I intended to post in defense of manpages. I love manpages. I think most open-source manpages are high-quality, and a few are really outstanding.
"It's easier to ask an AI" can be true without implying that manpages are bad.
However, "man" the tool does have issues, and one of them bit me just now.
So, I didn't know about openssh client escapes like ~?. I thought, "surely that's in the manpage?" I opened the manpage (in less) and searched for "\~\?". No hits.
Of course, escape characters are documented in the manpage, and the string "~?" does appear. Why didn't search find it? Because man, in its infinite wisdom, decided to render every instance of "~" as some bizarre unicode not-tilde, which is visually similar but totally impossible to grep for.
This has also bitten me in the past with dash. DASH. A character that is critically important when documenting invocation options. man loves to convert it into something that looks like dash, prints like dash, but doesn't come up in search.
I'm sure there is a way to turn this "feature" off, and I'm about to spend a bunch of time figuring out what it is. But this is documentation for command-line tools. Silently destroying our ability to grep it should NOT be the default.
Manpages are both good and awful. If you already know how to use the command and you just need a refresher on how to enable the diagnostic option for wolves, they're fine. OTOH if you don't know how to use the command they're terrible, just an endless catalogue of incomprehensible options scrolling past. Problem is that that's what Dennis and Ken created in 1929 for use on ASR-33s and it hasn't been updated since. 99% of the time when I need to use some new unfamiliar command I go to whatever DDG pops up, usually StackOverflow, for a guide on usage.
I don't know about it changing your characters, but you can search for a string verbatim by pressing ctrl-r at the start of the search pattern, like this:
yeah it sounds like your environment is doing some weird mangling of the encoding. Maybe consider adding an 'alias man="man -E ascii"' into your .bashrc
You might also consider seeing if it's any different in WSL. Or you could, you know, just use Linux ;P
I've been using the ~. shortcut for a while, but somehow escaped learning about the help menu.
Another neat thing I noticed while playing with it just now: there's an option to enter ~ twice to send a literal ~, but usually you don't have to do this when typing something like 'ls ~' in a regular session. Not only does the ~ have to be the first character on a line to start an escape sequence, but typing on a line, backspacing all the way to the start and then typing ~ also sends a literal tilde. It only triggers the escape sequence if the ~ is the chronologically first character after a newline (or first in the session), which is an unlikely thing to type into a shell in a normal session. Good choice of UI, both the character and the state machine.
> typing on a line, backspacing all the way to the start and then typing ~ also sends a literal tilde
for the younger readers, yes, because in terminal echo mode, "backspacing" does not clear your terminal line buffer, those characters backspaced are already sent on the line. if you ever seen a misconfigured terminal, it hints what's going on, like:
user@host$ ls ~/^?^?^?^?^?~/a.out
^? is backspace's control char.
that is ssh watches what you type, not what is on the screen (terminal).
The drawback is that if you think your session is hanging and want to bail with ~., you have to press enter, which might actually make it to the server and execute something.
I've been using SSH for ~15 years and never knew about these escape sequences. I'm eagerly awaiting my next hung session so that I can test `~.`. It's much nicer than my current approach of having to close that terminal window.
If hung SSH connections are common it's likely due to CGNAT which use aggressively low TCP timeouts. e.g. I've found all UK mobile carriers set their TCP timeout as low as 5 minutes. The "default" is supposed to be 2 hours, you could literally sleep your computer, zero packets, and an SSH connection would continue to work an hour later, and generally speaking this is still true unless CGNAT is in the way.
If you are interested there are a few ways you can fix this:
Easiest is to use a VPN, because the VPN's exit node becomes the effective NAT they usually have normal TCP timeouts due to being less resource constrained. Another nice benefit of this method is you can move between physical networks and your connection doesn't die... If you use Tailscale then you already have this in a more direct way.
Another is to tune the tcp_keepalive kernel parameters. Lowering the keepalive timeout to be less than the CGNAT timeout will cause keepalive probes to prevent CGNAT from dropping the connection even while your SSH connection is technically idle. For Linux I pop these into /etc/sysctl.d/z.conf, I have no idea for Windows or Mac:
This is really a misuse of these settings, they are supposed to be for checking TCP connections are still alive and clearing them up from the local routing table. Instead the idea is to exploit the probes by sending them more frequently to force idle connections to stay alive in a CGNAT environment (dont worry the probes are tiny and still very infrequent).
_time=240 will send a probe after 4 mins of idle connection instead of the default 2 hours, undercutting the CGNAT timeout. _intvl=60 and _probes=120 mean it will send 120 probes 60 seconds apart (2 hours worth) before considering the connection dead. This will keep it alive for at least 2 hours, but also allows us to have the best of both worlds so that under a nice NAT it keeps the old behaviour, e.g if I temporarily lose my network the SSH connection is still valid after 2 hours, but under CGNAT it will at least not drop the connection after 5 mins so long as I keep my computer on and don't lose the network.
There are also some SSH client keepalive settings but I'm less familiar with them.
Check Mosh. It supports these kind of cuts and it will reconnect seamlessly. It will use far less bandwidth too.
I successfully tried it with a 2.7 KBPS connection.
Depends on whether your sockets survive that, though. Especially on Wi-Fi, many implementations will reset your interface when sleeping, and sockets usually don't survive that.
Even if they do, if the remote side has heartbeats/keepalive enabled (at the TCP or SSH level), your connection might be torn down from the server side.
Yes, by generally I really mean all the defaults are pretty permissive, but I understand some people tune both TCP and SSH on their servers to drop connections faster because they are worried about resource exhaustion.
But if you throw up a default Linux install for your SSH box and have a not-horrible wifi router with a not-horrible internet provider then IME you can sleep your machine and keep an SSH connection alive for quite some time... I appreciate that might be too many "not-horrible" requirements for the real world today though.
Yes, this makes your connection more likely not survive client suspends. (ClientAliveInterval, which makes the server ping the client, will make it fail almost certainly, since the server will be active while the client is sleeping.)
Well, for different reasons, but you have similar issues with IPv6 as well. If your client uses temporary addresses (most likely since they're enabled by default on most OS), OpenSSH will pick one of them over the stable address and when they're rotated the connection breaks.
For some reason, OpenSSH devs refuse to fix this issue, so I have to patch it myself:
--- a/sshconnect.c
+++ b/sshconnect.c
@@ -26,6 +26,7 @@
#include <net/if.h>
#include <netinet/in.h>
#include <arpa/inet.h>
+#include <linux/ipv6.h>
#include <ctype.h>
#include <errno.h>
@@ -370,6 +371,11 @@ ssh_create_socket(struct addrinfo *ai)
if (options.ip_qos_interactive != INT_MAX)
set_sock_tos(sock, options.ip_qos_interactive);
+ if (ai->ai_family == AF_INET6 && options.bind_address == NULL) {
+ int val = IPV6_PREFER_SRC_PUBLIC;
+ setsockopt(sock, IPPROTO_IPV6, IPV6_ADDR_PREFERENCES, &val, sizeof(val));
+ }
+
/* Bind the socket to an alternative local IP address */
if (options.bind_address == NULL && options.bind_interface == NULL)
return sock;
This is a very common misconception. The issue is not IPv4 or CGNAT, it's stateful middleboxes... of which IPv6 has plenty.
The largest IPv6 deployments in the world are mobile carriers, which are full of stateful firewalls, DPI, and mid-path translation. The difference is that when connections drop it gets blamed on the wireless rather than the network infrastructure.
Also, fun fact: net.ipv4.tcp_keepalive_* applies to IPv6 too. The "ipv4" is just a naming artifact.
Mobile carriers usually have stateful firewalls for IPv6 as well (otherwise you can get a lot of random noise on the air interface, draining both your battery and data plan), so it's an issue just the same.
The constrained resource there is only firewall-side memory, though, as opposed to that plus (IP, port) tuples for CG-NAT.
Have been using that weekly since probably 20 years. Will change your life :)
My other favourite is I very often SSH with -v to figure out why the connection is hanging, you rapidly figure out if DNS is failing, the TCP connection doesn't open, it does open but no traffic flows at all or it opens and SSH negotiation starts but never finishes. You can learn a lot just from this about what is wrong.
And of course, you can use the ~v / ~V commands (as listed in the ~? menu) to increase/decrease verbosity after the connection is established.
That lets you `ssh -vvvv` to a host then once you've figured out the issue use ~V to decrease verbosity so that debug messages don't clutter your shell.
Also helps with auth failures, I've used it several times with co-workers who can't figure out why their ssh key isn't working. It lists the keys out and some extra information.
> It's much nicer than my current approach of having to close that terminal window.
You can also just kill the ssh process (say from another terminal). That way you get to keep your terminal window. And this works with everything "blocking" your terminal, not just ssh.
I last used this menu about 20 years ago when a dialup modem was the only way to roll, and have pretty much forgotten about it since the days of always-on direct to the desktop TCP/IP ..
Does not for me, not even with busybox sh and no funky escape codes in PS1 at all. It does with cat or yes running, so just something being output is not the problem… Hm.
So you're saying 'fish' intercepts it on the far end? The ssh server on the far end shouldn't be sending it to 'fish' until it knows what's coming next.
Is this a current-ish version of OpenSSH or some other client/server?
EDIT Interesting! I tested it with fish and it does indeed intercept it! Wonder how that works.
Same with me, I'll still instinctively go for ~. when a connection has hung / dropped (usually because of a NAT via a rebooted firewall), but never even considered how ~ doesn't normally cause an issue. Never knew it had to be immediately following a newline. Also never knew about the other options, ~^Z in particular looks useful.
I wonder if anyone still remembers the ctrl-[ sequence in telnet. I think I only ever used the quit command in that though.
No need – your local ssh client, which interprets the escape sequence, doesn't have any understanding of whatever mode the remote session might currently be in.
The point of the return is to prime it to accept the start of a new escape sequence. Presumably the idea is that `~.` is not completely unlikely to occur as part of text entered remotely, but less so at the beginning of a new line.
I typically just create a "new" connection in a separate tab when I want to add tunneling.
I put new in quotes because I use another little-known feature, "ControlMaster". Multiplexes multiple connections into one, it makes making " new" sessions instant (can also be configured to persist a bit after disconnecting). Also useful for tab-completing remote paths. It does not prompt for authentication again, though. And it's a bit annoying when the connection hands (can be solved with ssh -o close, IIRC).
> I use another little-known feature, "ControlMaster". Multiplexes multiple connections into one, it makes making " new" sessions instant
Is this what secureCRT used as well? I remember this being all the rage back when I used windows, and it allowed this spawn new session by reusing the main one.
I'm using that as well but had issues with tunneling where it creates the tunnel in the background and terminates and so you might not know the random port it assigned or I couldn't figure out how to un-tunnel it and tunnel again to the same port. Just bypassed the control master then.
"secret" not in the sense that it's hidden, but that most people won't know about it. Because approximately nobody actually reads man pages in their entirety, they just get in to find out how a specific flag works and then get out.
the difference is that knowing 2^8 is generally not useful to people who don't know it
this here is something that's pretty useful to most ssh users, yet seldom spoken of
a better analogy would be comparing it to calling a very good, but not well-known restaurant a secret place - using the word to mean a hidden gem rather than an intentionally hidden secret
You think that one is cool, go check out "~?" in IPMI "sol activate". From there you can deliver a serial break to the kernel, which then gets you to a third rarely seen menu from the kernel's console, which allows you to do kernel debugging of various sorts.
I use it when I need to crash a kernel on purpose to test kdump over the network.
You can also send commands to the simulated console of a VM under libvirt with "virsh connect". But I don't think you can send a break to the kernel with that.
The ~ (tilde) escape character in UNIX cu (call UNIX) did the same thing in the 1970s, when the remote system was on the other side of a telephone modem or other serial port. I think it's safe to say that ssh is a descendant of cu. (I don't mean that cu printed a help menu, I mean it had the tilde escape.)
it starts with a pretty common char, but almost never gets in the way to the point I forget it exists. Meanwhile docker -t uses ^P which I use all the time for history instead of arrow keys. It's possible to configure it, but it's not worth the hassle on servers. Really, really annoying.
This is not specific to ssh. Telnet and rlogin have similar things with ~ as the escape character. Back in the day it was common to send BREAK and other escape sequences when you were hardwired.
Wow, never knew this... That said I'm not sure if I'll remember using it as my muscle memory is already trained to kill hung processes via ctrl-b s, ctrl-k, ctrl-b x (in tmux).
Unlike CTRL ], at least ~. doesn't require that I press two modifiers at the same time ... CTRL ALTGR $. Because people who define those kinds of shortcuts never consider how they might work on non-QWERTY layouts.
Don't tell people this. In a minute you'll realize that this menu can only mean that ssh runs multiple channels, and so you can start up other things over an established ssh connection.
Then the firewall guys will realize that they really can't allow ssh ... let's just not go there. Working at a bank is annoying enough already.
When your firewall guys "block" ssh (or anything else), all you need to do is run your ssh server on port 443 and then `ssh -p 443 user@host`.
Running on 443 means it will blend in with https traffic, making it much more difficult for them to detect and block (I've never actually seen anywhere that can detect/block it - you'd need to be doing deep packet inspection on encrypted traffic) :)
Not on the same IP at the same time, but you could always add a second IP for ssh. Just configure sshd to only listen on one IP and <webserver> to listen on the other.
I have noticed it while running ~/bin/some_command. The ~ doesn't echo until I also type the /. It doesn't cause any misbehavior because there is no binding for ~/ but can be slightly surprising.
I find it odd that you would have commands in ~/bin but not have it be the highest priority in your PATH. I use ~/.local/bin, but would never type it because i wouldn't have bins that overlap shell commands and no other path would have priority.
Usually, it is. IIRC, this was when I was just setting up my environment on a new host, after I had populated ~/bin but before I restarted my shell to pick up PATH modifications.
I'd guess this is because it only works in ssh PTY sessions. So it would have no effect on tunneling or when piping arbitrary data through ssh to a non-interactive remote command (unless you use the -t switch to force PTY allocation even when stdin is not a TTY).
I was never able to properly parse large man pages, I'm so happy that llms can now prepare half a usable command without spending an hour reading a time without a single usage example.
I've also been running (neo)vim as a manpager. You get the same features as with vim (like easily copying text or opening referenced files/other manpages without using the mouse), but neovim also parses the page and creates a table of contents, which can be used for navigation within the page. It doesn't always work perfectly, but is usually better than nothing.
You can also run commands when a user authenticates, grab their keys from github.com/username.keys, validate they're a user in a specific github group, then let them connect by outputing the keys, otherwise nothing to deny them access.
It's really great for ops teams where you want to give ssh access and manage it from github teams without needing a complex system.
Honest question, why is ProxyCommand `fun`? What do I get out of ProxyCommand that i do not get out of setting the correct order for ProxyJump and doing an ssh finalhost -- domy --bidding?
ProxyJump is a newer functionality. There used to be only ProxyCommand. ProxyJump is a shortcut for the usual way to use ProxyCommand to connect through a bastion host but ProxyCommand is more flexible. For example with ProxyCommand you can run any command to connect to the remote host. ProxyJump only connects over ssh. I think I replaced all my ProxyCommand with ProxyJump because I don't need much else than the normal use case.
You can get a lot more out of ProxyCommand. For example, you can run SSH over non-IP protocols, such as serial, Bluetooth RFCOMM for embedded boards, or vsock for virtual machines without networking set up at all. The latter is built into and setup up automatically by systemd:
ProxyCommand allows you to use any command to setup a connection. Not necessarily an ssh command, like ProxyJump. It can be any command, as long as it receives on stdin and produces on stdout, it can act like a TCP connection.
ProxyJump is a special case of `ProxyCommand ssh -p <port> <user>@<host>`. Can't replace the `ssh` in there when using ProxyJump.
I came across ProxyCommand earlier this week, funnily enough. I have Cloudflare Zero Trust set up with an SSH service[0], and have the server firewall drop all incoming traffic. That helps reduce my attack surface, since I don't have any incoming ports open.
I use ProxyCommand in edge-case devices where key auth is not an option and the password is not controlled by me. ProxyCommand points to a script the retrieves the password from the vault, puts it on the clipboard for pasting, reminds me via stderr it's done so, and then proxies the connection.
Interesting. I might have such a use case. Do you have anything about best practices on how to automate grabbing passwords from vaults? Cuz it seems to me that the vault needs to be kept open or keep the vault password somewhere on disk.
`pgrep`/`pkill` HATE this trick! Learn how this renegade developer (relatively) easily exits their hanging SSH sessions without restarting their laptop.
I still struggle with the fact that I was (or delusional) an effective C programmer "back in the day" (before google etc) and all we had was "man" to look up std/x11 system calls.
Now I am dismayed with juniors who can't even be bothered to use google (or llms) to look up stuff on their first hiccup.
Hidden or undocumented features like this always have a strange appeal. Part of it is nostalgia for older software where small Easter eggs or experimental features would sometimes ship in production builds.
"It's easier to ask an AI" can be true without implying that manpages are bad.
However, "man" the tool does have issues, and one of them bit me just now.
So, I didn't know about openssh client escapes like ~?. I thought, "surely that's in the manpage?" I opened the manpage (in less) and searched for "\~\?". No hits.
Of course, escape characters are documented in the manpage, and the string "~?" does appear. Why didn't search find it? Because man, in its infinite wisdom, decided to render every instance of "~" as some bizarre unicode not-tilde, which is visually similar but totally impossible to grep for.
This has also bitten me in the past with dash. DASH. A character that is critically important when documenting invocation options. man loves to convert it into something that looks like dash, prints like dash, but doesn't come up in search.
I'm sure there is a way to turn this "feature" off, and I'm about to spend a bunch of time figuring out what it is. But this is documentation for command-line tools. Silently destroying our ability to grep it should NOT be the default.