I've got a 3 node Proxmox Virtual Environment (PVE) cluster running. I configured pangolin to access the first host from this cluster by IP. For redundancy I added the other 2 hosts. Doing so creates a problem with the built-in shell in PVE. The connection fails on every host. As soon as I disable the other 2 and only connect to the same host everytime, the shell works as expected
Environment
OS Type & Version: (e.g., Ubuntu 22.04) PVE 9
Pangolin Version: 1.15.4
Gerbil Version: 1.3
Traefik Version: 3.6
Newt Version: 1.9
Olm Version: (if applicable)
To Reproduce
add more than 1 PVE host for one ressource
try to open a shell connection through the PVE web interface connected through pangolin
disable the other connections in pangolin ressource settings --> console opens as expected
Expected Behavior
the shell connection functionality works regardless if more than one host from a given cluster is added to pangolin
Originally created by @Wri84 on GitHub (Feb 21, 2026).
Original GitHub issue: https://github.com/fosrl/pangolin/issues/2515
### Describe the Bug
I've got a 3 node Proxmox Virtual Environment (PVE) cluster running. I configured pangolin to access the first host from this cluster by IP. For redundancy I added the other 2 hosts. Doing so creates a problem with the built-in shell in PVE. The connection fails on every host. As soon as I disable the other 2 and only connect to the same host everytime, the shell works as expected
### Environment
- OS Type & Version: (e.g., Ubuntu 22.04) PVE 9
- Pangolin Version: 1.15.4
- Gerbil Version: 1.3
- Traefik Version: 3.6
- Newt Version: 1.9
- Olm Version: (if applicable)
### To Reproduce
- add more than 1 PVE host for one ressource
- try to open a shell connection through the PVE web interface connected through pangolin
- shell connection fails with error `TASK ERROR: command '/usr/bin/termproxy 5900 --path /nodes/pvenode01 --perm Sys.Console --vncticket-endpoint -- /usr/bin/ssh -e none -o 'BatchMode=yes' -o 'HostKeyAlias=pvenode01' -o 'UserKnownHostsFile=/etc/pve/nodes/pvenode01/ssh_known_hosts' -o 'GlobalKnownHostsFile=none' -t root@10.10.10.10 --' failed: exit code 1`
- disable the other connections in pangolin ressource settings --> console opens as expected
### Expected Behavior
- the shell connection functionality works regardless if more than one host from a given cluster is added to pangolin
Hi @Wri84
for this you can use the "sticky sessions" option in proxy settings of your resource.
<!-- gh-comment-id:3939778130 -->
@Lokowitz commented on GitHub (Feb 22, 2026):
Hi @Wri84
for this you can use the "sticky sessions" option in proxy settings of your resource.
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Originally created by @Wri84 on GitHub (Feb 21, 2026).
Original GitHub issue: https://github.com/fosrl/pangolin/issues/2515
Describe the Bug
I've got a 3 node Proxmox Virtual Environment (PVE) cluster running. I configured pangolin to access the first host from this cluster by IP. For redundancy I added the other 2 hosts. Doing so creates a problem with the built-in shell in PVE. The connection fails on every host. As soon as I disable the other 2 and only connect to the same host everytime, the shell works as expected
Environment
To Reproduce
TASK ERROR: command '/usr/bin/termproxy 5900 --path /nodes/pvenode01 --perm Sys.Console --vncticket-endpoint -- /usr/bin/ssh -e none -o 'BatchMode=yes' -o 'HostKeyAlias=pvenode01' -o 'UserKnownHostsFile=/etc/pve/nodes/pvenode01/ssh_known_hosts' -o 'GlobalKnownHostsFile=none' -t root@10.10.10.10 --' failed: exit code 1Expected Behavior
@Lokowitz commented on GitHub (Feb 22, 2026):
Hi @Wri84
for this you can use the "sticky sessions" option in proxy settings of your resource.
@oschwartz10612 commented on GitHub (Feb 25, 2026):
Thanks @Lokowitz reopen if this does not fix.