ssh

Created: 2016-02-29 11:15:20 -0800 Modified: 2023-12-07 08:48:14 -0800

Config (reference)

ssh

If you’re sick of doing “ssh-agent bash” followed by “ssh-add” (which you need to redo every time you start a new shell), then you should use ~/.ssh/config.

In short, you just need to do these steps:

vim ~/.ssh/config
Type this in

Host bitbucket.org

HostName bitbucket.org

User <your user account>

IdentityFile ~/.ssh/path_to_priv_key

Notes:

“Host” is what you would use to SSH. So, for example, if you set “Host chickens” at the top, then you could log into bitbucket by typing “ssh chickens”. If HostName isn’t set, then “Host” is the address of the remote. It accepts patterns, so if you want to use a “localvms” keypair to log in to 192.168 addresses, then type “192.168.._“.
You don’t have to specify a username if you’ll be logging in with the same one that you’re already using.
Even on Windows, if you modify ~~/.ssh/config, you’re allowed to use "~~" to represent %userprofile%.

ssh-agent (reference)

Note: on macOS, to not have to type passwords, follow these instructions.

Starting manually:

eval ”$(ssh-agent -s)“

ssh-add ~/.ssh/id_rsa

Starting from bash_profile:

function ensure_ssh_agent_running() {
env=~/.ssh/agent.env

agent_load_env () { test -f "$env" && . "$env" >| /dev/null ; }

agent_start () {
    (umask 077; ssh-agent >| "$env")
    . "$env" >| /dev/null ; }

agent_load_env

# agent_run_state: 0=agent running w/ key; 1=agent w/o key; 2= agent not running
agent_run_state=$(ssh-add -l >| /dev/null 2>&1; echo $?)

if [ ! "$SSH_AUTH_SOCK" ] || [ $agent_run_state = 2 ]; then
    agent_start
    ssh-add
elif [ "$SSH_AUTH_SOCK" ] && [ $agent_run_state = 1 ]; then
    ssh-add
fi

unset env
}
ensure_ssh_agent_running

Using multiple instances of the same HostName in ~/.ssh/config

Note: if you’re doing this for GitHub, consider looking at [[gh#Using gh as a credential manager]].

Host bitbucket.org
  User onaimada
  IdentityFile ~/.ssh/bitbucket_botlandinfra_deploy_key

Host test
  HostName bitbucket.org
  User onaimada
  IdentityFile ~/.ssh/bitbucket_botlandacctsrv_deploy_key

Then, when you’re cloning a repo, you can type out this command:

git clone git@test:onaimada/botlandrestserver.git

After you’ve cloned the repo, git pull should work normally. If you ever change the name of test, make sure to reflect that in your .git/config file.

To test this setup, you can use ssh -T, e.g.:

❯ ssh -T git@adamlearns.github.com

Hi AdamLearns! You’ve successfully authenticated, but GitHub does not provide shell access.

❯ ssh -T git@github.com

Hi Adam13531! You’ve successfully authenticated, but GitHub does not provide shell access.

Make sure that your remote uses the host that you have set, e.g.

❯ git remote -v

origin git@adamlearns.github.com:AdamLearns/Abbott.git (fetch)

If either identity is wrong on macOS, then run these commands (and possibly add them to your shell’s start-up script):

ssh-add --apple-use-keychain ~/.ssh/github_adam13531
ssh-add --apple-use-keychain ~/.ssh/github_adamlearns

When working with pull requests, the pushremote will need to be changed (e.g. here, I’m changing github.com → adamlearns.github.com): git config branch.smu-readme.pushremote git@adamlearns.github.com:smu4242/JumpRoyale.git

Troubleshooting

Running a command over SSH

I kept getting “No such file or directory”. Turns out it was a quotation marks issue.

I had:

ssh -i %keyLocation% admin@%overseerIp% ‘/usr/bin/mysqldump -u root -h db.bot.land —password=%dbPassword% botland ^> %dumpFileName%‘

I needed it to be:

ssh -i %keyLocation% admin@%overseerIp% “/usr/bin/mysqldump -u root -h db.bot.land —password=%dbPassword% botland > %dumpFileName%“

(notice: the ”>” is no longer escaped)

Troubleshooting

ssh: connect to host <IP_ADDRESS_REMOVED> port 22: Connection timed out

This is just a generic SSH message that usually indicates that the IP address is wrong or the host is offline, but this happened to me when trying to SSH from a Windows machine to my Raspberry Pi when my Pi was definitely online (because I could access the service it provided from outside of my home network) and the IP address that I was using internally was the same one that my router reported.

I tried a bunch of stuff with pinging and nmap and eventually figured out that it was a Windows Firewall problem from my PC. I only figured this out after ruling out a million other possibilities. Turning the firewall off and on again seemingly fixed everything (this is because the firewall is stateful, so restarting it cleared the state, although I never figured out how the rule got set in the first place).