port 8006 blocked

[BaRaD5]

I have no firewall and i can ping both nodes from each other using tailscale full domain.
I can access webui using the tailscale ip+8006.
What am i missing?

thanks in advanced

[willjasen]

I need more info to help out.

• When you say you have no firewall, do you mean an ACL with Tailscale?
• What exact errors do you run into?

[BaRaD5]

Hey!
I never used ACL in tailscale
when running tailmox.sh on both nodes, i get port 8006 is blocked

[willjasen]

Hey!
I never used ACL in tailscale
when running tailmox.sh on both nodes, i get port 8006 is blocked

If you have no ACL, then Tailscale is permissive by default (allow all).

It is possible that the tests that tailmox performs is mistakenly thinking that TCP 8006 is not responding when it is fine otherwise. Please share all of the output that tailmox returns from both nodes.

[BaRaD5]

cluster owner:

root@SkynetB:/home/pi# ./tailmox.sh
Checking if Proxmox is installed...
Proxmox VE 8.x detected.
Checking for required dependencies...
Starting Tailscale with --advertise-tags 'tag:tailmox'...
Some peers are advertising routes but --accept-routes is false
Waiting for Tailscale to come online...
This host's Tailscale IPv4 address: 100.100.111.11
This host's Tailscale MagicDNS name: dustbowl-proxmox.tail58de2.ts.net
Tailscale certificate services not found. Cloning repository...
./tailmox.sh: line 111: git: command not found
./tailmox.sh: line 115: cd: /opt/tailscale-cert-services: No such file or directory
./tailmox.sh: line 117: git: command not found
./tailmox.sh: line 118: ./proxmox-cert.sh: No such file or directory
./tailmox.sh: line 119: cd: /opt/tailmox: No such file or directory
Checking if all tailmox peers are online...
All tailmox peers are online.
Ensuring the local node can ping all other nodes...
Pinging 100.100.101.100...
Successfully pinged 100.100.101.100.
All Tailmox peers are reachable via ping.
Reporting peer latency...
Calculating average latency for 100.100.101.100 (25 pings with an interval of 0.1 seconds)...
Average latency to 100.100.101.100: 12.396 ms
Checking if TCP port 8006 is available on all nodes...
Checking TCP port 8006 on suijin (100.100.101.100)...
TCP port 8006 is not available on suijin (100.100.101.100).
Some peers have TCP port 8006 unavailable. Please check the network configuration.

the one joining:

root@suijin:/home/pi# ./tailmox.sh
Checking if Proxmox is installed...
Proxmox VE 8.x detected.
Checking for required dependencies...
Starting Tailscale with --advertise-tags 'tag:tailmox'...
Some peers are advertising routes but --accept-routes is false
Waiting for Tailscale to come online...
This host's Tailscale IPv4 address: 100.100.101.100
This host's Tailscale MagicDNS name: suijin-proxmox.tail58de2.ts.net
Tailscale certificate services not found. Cloning repository...
./tailmox.sh: line 111: git: command not found
./tailmox.sh: line 115: cd: /opt/tailscale-cert-services: No such file or directory
./tailmox.sh: line 117: git: command not found
./tailmox.sh: line 118: ./proxmox-cert.sh: No such file or directory
./tailmox.sh: line 119: cd: /opt/tailmox: No such file or directory
Checking if all tailmox peers are online...
All tailmox peers are online.
Ensuring the local node can ping all other nodes...
Pinging 100.100.111.11...
Successfully pinged 100.100.111.11.
All Tailmox peers are reachable via ping.
Reporting peer latency...
Calculating average latency for 100.100.111.11 (25 pings with an interval of 0.1 seconds)...
Average latency to 100.100.111.11: 12.662 ms
Checking if TCP port 8006 is available on all nodes...
Checking TCP port 8006 on SkynetB (100.100.111.11)...
TCP port 8006 is not available on SkynetB (100.100.111.11).
Some peers have TCP port 8006 unavailable. Please check the network configuration.

[willjasen]

Ah. There's an issue with the script installing the tailscale-cert-services repo and that's likely why. Give me a bit because I'll need to spin up my environment to account for such a scenario and update the code, then I'll let you know.

[BaRaD5]

thank you i appreciate this a lot!!!

[BaRaD5]

I installed git...
now i get this on both:

Scheduling Proxmox certificate renewal...
Checking crontab...
no crontab for root
no crontab for root
./tailmox.sh: line 119: cd: /opt/tailmox: No such file or directory
Checking if all tailmox peers are online...
All tailmox peers are online.
Ensuring the local node can ping all other nodes...
Pinging 100.100.111.11...
Successfully pinged 100.100.111.11.
All Tailmox peers are reachable via ping.
Reporting peer latency...
Calculating average latency for 100.100.111.11 (25 pings with an interval of 0.1 seconds)...
Average latency to 100.100.111.11: 15.256 ms
Checking if TCP port 8006 is available on all nodes...
Checking TCP port 8006 on SkynetB (100.100.111.11)...
TCP port 8006 is not available on SkynetB (100.100.111.11).
Some peers have TCP port 8006 unavailable. Please check the network configuration.

[willjasen]

tailmox has to be installed at /opt/tailmox so cd into the /opt directory, pull down the tailmox repo from there, then try running it again.

You should be running:
/opt/tailmox/tailmox.sh

Sorry, I know that requirement isn't exactly clear, but I wrote the whole thing in 36 hours, so some things were overlooked for time.

[BaRaD5]

36 hours is impressive!
ill try to do so in a few hours\tomorrow

[willjasen]

36 hours is impressive!
ill try to do so in a few hours\tomorrow

Impressive? Manic is more like it, coupled with a burning desire to prove other people incorrect. But thanks!

I will still update the code to account for these particulars soon.

[BaRaD5]

ok i have done it in the cluster creator

went like this:

root@SkynetB:/opt/tailmox# ./tailmox.sh
Checking if Proxmox is installed...
Proxmox VE 8.x detected.
Checking for required dependencies...
Starting Tailscale with --advertise-tags 'tag:tailmox'...
Some peers are advertising routes but --accept-routes is false
Waiting for Tailscale to come online...
This host's Tailscale IPv4 address: 100.100.111.11
This host's Tailscale MagicDNS name: dustbowl-proxmox.tail58de2.ts.net
Tailscale certificate services already cloned.
Proxmox is installed, continuing...
Tailscale is installed, continuing...
jq is installed, continuing...
Attempting to generate Tailscale certificate for dustbowl-proxmox.tail58de2.ts.net
Public cert unchanged at dustbowl-proxmox.tail58de2.ts.net.crt
Private key unchanged at dustbowl-proxmox.tail58de2.ts.net.key
Setting custom certificate files
Restarting pveproxy

finger prints go here

Scheduling Proxmox certificate renewal...
Checking crontab...
0 4 1 * * /opt/tailscale-cert-services/proxmox-cert.sh;
Checking if all tailmox peers are online...
All tailmox peers are online.
Ensuring the local node can ping all other nodes...
Pinging 100.100.101.100...
Successfully pinged 100.100.101.100.
All Tailmox peers are reachable via ping.
Reporting peer latency...
Calculating average latency for 100.100.101.100 (25 pings with an interval of 0.1 seconds)...
Average latency to 100.100.101.100: 82.187 ms
Checking if TCP port 8006 is available on all nodes...
Checking TCP port 8006 on suijin (100.100.101.100)...
TCP port 8006 is not available on suijin (100.100.101.100).
Some peers have TCP port 8006 unavailable. Please check the network configuration.

[willjasen]

In regards to the two issues of checking for git and ensuring that the script is running from the /opt/tailmox directory, I have committed these fixes into the "proxmox-v9" branch and hope to have the next version of tailmox (v1.1.0) released soon.

[willjasen]

I have committed a fix for the third issue of:

Checking TCP port 8006 on suijin (100.100.101.100)...
TCP port 8006 is not available on suijin (100.100.101.100).
Some peers have TCP port 8006 unavailable. Please check the network configuration.

The tailmox.sh script was looking for other Tailscale peers not including the host where the script is running for the tag of "tailmox", so upon the first clean run, that error was being encountered. The script will now continue if there are no peers so that it proceeds to ask if a new cluster should be created.

In accordance, the revert_test_vms.sh can now remove existing hosts in Tailscale with the tag of "tailmox" by using the "--api-key" parameter. This should only be needed when testing.