mirror of
https://github.com/balena-io/balena-cli.git
synced 2024-12-24 15:56:38 +00:00
129 lines
6.5 KiB
Markdown
129 lines
6.5 KiB
Markdown
# FAQ & Troubleshooting
|
|
|
|
This document contains some common issues, questions and answers related to the balena CLI.
|
|
|
|
## Where is my configuration file?
|
|
|
|
The per-user configuration file lives in `$HOME/.balenarc.yml` or `%UserProfile%\_balenarc.yml`, in
|
|
Unix based operating systems and Windows respectively.
|
|
|
|
The balena CLI also attempts to read a `balenarc.yml` file in the current directory, which takes
|
|
precedence over the per-user configuration file.
|
|
|
|
## How do I point the balena CLI to staging?
|
|
|
|
The easiest way is to set the `BALENARC_BALENA_URL=balena-staging.com` environment variable.
|
|
|
|
Alternatively, you can edit your configuration file and set `balenaUrl: balena-staging.com` to
|
|
persist this setting.
|
|
|
|
## How do I make the balena CLI persist data in another directory?
|
|
|
|
The balena CLI persists your session token, as well as cached images in `$HOME/.balena` or
|
|
`%UserProfile%\_balena`.
|
|
|
|
Pointing the balena CLI to persist data in another location is necessary in certain environments,
|
|
like a server, where there is no home directory, or a device running balenaOS, which erases all
|
|
data after a restart.
|
|
|
|
You can accomplish this by setting `BALENARC_DATA_DIRECTORY=/opt/balena` or adding `dataDirectory:
|
|
/opt/balena` to your configuration file, replacing `/opt/balena` with your desired directory.
|
|
|
|
## After burning to an sdcard, my device doesn't boot
|
|
|
|
- The downloaded image is not complete (download was interrupted).
|
|
|
|
Please clean the cache (`%HOME/.balena/cache` or `C:\Users\<user>\_balena\cache`) and run the command again. In the future, the CLI will check that the image is not complete and clean the cache for you.
|
|
|
|
## I get a permission error when burning to an sdcard
|
|
|
|
- The SDCard is locked.
|
|
|
|
### I get EINVAL errors on Cygwin
|
|
|
|
The errors look something like this:
|
|
|
|
```
|
|
net.js:156
|
|
this._handle.open(options.fd);
|
|
^
|
|
Error: EINVAL, invalid argument
|
|
at new Socket (net.js:156:18)
|
|
at process.stdin (node.js:664:19)
|
|
at Object.Interface.createInterface (C:\cygwin\home\Juan Cruz Viotti\Projects\balena-cli\node_modules\inquirer\node_modules\readline2\index.js:31:43)
|
|
at PromptUI.UI (C:\cygwin\home\Juan Cruz Viotti\Projects\balena-cli\node_modules\inquirer\lib\ui\baseUI.js:23:40)
|
|
at new PromptUI (C:\cygwin\home\Juan Cruz Viotti\Projects\balena-cli\node_modules\inquirer\lib\ui\prompt.js:26:8)
|
|
at Object.promptModule [as prompt] (C:\cygwin\home\Juan Cruz Viotti\Projects\balena-cli\node_modules\inquirer\lib\inquirer.js:27:14)
|
|
```
|
|
|
|
- Some interactive widgets don't work on `Cygwin`. If you're running Windows, it's preferrable that you use `cmd.exe`, as `Cygwin` is [not official supported by Node.js](https://github.com/chjj/blessed/issues/56#issuecomment-42671945).
|
|
|
|
## I get `Invalid MBR boot signature` when configuring a device
|
|
|
|
This error, accompanied with something like: `Expected 0xAA55, but saw 0x29FE` usually indicates a corrupted device operating system image in the cache, due to bad a internet connection during the download process.
|
|
|
|
Try clearing the cache with the following command and trying again:
|
|
|
|
```sh
|
|
$ rm -rf $HOME/.balena/cache
|
|
```
|
|
|
|
Or in Windows:
|
|
|
|
```sh
|
|
> del /s /q %UserProfile%\_balena\cache
|
|
```
|
|
|
|
## I get `EACCES: permission denied` when logging in
|
|
|
|
The balena CLI stores the session token in `$HOME/.balena` or `C:\Users\<user>\_balena` in UNIX based operating systems and Windows respectively. This error usually indicates that the user doesn't have permissions over that directory, which can happen if you ran the balena CLI as `root`, and thus the directory got owned by him.
|
|
|
|
Try resetting the ownership by running:
|
|
|
|
```sh
|
|
$ sudo chown -R <user> $HOME/.balena
|
|
```
|
|
|
|
## Broken line wrapping / cursor behavior with `balena ssh`
|
|
|
|
Users sometimes come across broken line wrapping or cursor behavior in text terminals, for example when long command lines are typed in a `balena ssh` session, or when using text editors like `vim` or `nano`. This is not something specific to the balena CLI, being also a commonly reported issue with standard remote terminal tools like `ssh` or `telnet`. It is often a remote shell configuration issue (files like `/etc/profile`, `~/.bash_profile`, `~/.bash_login`, `~/.profile` and the like), including UTF-8 misconfiguration, the use of unsupported ASCII control characters in shell prompt formatting (e.g. the `$PS1` env var) or the output of tools or log files that use colored text. The issue can sometimes be fixed by resizing the client terminal window, or by running one or more of the following commands on the shell:
|
|
|
|
```sh
|
|
export TERMINAL=linux
|
|
stty sane
|
|
shopt -s checkwinsize
|
|
bind 'set horizontal-scroll-mode off'
|
|
```
|
|
|
|
Terminal multiplexer tools like GNU `screen` or `tmux` are sometimes reported to fix the issues, though at other times they are reported as the _cause_ of the problem. They have their own configuration files to take into account.
|
|
|
|
Further reference:
|
|
* https://stackoverflow.com/questions/1133031/shell-prompt-line-wrapping-issue
|
|
* https://superuser.com/questions/46948/any-way-to-fix-screens-mishandling-of-line-wrap-maybe-only-terminal-app
|
|
* https://unix.stackexchange.com/questions/105958/terminal-prompt-not-wrapping-correctly
|
|
* https://unix.stackexchange.com/questions/529377/terminal-long-line-wrapping
|
|
* https://github.com/microsoft/WSL/issues/1436
|
|
|
|
If nothing seems to help, consider also using a different client-side terminal application:
|
|
* Linux: xterm, KDE Konsole, GNOME Terminal
|
|
* Mac: Terminal, iTerm2
|
|
* Windows: PowerShell, PuTTY, WSL (Windows Subsystem for Linux)
|
|
|
|
## "Docker seems to be unavailable" error when using Windows Subsystem for Linux (WSL)
|
|
|
|
When running on WSL, the recommendation is to install a CLI release for Linux, like the standalone
|
|
zip package for Linux. However, commands like "balena build" that contact a local Docker daemon,
|
|
like the Docker Desktop for Windows, will try to reach Docker at the Unix socket path
|
|
`/var/run/docker.sock`, while Docker Desktop for Windows uses a Windows named pipe at
|
|
`//./pipe/docker_engine` (which the Linux CLI on WSL cannot use). A solution is:
|
|
|
|
- Open the Docker Desktop for Windows settings panel and tick the checkbox _"Expose daemon on tcp://localhost:2375 without TLS"._
|
|
- On the WSL command line, set an env var:
|
|
`export DOCKER_HOST=tcp://localhost:2375`
|
|
Alternatively, use the command-line options `-h 127.0.0.1 -p 2375` for commands like `balena build` and `balena deploy`.
|
|
|
|
Further reference:
|
|
|
|
- https://techcommunity.microsoft.com/t5/Containers/WSL-Interoperability-with-Docker/ba-p/382405
|
|
- https://forums.docker.com/t/wsl-and-docker-for-windows-cannot-connect-to-the-docker-daemon-at-tcp-localhost-2375-is-the-docker-daemon-running/63571/12
|