc upsName@localhost just times out, but starting nut-server runs fine and that appears to connect to the UPS. I've tried google, tried rebooting, am now on 5.11 kernel, no luck with any of those. I'm on the latest apt update, so possibly a bug somewhere there?

 

[holr]

Hello, I only went as far as filing the bug and left it there, sorry. I've had to fall back to novnc until the limit is altered. I only experienced the error with 100 or more spice enabled vms running simultaneously. Sorry I can't be more help!

 

[cogis]

I experience this error with proxmox 7.3-6 after setting my node in cluster mode
even with 0 VM running, and no process using port in range 61000-61099

edit : solution is here : https://forum.proxmox.com/threads/error-unable-to-find-free-port.14729/post-78258