Linux help!

[Rim]

I'm running RedHat 9 and it's stuck on what I assume is the Linux version of the Blue Screen of Death (unresponsive to keyboard). It's at a remote site so power cycling is tricky.

 

Is there a way I can tell it to reboot remotely (from another RH box)?

 

For you Linux gurus, I'm running BIND, SendMail, Apache, Nagios, and Cacti on the box.

 

There's a bunch of tech code on the screen like:

 

EIP is at vfs_fstat [kernel] 0x31 (2.4.20-28.9smp)

Process snmpget (pid: 23529, stackpage=e6753000)

Call Trace:

sys_socketcall [kernel] 0x34 (0xe6753f80)

do_page_fault [kernel] 0x34 (0xe6753fb0)

error_code [kernel] 0x34 (0xe6753fb8)

system_call [kernel] 0x34 (0xe6753fc0)

[deanmass]

Assuming you are able to get into it at the command line, type (from sbin) "shutdown -r now"

 

Use 'shutdown /man to read all the options for the shutdown command

-r means 'reboot'

'now' is do it NOW dammit..

 

Syntax is 'command switches modifier

 

If it is at a remote site, hopefully it will come back up in a way that allows you in to look at logs and such.

 

After you get it back up, you need to lock it down better (prolly) if telnet is enabled, etc. Nothing like that should really run remote from root, but read up first, cause I am just getting my feet wet on the *Nix sys stuff.

 

Post if you get it back online..I am curious if it works out.

[phaeton]

if it kernel panics like that, you won't be able to do ANYTHING but physically hit the power switch.

 

Otherwise, if it's still running but acting oddly, you can usually ssh or (ugh) telnet in and restart the services that have stopped, or restart the machine itself if necessary.

[deanmass]

Phaeton, is that screen the official 'kernal panic' screen? Will that kill any external connections? (guessing yes, but not sure...)

 

Rim, sorry, I guess I wasn't thinking no KB...err...DUH...

[phaeton]

Sorry, i guess that my post was really vague and not helpful at all, heh...

 

Kernel panics look slightly different depending on what caused them, but clonk here to see an image i pulled off the web.

 

You'll probably never see this remotely, as it will dump its core to the first virtual terminal (vty0) and that's it.

 

if you can get into the machine via term, then it hasn't panicked. You can get to it like:

 

ssh -l

 

If it's sitting behind a router with a non-routeable ip address then you're probably SOL unless there's another machine you can get into that can get into it.

 

If you can't ssh into it, try telnet:

 

telnet

 

All of this assumes that you have a user account on the system, and the next assumes that you have wheel priveleges as set in /etc/group:

 

Type "su", and it will prompt you for the root password. If you know it, enter it. If you don't, there's probably not much else you can do.

 

If you can become as root, check any mail messages first ("mail") as they may contain warnings about doing the following:

 

shutdown -r now

 

If you're lucky, it'll come back up. If you're not, it won't. You can ping it if you're sweating bullets:

 

ping

 

and if you start getting a response that's a start..

 

Hope this helps, but the problem you're having is kinda vague...

 

is it not responding to keyboard at location, or did something happen like your telnet session stop responding? It may have just timed out, and you might be able to reconnect.

[Rim]

It was very similar to a kernel panic, though I didn't actually see the words "kernel panic" like the screen cap on you link, phaeton.

 

This is for work and it's something I inherited and I am still a Linux newbie. We don't have telnet enabled, but SSH is, though I couldn't SSH. Pings timed out when I usually get replies. We also have a remote KVM solution at the remote site so I was able to see the screen, but it didn't respond to any keyboard commands - I couldn't do Alt-F2 to bring up a different terminal screen.

 

I am way more familiar with Windows and I know in Windows, if a machine is unresponsive, I can try to reboot it from a different Windows machine (works sometimes). I didn't know if Linux had a similar function (outside of SSH/telnet).

 

Anyone have good books/links/whatever that will get me up to speed quick? I've seen many books but many focus on installing Linux. The distros I've tried, installation has become pretty straightforward. I need something to get me up to speed on troubleshooting Linux - even just the basics, like finding how much space is left on a volume. Also, if there are any maintenance I should be doing on Linux, I would like to find out what and how to do it.

 

If it wasn't for Webmin (webmin.com), I would be so useless when dealing with Linux. Webmin is a web-based interface that allows you to do a lot of procedures on a Linux box from editing files to installing and configuring apps and I love it. The downside is, it's a crutch, so if you put me in front of a Linux box without Webmin, I am pretty lost.

[phaeton]

http://www.icon.co.za/~psheer/book/

 

It's probably the best and least-distro-centric guide out there. The treeware version is worth having around too.

 

And for space left on a volume, try

 

df -h

 

You don't always see the term "kernel panic", sometimes it'll puke before it gets there. Do you ever have physical access to this machine?

 

Linux isn't perfect, but it's pretty dern robust.. if this is happening a lot, there could be issues like a failing hdd or (more likely) cpu fans that are going. We had a rash of stock AMD CPU fans (the ones they require you to use to maintain your warranty) dying on us, and that's just how they'd act. The processors would overheat and shut themselves off. The machine would just quit mysteriously with no output.

 

We'd go in and restart it, and it'd come up fine (cores cooled off by then ) then in 5 minutes they'd quit again. Kinda funny now, but a real pissor then..

[Franknputer]

Webmin rox! BTW - a shorter version of the reboot command is 'init 6'. The 'init' command will switch runlevels - if you're in graphical mode, you're in runlevel 5, normal command-line mode is runlevel 3...runlevel 0 is shutdown, and runlevel 6 is restart.