> ## Documentation Index
> Fetch the complete documentation index at: https://docs.axottle.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Diagnostics of problems

> How to localize a login failure, SSH, SFTP, agent, playbook, schedule, or Telepúrte update.

First, determine at what stage the failure occurred. Entering the panel, visibility of the node, SSH connection, file operation and agent assignment check different conditions. The success of one stage does not confirm the others.

## Quick map.

| Symptom | Start with | |
| - | - | - |
| Panel does not open | Telepúrte Service, HTTPS, DNS, reverse proxy and `public_url` | |
| Local Login Operating, SSO No | User Axottle, MFA, Communication and Fresh Synchronization | |
| Node visible, terminal unavailable | Host key, SSH preparation, grant, system login and bastion | |
| Online Agent, Operation Prohibited by | Action Support, Local Listings and Linux | Rights |
| SSH works, Web SFTP no | SFTP, Web route, root directories and file bans | |
| Editor does not save | Modification of the original file, size, rights and result of the | validation |
| Playbook remains in line with | Agent, Policy, Author Rights and | Assignment Term |
| Schedule Active, Launch No | Time Zone, Offensive History, Crossing and Target Availability | |
| Terminal Record No | Viewing rights, shelf life, cleaning and session type | |
| Update Not Completed | Update Log and Recovery Status | |

## Panel and entrance

On the server, check the service and log:

```bash theme={null}
sudo systemctl status telepurte --no-pager
sudo journalctl -u telepurte --since '30 minutes ago' --no-pager
sudo -u telepurte telepurte check --config /etc/telepurte/config.json
```

Convert the exact HTTPS-origin to `public_url`, request domain, and save Host to reverse proxy. In the absence of `public_url`, web login is not enabled. Do not create a new database or master key to fix the login error.

For the local user, check the lock, log-in time and password, TOTP device time and attempt limit. For Axottle, external log-in activity, MFA and user list freshness. If the local owner has lost both factors, use [CLI restore ](https://docs.axottle.com/en/telepurte/backups.md#восстановление-владельца).

## Node and SSH

Open the node card and successively check the presence, address, key, SSH availability and authorized system login. Changing the IP or host key requires a new trust check; the old resolution should not be imperceptibly transferred to another machine.

Then check the bastion and the transport. You need a direct route to the sshd. You need a live tunnel and the right local port for the agent. You don't automatically switch to the direct SSH if the tunnel fails.

If the rejection applies only to one user, check all roles, direct grant, deadline, project area, bans and session limits. With a personal login, you also need a confirmed registration preparation on this node.

## Agent and playbook

Online does not guarantee support for a new command. Verify the version and the ability to perform the desired action. For systemd, check the exact unit, for the file, the full path, for the package, the exact name. Permission in the policy does not replace OS rights and agent service limitations.

`policy_changed` requires a new preview. `policy_denied` fixes a specific resolution. For `script_failed`, check the code and output; for `ansible_unavailable`, the Ansible installed on the node itself.

`execution_uncertain` means that the confirmation of the change is lost. First read the actual state of the server, then decide to repeat. Don't run the entire fleet again until the partial result is disassembled.

## Files and editor

Compare the system user in SSH and files. Check the authorized root directory, action, path, Linux rights and object type. Both paths are important for renaming, and void for the directory when deleted.

If a file has changed after opening, save your edits separately, re-read the current version, and combine the changes. A save error is not a permission to replace someone else's version. The 16 MiB preview and 8 MiB editor limit is less than the file transfer limit possible.

With a running terminal and broken files, check for reverse proxy routes of the file Web gateway, not just the WebSocket terminal.

## Schedule

Open the offensive history, not just the general list. `skipped_overlap` stands for the previous mission still working, `skipped_missed` stands for the handled pass, `skipped_window` stands for the closed window. `recipe_changed` requires a new schedule for the current version.

After changing the team composition, check the exact saved goals: the old schedule does not pick up new nodes. If one target is not available, the preliminary check of the entire offensive can be completed without creating a partial task.

## What to save for parsing

Write down the time and time zone, the panel/agent/bastion version, the node or task ID, the exact error code and step. Save the results of each target for the task, update the transaction log. Eliminate passwords, tokens, guest links and private keys before transferring materials.

## Verification

After the fix, repeat the failed path with the same user and node: input, SSH, a specific file action, or a new task preview. Then check for the final result and no new errors. Don't limit yourself to the updated green status.

## Rollback

Return the last modified setting if it caused the problem, and repeat the check. Before changing the keys, base, or identity, save a copy and select the standard restore. The actions performed on the servers are returned separately from the panel setting.

Check the versions of all the components involved and the actual settings of your installation.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.