Software :: Shutdown Of Server Over A Serial Connection?
Jan 18, 2010
We have an APC UPS connected to our main server via a serial cable, with the APC Powerchute Business Edition up and running nicely. We also have another server which is protected by the same UPS I started looking at an APC interface expander which would allow shutdown of multiple servers from the same UPS, however it appears it's not standard serial cables but a whole proprietory nightmare.possible to hook both servers up with a serial cable and have the master shut the other one down through a shell script? Both boxes are CentOS 5.4
I have installed a cluster computer with 10 nodes . The manufacturer is HP . All nodes and the master node have redhat enterprise linux installed in them . When I shutdown the nodes from the master terminal using "shutdown -h now" they get shutdown . But they dont get completely turned off . This issue bothers me when the power supply is given , all nodes boot up simultaneously generating a huge heat .
Thing to note : When we shutdown our PC they get completely turned off . When the power supply is given , a press on the Power On button is required to boot the system. But , why does it not happpen in the case of cluster? Is there any other way of completely turning off the nodes from the master terminal ?
I have been tol that for debugging purposes it is often useful to have a serial connection to a computer. I have some diskless workstations thats freezes during boot. I guess X has someting to do with this, but I'm not sure. Since the workstations are diskless, syslog is not stored locally so I cant se what is happening after the NiC stops working. When the worstation freezes, screen is going black, all lights on the keyboard turns on, lights on NiC is going black. It is not possible to ping the workstation.
I have a Sun server that could be only configured via serial interface. It has one serial port with a RJ 45 connector. Although my laptop has not serial interface. Is there a way to use C-Kermit or any software to establish a serial connection via the ethernet port. It should be some kind of virtual serial port that should transmit the information over the ethernet port, without encapsulating it in ethernet frames.
I have a Cisco Switch that is connected to my computer via serial. I simply want to configure the switch via the serial connection from within a terminal. I edited the /etc/inittab to include agetty for ttyS2 and ttyS3. The switch is hooked up to one of the two.
I ran
Code:
init q
to re-read the inittab
Ran a
Code:
ps -ef
and I can see the two agettys I also have run a dmesg to confirm that my computer has detected ttyS terminals. The requirements for the cisco switch connection are the following:
I've been reading too much and understanding too little. What I have: Lucid laptop & desktop. Both have ethernet > hub > internet. What I want: connect them directly by the serial ports so the laptop can access the desktop's drives, and the desktop can be retired from the internet. I've got an old DB-25 null-modem, last used for multiplayer Doom. The laptop's dock has DB-9, so the connections is desktop > null-modem > Belkin DB-25 - DB-9 cable > laptop.
I had hoped this post would get me somewhere:[URL].. No such luck. When I ping 'grin.local', the connection is through the internet. If I disconnect the hub, i cannot ping.
Dunno if there's something more I should do to enable serial connection in Lucid, or if my null-modem is no longer good. I'm also rather curious how "grin.local" manages a connection through my ISP's system. These are DHCP. Shouldn't that have required the full numerical IP?
(It's nothing to do with the hub. Checked that by connecting a third box to the hub. My ISP only hands out two IP. I couldn't ping the third box.)
Hello. I want to use my Debian box as an internet connection for a Win95 laptop. The laptop is old enough that the only port I have available to connect to the internet from is the serial port. I have heard that a serial port redirector will allow it to access the internet through my debian box, but I have no clue how to set one up. Any advice. Thanks
My computer for a project I'm working on will be using Ubuntu with Perl and VLC from the repositories, if that matters at all. I'm not sure what I should/could use to accomplish this, but essentially I'm looking for Perl to be able to send predefined data via a serial connection based on the time within a song that VLC is playing.
A couple of the things that I've thought about after Googling a bit are SMPTE and subtitles/lyrics; I can't stop thinking that there's got so be a simple way that I'm just overlooking, though - most of the stuff I've found is for over-the-network applications, and that's overkill for this situation. All the subtitles/lyrics would have for text is hexadecimal values; all the SMPTE would do is keep Perl and VLC (playing audio) in-sync with each other. I would need either one or the other, but not both. Or, I could use something else...
I'm envisioning something like the setup below, I just don't know what else I need to make it happen.
Files: song.mp3 timing.txt script.pl
When I run "./script.pl song.mp3 timing.txt", VLC will open, and start playing "song.mp3", and the variable "$mils", for milliseconds elapsed, will be equal to zero. For each line in "timing.txt", the "script.pl" will assume the portion before an underscore is the "$time" for that line (time in milliseconds), and "script.pl" will assume that the portion after the underscore is the "$data", a three-byte hexadecimal value (such as "ff101f", or "011c00"), for that line.
I think that I will need to use two arrays, "@time" and "@data", such that the given line in "timing.txt" would read the same as printing: "$time[$line]".'_'."$data[$line]"." " (if "$line" is the line number). My problem: jitter/stuttering playback. If VLC stumbles (which it sometimes does), and is a about half a second behind where it "should" be, but Perl didn't stumble, then I've got an out-of-sync issue.
Since Perl can't just count the time by itself to avoid synchronization issues, I've got two main ideas for getting this done: One way: Perl has to "read" where VLC is at, and use that as the "$mils" value. Then, if VLC delays/stubmles, so will Perl, but everything will still be in-sync. Alternative way: VLC has to "listen" to what Perl says "$mils" is at any given moment, essentially having Perl be in control of VLC's timing. This would make VLC "skip ahead" (to where it "should" be) if it starts to fall behind, thus keeping the two in sync.
I don't know which, if either, of these ways is best for my particular need. I would need "script.pl" to send "$data[$line]" over the serial connection when "$mils" == "$time[$line]". I already know how to send serial data from Perl, read files via Perl, and convert the hexadecimal number (if needed). I don't know how to "link" Perl and VLC so that they stay synchronized, and I am quite lost as to where to go from here.
I am trying to get two way serial communications going between a Windows XP system and a Linux system (RHEL 5).I have /sbin/agetty -L 9600 ttyS0
in /etc/inittab. I am using a generic USB to serial adaptor on Windows (Unitek) and a null modem cable. I have putty configured for 9600 baud, 8 bits, no parity, one stop bit, no flow control.I get the login prompt from agetty in the putty window but input does not work; I see weird characters in the putty screen. I can echo output into the device from windows and see it, but
cat < /dev/ttyS0. just prints out weird characters from what I type.
On my CentOS 5.2 install I've installed a dual-port serial card based on the NetMOS 9835 chipset. I've followed the serial instructions, but everything they tell me to do is already done - see setserial outputs below. I've compared these values to the lspci output (below, NetMos card is at the bottom of the output) for the card and it appears right. Problem being, I can't use the serial ports. Oh and I did use search, and looked at every 9835-related article before posting this :)
I am hoping to be able to get an old serial-touchscreen to work with a usb/serial adapter. I had this touchscreen working some years ago on different hardware. I would like to hook it to the machine I am setting up as a multimedia host with mythtv among other things.
Following the instructions I left behind when I got this to work way back when does not work. See [URL] ....
This info is a work around to get the xserver to see the touchscreen. [URL] ....
I do know that the touchscreen works as I am able to get garbage on the screen as in the first part of my howto. But I have not had any success getting xorg to see it. I wish I had posted a copy of the xorg.conf at the time, but......
There is an issue with the current xserver in testing that I am hoping the next update (in unstable) will fix when it gets pushed to testing. That is that Code: Select all# X -configure fails with a segfault. So I am not able to generate the xorg.conf needed to get it to work. I was going to post a bugreport, that is when I found out there is an update in unstable, so I am waiting for it to get pushed at the moment.
Anyway I am hoping that I can link /dev/ttyUSB0 to /dev/tty0 and get it to work. I would like some kind of guess as to my chances of success before I go to all the work of getting the monitor hooked to the host.
Linux printing appeared to be working fine up until yesterday. Today typing lpq gives the following: lpq Printer 'sdst@other.domain' - cannot open connection - Connection timed out Make sure LPD server is running on the server
The /etc/cups/printers.conf file is properly set, the printers appear in localhost:631 and they are printing test pages. However, all command line print commands seem to be trying to print to sdst@other.domain I don't know why printers.conf is being ignored and why and how sdst@other.domain was added. Seems like it might have been auto-discovered?
# dit: sdst@other.domain was mentioned in /usr/local/etc/lpd.conf I'm not sure why lpd.conf is being used instead of /etc/cups/printers.conf
I have a bit of a dilemma. I'm attempting to use a microcontroller to send MIDI messages over a virtual serial port. I want to (eventually) read them with Mixxx. I'm using an FTDI chip, so I get a /dev/ttyUSB0. I've tried spikenzielabs' Serial-Midi program, but it doesn't see any serial ports. NOTEMIDI looks really old and won't compile on my 10.04 LTS machine.
We have a new machine with RedHat enterprise 5 on it. I need to connect a serial cable to the serial port and talk to another system (old alpha system) instead of using a VT connected to the alpha.Does RedHat come with anything like Keaterm/hyperterm/etc etc?
I'm a frequent user of the shared connection feature of SSH. Usually i have two or three connections open at the same time through the connection sharing feature, but sometimes when i try to start another i get the message "Shared connection to <server< closed". All the connection that are established through the connection sharing feature are still active and not closed, but when i try to create a new one i get that message.. i've even got that message when i have only had two connections through the shared connection feature.
I've tried to search google but it doesnt seem like anyone else has had the same kind of problem, does someone know why i sometimes get this error ?
I've got a home server running ubuntu 10.0.4. It have no monitor or keyboard. So I want to access it from a serial console. (I use SSH currently but if something goes wrong, I need to be able to do something... so serial console seems a fine thing). OK, so on the server, I tell grub to output everything on ttyS0 and opened a getty on ttyS0. On the client side (Laptop running Ubuntu 10.0.4 with pl2303 USB to Serial adapter) I use Minicom.
Home server restart:I see grub menu, linux kernel messages, login prompt. But, I can't do anything:In grub or at the login prompt, no key press is recognized by the server.I've got write permission to /dev/ttyUSB0. I tried a different getty (mgetty)with no success.I stopped the getty on ttyS0 on the server. And run minicom.
From the server's minicom, if I type something, it appear on the client screen. From the client's minicom, if I type something, some garbage appear on the server screen :(One of this things for each char typed). If I short-circuit pin 2 and 3 of the client serial port, it echoed what I type. The two serial ports are connected with a nullmodem cable. (seems to be a full handshaking cable). Of course, same serial port config for both minicom, grub and getty. Hope I'm understandable enough. (English not my primary language). Can somebody help me make this *$*#! serial console working right ? I'm on it for 10 hours now with no success...
Edit : I nearly forgot to tell you I had to update the standard ubuntu kernel to 2.6.34-020634rc1-generic because there was a bug in the pl2303 driver in the stock kernel (2.6.32-25-generic) which prevented me from doing anything (even receiving).
I am wanting to set-up serial communications between my server (ubuntu 9.04) and a picaxe chip (a pic with a bootloader) I need to send data in the format of "1,1,1" or "1,1,0" at 9600 baud 8 data bits, no parity, 1 stop bit and no flow control. the data must be sent as numbers not ASCII I have tried Code: echo 1,1,0 > /dev/ttyS0 Any simple shell script or even python.
Here's a little tool that does this: Without arguments, updates the SOA serial in a zonefile to the current date. If the date was already updated, just updates the revision number (incrementing up to 99, and then again 01). Uses RFC 1912* recommended format. With $1 == <two digit number>, auto updates (if necessary) just the date part and uses your provided revision number. With $1 == <eight digit number>, uses that as a date (no validation of any kind), and just auto update the revision number With $1 == <full serial>, will just replace whatever the serial is with the provided serial, without any validation
* YYYYMMDDRR (4-digit year, 2-digit month, 2-digit day of month, 2-digit revision number) This script + keeping SOA/NS/MX/CNAME RRs in a common file $included from other files with $ORIGIN and A/PTR/TXT RRs, made everything way easier to manage, enabling me to script some zone switchers, automatic failover/redirection of DNS on WAN changes, etc, etc. I think this stuff might be cool to integrate with something like this script and make nice CLI toolset for bind. Looking forward to implement it.
I'm working on program that is going through setting different baud rates onto a config file. After I set a new value i want to check if it's the correct one by reading from the serial port on the server unit. I know what to expect if it's the correct baud rate so that's no problem, but searching through the internett i've yet to find a compact solution to my problem.
I have a debian machine with an apache2 webserver. I am able to start the machine from the internet (power plug board with webinterface) but I don't know how to shutdown the server automatically if nobody uses the website anymore. It is a homeserver which should only run if needed.
Unfortunately I don't know much about Linux, apache, php, cgi-scripts, cronjobs and other things that might be useful. But I googled a lot and have now an idea how it could work. It seems complicated to me and so I want to ask you guys what you think about it and if perhaps you could give me a hint.
The idea: A cronjob starts every ten minutes a php script that checks when the index.html of the welcome-page was opened the last time (fileatime()). Lets say it was opened last time at 3pm. Then the php script adds for example 60 minutes to that time (=4pm) and calculates how much time is left to 4pm. This time is saved in the variable $timeleft. Now the php script compares if $timeleft is less then 10 minutes (=600 seconds). If it is, the php script starts a cgi script that will shutdown the server. The cgi script will login as root (I think that can be done with the "expect"-command url and then enter the command "shutdown -h now"
Is all this realizable? Isn't there a better and/or easier way to do?
I am running an Ubuntu Server on a VirtualBox VM running on my windows machine. So I've created a self-signed certificate using the following tutorial: [URL]
From this tutorial I'm left with 3 files: server.key server.csr server.crt
Then I found this very similar tutorial that has an extra bit on installing the certificates in apache: [URL] So I followed it's instructions which boil down to this:
[Code]...
So I'm thinking this should work now. However in Chrome I get: SSL connection error Unable to make a secure connection to the server. This may be a problem with the server, or it may be requiring a client authentication certificate that you don't have. Error 107 (net::ERR_SSL_PROTOCOL_ERROR): SSL protocol error. IE8 gives me a typical "Internet Explorer cannot display the webpage" Note that [URL] fails while [URL] works fine, so it's definitely something in my ssl setup I'm thinking.
vsftpd not working on my CentOS system. I tried logging in using FileZilla but the error message was: Connection attempt failed with "ECONNREFUSED - Connection refused by server".
I can make a connection to the vpn server, the connection starts, but nothing happens! My IP address remains the same as previous! There is nothing added to my Knetworkmanager, I'm a beginner I should use a pcf file for my vpn connection. I use it properly, I'm sure because the connection starts and an icon is added to my panel and remains until I disconnect.
I checked it via ifconfig -a, the last part (which is for vpn) is:
Recently I installed vncserver (tigervnc) on my desktop. Ever since my computer refuses to shutdown normally. At shutdown the following message pops up: Quote: System policy prevents stopping the system when other users are logged in Then I have to enter the root password to shutdown. If I stop vncserver before, the computer shuts down normally.
