ToxicFrog posted:

^^ Cool. I wonder what expect is doing that normal IO redirection isn't; attempting to do this with pipes results in ssh ignoring the input attached to stdin and hijacking the terminal to ask for a password anyways.

Sepist posted:

I figured there was no way to feed the password but since this entire thing has to be manned I'm not concerned about that, thanks!

ShoulderDaemon posted:

Or you can create the 2 disk RAID 5 in a non-degraded state, which is also allowed and is in fact the default. It is in that case functionally identical to a mirror, has the capacity of only one disk, and you do not lose anything if either of the disks dies.

Puck42 posted:

I've done it in the past using a perl script and the SSH module.

http://search.cpan.org/~turnstep/Net-SSH-Perl-1.34/lib/Net/SSH/Perl.pm

notMordecai posted:

I have a pretty simple question regarding a new 10.04 I put on an older PC. I haven't used Ubuntu since 6.0 so I am really rusty on this sort of thing.

I installed Ubuntu 10.04 using the windows installer on an external drive (E:\) I use to hold nothing but my media (movies, music, books, etc). Upon instillation, I gave it a small 20GB partition and installation went fine.

When I booted into Ubuntu, I was a little dismayed to notice that Ubuntu cannot recognize any of the media surrounding the 20GB partition on the 2TB drive I installed it on. Was this supposed to happen and is there a way around this? I kind of just... "expected" ubuntu to notice the rest of the E:\ but then again I could just be stupid.

I can easily just uninstall ubuntu and put it on the native C:\ but would kind of like to avoid the hassle if I can.

FISHMANPET posted:

Does the device show up in the places menu? It might just not be mounted.

Ubuntu for sure sees your drive, because Ubuntu is installed on a 20GB File that lives inside that partition. 'sudo fdisk -l' will list all the disks attached to your machine, does it show up there? 'mount' will list all mounted devices. Is it mounted but mounted somewhere weird? I'm not sure exactly how wubi interacts with the device it lives on.

PlesantDilemma posted:

Sepist posted:

^^^^ Your 2Tb drive's mount point is /host, you should be able to view the contents just by cd /host or go to that directory from the root directory..

Sepist posted:

I'm not really a GUI person, can you drag and drop the folder to the taskbar drop down?

notMordecai posted:

Good lord for some reason that never came to mind. I must really be out of my element right now. Thanks.

Sepist posted:

It sounds like it's trying to open a non-existing device, can you verify /dev/dsp is not missing somehow?

Also in regards to the sound not working, if you do "alsamixer -c0" is your line in set to 0? I had that problem with my line out in fedora, I went into alsamixer, saw that it was set to to 0, raised it back to 100 and ran "alsactl store" and was back in business.

Rocko Bonaparte posted:

I am having a problem keeping some commands running on a remote machine that I initiated through SSH using an SSH library. In this case I'm really playing with Chilkat SSH, but I've seen this around before with other libraries. If it becomes a big deal, I can throw something together in Perl so we have something everybody could work with.

I've been comfortable with using nohup to basically background commands after I log off a box, but that only seems to work interactively when I am logging in from something like PuTTY, Cygwin, or an xterm window on another Linux machine. Whenever I try to use a library to use SSH like a shell, I find that it stops the command in place as soon as I disconnect.

So let's say I want to run: sleep 20; echo Hooray > /tmp/hooray.txt, and assume I disconnect afterwards, what are some tricks to make this work?

If I was doing this manually, I could:
1. Run: nohup sleep 20; echo Hooray > /tmp/hooray.txt
2. CTRL-Z
3. logoff, whether elegantly or not.

And even if I get all that in during the sleep command, it'll eventually finish it for me.

Some things:
1. I also considered the "batch" command, but I am not to familiar with it and so far it's had the same problem.
2. Reading around, it looks like the standard I/O pipes come into play. Like, I should pipe /dev/null as standard in or somesuch. Maybe that's it but I wonder why I have to do it.
3. I've seen hacks involving wrapping the commands up into scripts that I was hoping to avoid, mostly so I can understand what the underlying problem is. This is something of an annual problem for me that I wanted to nip in the bud.
4. I can't rely on GNU screen being installed on the remote machine.
5. I'd prefer if you had tried this before using an SSH library to achieve this effect programatically. And if you have had success I would like to know the language, library, and anything about the implementation so that I could compare. It's that doing this manually seems to work fine, so there are a lot of manual methods that ultimately don't work this way.

Ziir posted:

I was given a laptop with Ubuntu installed on it I would like to just set it aside in my room and leave it permanently on to use as a server or something that my MBP can connect to. I understand I need to install OpenSSH and I did that, but I'm completely lost now because I can't find a "dummy's" guide to this. The Ubuntu wiki page for setting up an OpenSSH server (https://help.ubuntu.com/8.04/serverguide/C/openssh-server.html) is pretty much useless I think.

Can anyone tell me where I can find a guide to basically walk me through the process or just post here telling me what to do and how I can secure it so that only my MBP can connect to it?

OpenSSH is installed and running right now and when I try to connect to it from my MBP by typing ssh useraccount@ipaddress it tells me my connection is refused.

The end result is I'd like to install a VNC server as well that I can turn on via SSH and essentially have access to a Linux box full time.

Ziir posted:

I installed openssh-server.

Moose Milkie posted:

do a sudo netstat -natlp | grep 22 and see if anything comes up.

enotnert posted:

If the disconnect is whats borking it have you tried using Screen?

Rocko Bonaparte posted:

I had mentioned that in my first post. I can't rely on having screen on these machines because they're little barebones things.

Rocko Bonaparte posted:

I am having a problem keeping some commands running on a remote machine that I initiated through SSH using an SSH library. In this case I'm really playing with Chilkat SSH, but I've seen this around before with other libraries. If it becomes a big deal, I can throw something together in Perl so we have something everybody could work with.

I've been comfortable with using nohup to basically background commands after I log off a box, but that only seems to work interactively when I am logging in from something like PuTTY, Cygwin, or an xterm window on another Linux machine. Whenever I try to use a library to use SSH like a shell, I find that it stops the command in place as soon as I disconnect.

So let's say I want to run: sleep 20; echo Hooray > /tmp/hooray.txt, and assume I disconnect afterwards, what are some tricks to make this work?

If I was doing this manually, I could:
1. Run: nohup sleep 20; echo Hooray > /tmp/hooray.txt
2. CTRL-Z
3. logoff, whether elegantly or not.

And even if I get all that in during the sleep command, it'll eventually finish it for me.

Some things:
1. I also considered the "batch" command, but I am not to familiar with it and so far it's had the same problem.
2. Reading around, it looks like the standard I/O pipes come into play. Like, I should pipe /dev/null as standard in or somesuch. Maybe that's it but I wonder why I have to do it.
3. I've seen hacks involving wrapping the commands up into scripts that I was hoping to avoid, mostly so I can understand what the underlying problem is. This is something of an annual problem for me that I wanted to nip in the bud.
4. I can't rely on GNU screen being installed on the remote machine.
5. I'd prefer if you had tried this before using an SSH library to achieve this effect programatically. And if you have had success I would like to know the language, library, and anything about the implementation so that I could compare. It's that doing this manually seems to work fine, so there are a lot of manual methods that ultimately don't work this way.

amishpurple posted:

Put your desired commands in a shell script and then "nohup ./script &" and you should be good.

Misogynist posted:

What the hell is wrong with you people?

Accipiter posted:

This was a pretty important part of Moose's post:

Ziir posted:

code:

rudolf@Clausius:~$ sudo netstat -netlp | grep 22
tcp        0      0 0.0.0.0:22              0.0.0.0:*               LISTEN      0          14593       1988/sshd       
tcp6       0      0 :::22                   :::*                    LISTEN      0          14595       1988/sshd       

I've turned on public key encryption and generated a key for the server machine and now I'm able to do ssh localhost from the server and everything seems to work fine. On my client machine I've also generated a key and then I moved the public key (ziir.pub) to the server and added it to authorized_keys and restarted OpenSSH (dunno if this was necessary) but connection is refused. Any ideas?

code:

Navier-Stokes:~ ziir$ ssh 10.0.1.4
Permission denied (publickey).

Navier-Stokes:~ ziir$ ssh rudolf@10.0.1.4
Permission denied (publickey).

I'm trying to connect from within my network for now and I'll deal with connecting from outside when I get everything working nicely.

Ziir posted:

code:

rudolf@Clausius:~$ sudo netstat -netlp | grep 22
tcp        0      0 0.0.0.0:22              0.0.0.0:*               LISTEN      0          14593       1988/sshd       
tcp6       0      0 :::22                   :::*                    LISTEN      0          14595       1988/sshd       

I've turned on public key encryption and generated a key for the server machine and now I'm able to do ssh localhost from the server and everything seems to work fine. On my client machine I've also generated a key and then I moved the public key (ziir.pub) to the server and added it to authorized_keys and restarted OpenSSH (dunno if this was necessary) but connection is refused. Any ideas?

code:

Navier-Stokes:~ ziir$ ssh 10.0.1.4
Permission denied (publickey).

Navier-Stokes:~ ziir$ ssh rudolf@10.0.1.4
Permission denied (publickey).

I'm trying to connect from within my network for now and I'll deal with connecting from outside when I get everything working nicely.

enotnert posted:

It's fuckign early, and I'm up doing unrelated bullshit, but if you can't hit it within your network, I'd check your iptables/firewall to make sure it's not boning it somewhere.

Accipiter posted:

Permission Denied is NOT the same thing as Connection Refused.

Your key config isn't working. SSH is responding on the network just fine.