Open NiceGuyIT opened 1 year ago
The "symptom" is the Connect greyed out in MeshCentral. journalctl
reports the following.
Aug 21 07:27:44 ns-v18-tactical meshcentral[2486]: Agent bad web cert hash (Agent:e583a0455e != Server:1a57156b69 or 1a57156b69), holding connection (127.0.0.1:36028).
Aug 21 07:27:44 ns-v18-tactical meshcentral[2486]: Agent reported web cert hash:e583a0455e0e7378449c502f4431fe8ddda4cd86f24bbe0b806cb3eef713ebd094aec202b4744692daf2f91040a2f5aa.
How would you suggest documenting it? Disclaimer in the unsupported first page?
My PR essentially addresses this issue. The error isn't mentioned per se but if users follow my writeup to a T they shouldn't see this error.
My PR essentially addresses this issue
Which PR?
Commit 716c0fe introduced the "management" of
/etc/hosts
to handle the scenario of cloud init files. This change assumes a reverse proxy is not used and will cause problems for people using MeshCentral with reverse proxies.In my environment,
mesh.example.com
points to the reverse proxy to get the TLS cert, not127.0.0.1
. Since$meshdomain
is included in the checks/fix, updating might "break" the (unsupported) reverse proxy configuration.I don't expect the code to be updated to account for unsupported configurations. This issue is to update the documentation.