Unusual request

The problem: One of my machines seems to have developed some problem. I
can't physically go to it and see what is going on. My attempts to connect
to it via ssh just time out - but, it responds to my pings
without any noticeable delay, when such pings are issued from the other
boxes in the same LAN.

The question: Is it possible, under such circumstances, to get this
machine to reboot remotely? The only communications service that I keep
open in it is ssh. Also, at the time the problem developed it was running
X windows.

Steven Jones wrote:
> The problem: One of my machines seems to have developed some problem. I
> can't physically go to it and see what is going on. My attempts to connect
> to it via ssh just time out - but, it responds to my pings
> without any noticeable delay, when such pings are issued from the other
> boxes in the same LAN.

Are you sure that sshd is running or runnable on that machine? On my
distribution for example, Debian Sarge, the ssh server is in a different
package (openssh-server) from the ssh client (openssh-client).

And if sshd is running, your syslog (/var/log/syslog or
/var/log/messages) should tell you why ssh doesn't connect.

>
> The question: Is it possible, under such circumstances, to get this
> machine to reboot remotely? The only communications service that I keep
> open in it is ssh. Also, at the time the problem developed it was running
> X windows.

You need to login somehow first.

Robert

"Steven Jones" <(E-Mail Removed)> wrote in message
news(E-Mail Removed)...
> The problem: One of my machines seems to have developed some problem. I
> can't physically go to it and see what is going on. My attempts to connect
> to it via ssh just time out - but, it responds to my pings
> without any noticeable delay, when such pings are issued from the other
> boxes in the same LAN.
>
> The question: Is it possible, under such circumstances, to get this
> machine to reboot remotely? The only communications service that I keep
> open in it is ssh. Also, at the time the problem developed it was running
> X windows.
>
Steven,

Some systems have the ability to be serviced remotely, even
rebooting, and some do not.

Example: Any HP system with "iLO"

( Integrated Lights Out )

In these systems there is an extra network connection
that is serviced by components that are integrated
into the node. The iLO interface is serviced by ROM
based code, and is always on. You can interface
to the iLO over a network, then login, or reconfigure,
or reboot, the node.

These types of systems are designed for rack-mount
server environments, where remote management is
a requirement.

For managing systems that do not have an iLO.... I typically
use a serial console, and if that fails, then its time for
a network controlled power switch.

Enjoy,
Postmaster

On 2006-02-19, Steven Jones <(E-Mail Removed)> wrote:
> The problem: One of my machines seems to have developed some problem. I
> can't physically go to it and see what is going on. My attempts to connect
> to it via ssh just time out - but, it responds to my pings
> without any noticeable delay, when such pings are issued from the other
> boxes in the same LAN.
>
> The question: Is it possible, under such circumstances, to get this
> machine to reboot remotely? The only communications service that I keep
> open in it is ssh. Also, at the time the problem developed it was running
> X windows.


Well, I can't help you *this* time, but I have something that *might*
be useful for *next* time. I personally have never used this, as I
have no need for it, but I saw if referenced on Harald Welte's blog,
and recommended it for a friend who was having similar issues with a
server, and it turned out to be very useful to him.

http://marek.terminus.sk/prog/ipt_sysrq.shtml

If you're having frequent problems with this occurring, then you need
to spend some time focusing on troubleshooting your (likely) hardware
issues, but that ipt_sysrq target might get you by until then...

RW

--

http://rlworkman.net

On Mon, 20 Feb 2006 02:26:18 GMT, Robby Workman
<(E-Mail Removed)> wrote:


>Well, I can't help you *this* time, but I have something that *might*
>be useful for *next* time. I personally have never used this, as I
>have no need for it, but I saw if referenced on Harald Welte's blog,
>and recommended it for a friend who was having similar issues with a
>server, and it turned out to be very useful to him.
>
> http://marek.terminus.sk/prog/ipt_sysrq.shtml
>
>RW

ipt_sysrq won't 'make' on any of my Slackware installs (10.0 and 10.2)
and the author has not replied to Email.
--
buck

On 2006-02-21, buck <(E-Mail Removed)> wrote:
> On Mon, 20 Feb 2006 02:26:18 GMT, Robby Workman
><(E-Mail Removed)> wrote:
>
>
>>Well, I can't help you *this* time, but I have something that *might*
>>be useful for *next* time. I personally have never used this, as I
>>have no need for it, but I saw if referenced on Harald Welte's blog,
>>and recommended it for a friend who was having similar issues with a
>>server, and it turned out to be very useful to him.
>>
>> http://marek.terminus.sk/prog/ipt_sysrq.shtml
>>
>>RW
>
> ipt_sysrq won't 'make' on any of my Slackware installs (10.0 and 10.2)
> and the author has not replied to Email.


I seem to recall that he had to edit some of the source code - I think
it was looking for iptables headers in the wrong directory. Email me
privately (see headers) and I'll forward your information to him.

RW

--

http://rlworkman.net