Chrome OS - Crostini fails to start with vmshell error 20

chrome-os-crostini-error-20.md

Chrome OS - Crostini fails to start with vmshell error 20

This I found out to be an sshd problem. Somehow the systemd service cros-sftp of the container was not working.
All the below steps are from crosh terminal.

Start the container from crosh

From crosh terminal (Alt+Ctrl+T):

vmc start termina

Wait for sometime till the green color prompt comes. Then to list your containers, type lxc list. In most cases it is only one container named "penguin", so you can issue the below command:

lxc start penguin

Finally login to the root shell of the container:

lxc exec penguin -- /bin/bash

Sources:
https://superuser.com/questions/1617731/how-to-fix-the-terminal-and-linux-files-for-error-20-on-a-chromebook
https://www.reddit.com/r/Crostini/comments/cij85h/how_to_login_to_container_from_crosh/
Login to root without password: https://www.reddit.com/r/Crostini/comments/e8749l/what_is_the_default_password_for_the_initial/
Extensive material: http://pauldone.blogspot.com/2019/12/my-notes-chromeos-crostini.html

Find the problem.

Most problems can be found out by reading the journal, type this to check it journalctl -xe | more. Alternately from crosh as well we can check the problems from the file home/chronos/user/log/chrome or var/log/messages
Source: https://www.reddit.com/r/Crostini/comments/m0r9dg/quick_tip_more_info_on_crostini_error_codes_is/
In our case it is known to be a problem of the cros-sftp service which is probably dependent on the sshd service.
Check the service:

systemctl status cros-sftp

If the service is disabled, enable it.

systemctl enable cros-sftp

Start the service

systemctl start cros-sftp

Failing with cros-sftp.service: Start request repeated too quickly

Edit the below file:

vim /etc/systemd/system.conf

Uncomment the line (remove the starting #): DefaultStartLimitBurst. Set the value to a large number like 50.

DefaultStartLimitBurst=50

Save it. The restart the systemd daemon.

systemctl daemon-reload

Again try starting the cros-sftp service

systemctl start cros-sftp

Sources:
https://www.reddit.com/r/Crostini/comments/8p351n/linux_files_mount_error_anyone_else_having/
https://serverfault.com/questions/845471/service-start-request-repeated-too-quickly-refusing-to-start-limit

Fails again with ExecStartPre=/usr/sbin/sshd -t -f /dev/.ssh/sshd_config (code=exited, status=25

Try to start the mentioned service manually:

/usr/sbin/sshd -t -f /dev/.ssh/sshd_config

If it fails with the error Privilege separation user sshd does not exist we need to create the sshd user.

Create sshd user if not present.

Source: http://linuxhostingsupport.net/blog/starting-sshd-privilege-separation-user-does-not-exist#:~:text=Starting%20sshd%3A%20Privilege%20separation%20user%20does%20not%20exist,-The%20error%20message&text=It%20indicates%20that%20the%20user,sshd'%20user%20on%20the%20server.&text=If%20it's%20a%20dedicated%20server,in%20before%20the%20problem%20occurred.

vim /etc/passwd

Add the line: sshd:x:74:74:Privilege-separated SSH:/var/empty/sshd:/sbin/nologin then save and exit.

vim /etc/group

Add the line: sshd:x:74: then save and exit.
Now try starting the service, it should not complain anymore.

/usr/sbin/sshd -t -f /dev/.ssh/sshd_config

Also check the cros.sftp service, it should be green.

systemctl status cros.sftp

All done, stop the container from crosh and start it from chrome os Terminal app.

Press Ctrl+D to logout of the container root shell.
In the green prompt type:

lxc stop penguin

Press Ctrl+D again to logout to yellow crosh shell. Type:

vmc stop termina

After this the terminal can be closed. Now open the Terminal app from chrome os launcher and all should work fine.