Ubuntu Servers :: Epic Battle To Daemonize Rtorrent On Headless Server
Mar 9, 2010
I spent most of the day yesterday trying to make this work and haven't gotten really anywhere with it. Building from source quietly failed although it built without error and I got checkinstall to pack it up it didn't actually work.So I rolled back to the repository version of rtorrent.As I understand it, the things I need to do are (I am thwarted at every step):create user with --disabled-password optioncreating the user is no problem except they seem to have no shell, and running any command while su'd as that user throws an error like Code: Cannot open your terminal '/dev/pts/1' - please check.
I found somewhere that changing the permissions on /dev/pts/N solves this problem however N can change after a reboot, or probably following other events so that's likely not the right fix.create init script to run rtorrent as user in previous stepI found an example that seems like it works but I wont know for sure until I get the prior step worked out fully.
I have been having some trouble configuring a means to remotely administer an Ubuntu Sever 10.04 machine i recently built and am hoping for some guidance. I'm sorta new to linux so please forgive my ignorance of conf files and the command line . I am looking to have this server serve as...VPN Server, Minecraft Server, KVM Server, Torrent and Dyn-DNS box.I have managed to install all the KVM settings including bridged networking running (Personal victory there) and sorta have Minecraft running (only need to figure out how to make it launch on boot). Torrents already taken care of since Transmission came along with the ubuntu desktop install .The big thing that is slowing me down is remote management. I installed ubuntu desktop (thankfully its Gnome 2), i know that most people don't like it but its what i'm most familiar with. I tried running VNC but always needed to login on the physical machine before I could use that which defeats the purpose. I am now using xrdp which frankly the performance is terrible compared to what i was getting with the built in VNC server and still needed the monitor on and active to connec
I've got a headless server (32 bit 9.04/9.10, can't remember) that started refusing SSH and HTTP connections a few days ago, and I'm just getting around to hopefully fixing it. The problem is that I can't even connect to the darn thing. I've plugged in a mouse, keyboard, and monitor, but the monitor tells me that it has no signal. I know the monitor works with the computer, as I used the same one to set up the server.
Just for kicks, I even tried to put in a live CD, but that didn't help. Any ideas on how I can connect to this thing?
I have a very specific requirement outside the typical use case of "enterprise server". If possible, I need to run a FUSE filesystem on a headless server, and share it via Samba.Physical security is a non-issue.I see two challenges with this, that Google and I have been unable to figure out: I'm assuming that in order to mount a FUSE system, a user must be logged in. (Hence the "U" in "FUSE".) So the first challenge is how to get Ubuntu to automatically login, on the physical console, to a user terminal session. (I know this can be trivially done in GNOME, but this won't have X-Windows or GNOME installed.) This has to happen automagically after boot-up; manually logging in locally or via SSH isn't an optin.Second challenge (?), how to share a FUSE file system via Samba?
I upgraded my Ubuntu 8.04 LTS server to 10.04 LTS - rebooted and things ran good. I knew I had a few problems, but DHCP and DNS were booting up and working fine for me.I went to take a look at Webmin, but apparently that's been removed from support for Ubuntu 10.04 - so I thought, I'll use eBox. I realized that Ebox didn't have the network module turned on, so I tried to turn it on. I found out the installation was missing scriptaculous, so I got that and installed that and then turned on the network module.I then rebooted the machine - and now nothing. It hangs when I get to the load for Apache. I would like to disable the network module for eBox, but can't find documentation for how to do that from the command line - which I have to use because the damn box won't boot and I'm running from a Live CD.
This is a silly question maybe, but how can I turn my passport drive off after has being unmounted? I am assuming that's the right order, before removing: umount and then turning off. By the way, I am talking about Ubuntu Server, not regular Ubuntu desktop via GUI.
Background: I need some help regarding install X server/apps on a headless(ie. no monitor, no graphic device) server. The server is in fact a virtual one by Amazon EC2. So there's no monitor nor any graphics hardware. Fedora server (6, i believe) is pre-packaged. The problem is that I want to be running a server app with a GUI. The app won't start without the GUI (and I probably need to tweak a few things through its GUI too).
I plan on setting up very very bare minimal X on the server and then uses NX for remote access. Can somebody shed some light here?
Simply put, my questions are:
1) What would be the minimal list of package i need to install? 2) Where can I find docs about installing and setting up NX? I could only find very fragmented/outdated docs about it.
I have both ubuntu and kde-desktop installed, and after trying KDM, switched back to GDM which worked fine for a while (although KDE changed my other gnome settings like usplash and cursors which took a while to change back). A day later, though, when I start up my system, the screen flickers a few times then shows me a message that an x-server is already running, so hit no to try loading it on '0' again or yes to try another number.
If I hit no, it will briefly show me the KDE login screen refresh a few times and go back to the menu. If I hit yes, it refreshes about 7 times and finally shows the gnome login screen. This process takes a long time and I'm not sure is great for my screen.
I've just installed rutorrent following this guide: [URL] As I load the rutorrent site on my server it actually kills rtorrent with this error: Caught Segmentation fault, dumping stack: Stack dump not enabled.
Today I was trying to upgrade mysql and php.PhP went well, thats at the latest version, but the mysql went to 5.5x from REMI. This killed all my websites, no DB and worse.Well I typed the wrong this (totally my fault - serious EPIC FAIL)i typed yum remove mysql not downgrade...This removed plesk and a load of other stuff, and now im screwed.
I've been trying to get the XML RPC interface on rTorrent working for some time now, without success. I'm running Debian 5.0 'Lenny' (stable) i386 on a Intel Atom, and need to let an external application communicate with rTorrent via XML RPC.
I've hand compiled the XML RPC C (1.06) libraries and tried the default packaged version for Lenny, and tried almost all possible combinations of the libtorrent10/libtorrent11 libraries and the rtorrent 0.7.9 through to the latest 0.8.x release. As far as I can see the compilation of both the XML RPC libraries and libtorrent work flawlessly (using no special parameters for the 'configure' scripts). I've compiled rTorrent with the --with-xmlrpc-c syntax, which is also accepted, and it compiles flawlessly.
Code: system:~# xmlrpc 127.0.0.1:5000 system.listMethods Error: Curl failed to perform HTTP POST request. curl_easy_perform() says: Empty reply from server (-504) However, rTorrent (running in a shell) reports
I'm starting several VMs on a remote headless ubuntu server via ssh using:
Code: VBoxHeadless -s <vmname>
The VM starts up okay, and I'm left with an occupied command terminal on my local ubuntu machine. So, if I want to start up several VMs on the remote server, I have to open up several command terminals and end up with as many occupied terminals once they're all running. When the local terminal is closed, the remote VM is also shutdown.
Is there a better way to do this without the remote VM being dependent on the local terminal? I'd like to remotely startup the VMs and be able to close the local terminals without shutting down the VMs. I'm sure this must be possible, I just don't know how to do it.
I have a cron job to start spideroak when the server boots code...
Command line arguments not allowed during New User Setup Interestingly enough.... the line saying "Command line arguments not allowed during new user setup" isn't from my script. Any chance that has something to do with it?
I am running rtorrent on a remote server mainly for seeding torrents. However, I am also using it to download new media.
I wish to share the downloaded torrent media on the web server. This works great for torrents that have finished downloading, however, for torrents that are in the process of downloading, such as large multi-GiB torrents which have slow seeders or no seeders, this creates an issue.
The problem is that I have set the rtorrent default download directory to be visible on the web server. Which is what I want, so that people can get the downloaded media via a web server. Then, when I drop a new torrent into the directory that rtorrent watches, it will start downloading the media to the web servers visible directory. When it finishes the download, then it works the way that I want. So that the media is available both as a web download and also as a torrent download.
The issue is that when a download is in progress, media that is available on the web server is incomplete.
I can manually close a torrent and manually move the media to a different location, then manually tell rtorrent to use the new location for the torrrent (^o)... but that is doing it manually... and if I need to do this with a large number of torrents, it takes a very long time.
I want to automate that so that I don't have to mess with it at all. A lot of other software such as utorrent and such can do this automatically, but it seems that feature is missing from rtorrent.
rtorrent does offer a scripting function in it's config file that can start a script on completion of a torrent file, however, I have not been able to figure out how to automaticaly stop a torrent seed, move the torrent media to a different location and then tell rtorrent the file is in the new location, then start up the torrent again. Moving a torrents associated media to a new location also initiates a hash check when the torrent is started.
I am not going to use different software. Since rtorrent is shell based and uses curses, it doesn't require GUI features. That saves something like 150 megs of memory for disk caching, etc. rtorrent kicks MUCH ass for this purpose and it uses a tiny amount of system resources.
I use screen with rtorrent so that it will run without having to be logged into the system.
I also need to limit the bandwidth on Apache so that people can download with the web if they absolutely must, but I would prefer that they used the torrent files. So, I would like to slow down their downloads quite a bit so that more bandwidth is used with rtorrent.
As with most Linux issues, there is more than one way to do something. The problem is, I can't even figure out one of those ways right now.
I'm new to ubuntu and I wanted to setup a headless server (Asus Hummingbird, 2GB RAM, Ubuntu Server 11.04). The installation was successful but: After I tried starting my server without any connected monitor, the server did not start -> I run into a kernel panic (if I connect the monitor after a while I can see the messages on the screen, Keyboard lights flash...). What do I have to do to get the machine running without any monitor? My second problem is, that the Network is always powered down on system power down -> no WOL is possible. Is there an easy way of enabling the WOL function? I tried several things from different tutorials but nothing worked...
I have now had an Ubuntu desktop server on a very crappy server (network filesystem with enourmus latency) for a year and as the subscription will soon run out I opted for a slightly better one, still crappy but at least a "complete server" (cpu+mem+HD).And a headless installation of Ubuntu Server 10.10 "Maverick Meerkat" instead of my Desktop version.I'm not using it for web pages or other 'ordinary' things though, I'm running my homemade MMORPG off of it and this new headless situation is a bit unclear to me.
I created an account and password with my root account, a folder with the username was created in /home/ and I uploaded my mmorpg executable there with an ftp client.I tried to change the whole folder (being root) to 777 but it didn't work either.
I have an HP MSS 487 (headless). I was going to update to WHS2011 but was thinking of trying Ubuntu Server first. My MSS has a USB slot, but I'm not sure how to start and run Ubuntu Server from that USB stick. I can use remote desktop to access the box. I'll also need to install something on my other pcs/macs so that I can get the backups working. I didn't see anything that is similar to the whs connector software (client). Would I need to install a full Ubuntu client to do the backups?
I have the latest download of Ubuntu Server 10.04 installed. I have not installed a gui for it. I would like to be able to do this from command line if possible. I will be running this server headless but I would like to be able to login remotely if necessary. My hunch is I will need a gui to be able to log in remotely?
I have three Internal drives in my server. 80GB, ext4 formatted, this has the server os loaded on it. The other two drives are 160 and 500GB formatted ext4 as well but they are empty. I have two Ext USB drives with all my stuff I don't want to lose. Both are formatted NTFS. So I need help but keep in mind I am going to run the server headless, but would like to login remotely if necessary without a gui.
My other home computers will dual boot windows or linux, so I need to know what is the safest and best way to format the two drives in my server. ex4 or ntfs? Will I be able to run server without a gui as headless? I just installed Samba4 and starting down the learning curve by googling how to set it up. My file sharing is simple. Nothing special within the Lan, but want to keep hackers out from the outside world. I use Logmein on my windows computers and may want to setup a remote login for the server at some point. Right now my big this is setting up the drives, shares, and going headless. I would like to be able to just turn the thing on and not worry about it.
I've been trying to get CUPS working on a headless server, in order to share the printers connected to it with the rest of the network. The problem is that I can't add printers via the web interface, or rather, that I can log in, as it will not authorize my account. I've tried lppasswd, adding cupsys to the group 'shadow'(only to have the system tell me that there is no user called cupsys), and been editing the cupd.conf for hours. On top of all tha, I can't even get to the web-interface without running "/etc/init.d/cups restart"!
i have been scratching my head and repeatedly banging it against a wall
i have a lan: m0n0wall router ubuntu server (headless/no mouse/kbd/speakers, etc in attic) 4 desktop pc's 1 dlink dsm 520
my server is exactly that it has mysql/apache/mediatomb/php/cacti/phpmyadmin/samba/nfs/mpd) ignore all other service running on server & focus on MPD. all 4 desktop pc can connect to mpd and they can see the library. BUT they cannot play any track as it states there was an error with the audio device!i have tried all the examples on the web and none work what do i do? how do i configure it to use the LOCAL soundcard NOT the one that is not present in my server? all the examples are set to use the LOCAL(one on machine running MPD) sound card which does not work until i connect a speaker to my server which again is located in my attic and only has a pc speaker so this is NOT what i want. all 4 desktops do not have mpd installed as they do not need it. all media is stored on my server the gui's that connect to mpd do not have any configuration for a sound card
I've installed SliTaz Linux onto a headless server with the following partition scheme:
/dev/hdb1 ext3 main system /dev/hdb2 ext3 currently empty /dev/hdb3 swap
I would like to further configure this machine so that (1) a VirtualBox virtual machine is automatically started during boot with no user intervention, and (2) /dev/hdb1 /dev/hdb2 are encrypted with LUKS or the like.
Because /dev/hdb1 and /dev/hdb2 will be encrypted, I will need to decrypt them with my passphrase during boot. But since this will be a headless server, I need to SSH in before the main OS loads, i.e. I probably need to SSH into initramfs? So far, I've found (1) VBoxTool which starts a virtual machine during boot, and (2) and early-ssh which installs dropbear SSH server into initramfs. However, since all my partitions (/dev/hdb1, /dev/hdb2) will be encrypted, I need to decrypt them before the boot process can continue after initramfs. This seems to imply that I need to modify the initramfs process so that I can (1) login via early-ssh mentioned above, (2) enter my decryption passphrase, then (3) let the normal boot process continue.Therefore, my question is: how do I find and edit my initramfs so that I can add the needed decryption tools (what are they?) and be prompted to enter my passphrase via SSH during boot?
setting up a headless torrent server for the flat. The webui works great on transmission except I want to schedule it to pause during daylight hours. Essentially I only want it to be downloading between the hours of 2am and 8am.
Currently I cannot get it to pause. Only throttle the speed. Setting the speeds to zero appears to set it to unlimited speed. And we don't really want it running 1 kb/s constantly (lowest possible setting) during the day.
Server is running 8.04 LTS using Transmission 1.83. Currently I have the Gnome desktop installed and using remote desktop viewer to set some things up via a gui. CLI is faster... if you know what you're doing, which I don't really, so I like to fiddle. I did think about using Crontabs to kill the process at 8am and start it at 2am, however this is pointless as we'd like to be able to setup torrents via the webui during the day.
I have a headless ubuntu server (10.04 x64) in the attic that I use as mostly as a file server for media streaming and backing up other PCs on the home network. The box doesn't have a display, keyboard or mouse connected and is in a location where it's physically difficult to get at.
I'm mostly looking at a solution to protect the system if it were physically stolen(hopefully this is very unlikely) so that I could have a reasonable confidence that no one could boot a live CD and steal any private documents on there.
I don't feel like I need to encrypt the root file system (and I don't want to reinstall!) as most of the data on the system isn't sensitive (MP3s, video files etc).
I would like to protect the backups of the other systems as they have some personal documents etc on them,
I need to also balance this with convenience as I need the backups to continue to be automatic or they most likely won't get done at all.
Currently I have a samba share called "backup" that is password protected and points to /media/backup where a 2tb drive is mounted, this drive is used just for backups from the other systems so I would like to encrypt the whole drive.
I realise that I can't have the drive mount completely automatically as it will then be accessible to anyone who can boot the box. The compromise I would like would be that the drive is encrypted and then either:
1.) I log in via ssh and enter a command to mount the drive which prompts for a passphrase - drive stays mounted after I log out until manually unmounted or the system is rebooted
or
2.) On boot something runs automatically which then prompts me for the passphrase when I next connect via ssh - again the drive stays mounted after I log out until manually unmounted or the system is rebooted
I would prefer 2 as then I am less likely to forget to mount the drive after a reboot.
I have seen several tutorials using LUKS to encrypt the drive and manually mount but am wondering if it's possible to achieve 2. Also wondering if there are any implications for samba if the drive is not available straight away at boot, would I need to restart anything after mounting the encrypted drive for example?
I have a strange problem. I have set up a server to run without a keyboard, mouse, or video. If I boot it connected to a monitor, there is no issue. However, if I boot it without a monitor attached, the CPU just runs at 100%. Below is a shot of my HTOP from the machine: The weird part is that if you sort the processes by CPU%, they don't actually total 100%. I was trying to see what was pushing the CPU so high, but most of the time the running processes only total between 5 and 20%. Has anyone else seen the problem? Any idea what's happening? I'm worried that if I leave it running like that, the processor could burn up. My temporary work around is to boot it up with a monitor, then disconnect the monitor, but that's really inconvenient.
I'm currently using the default Gnome desktop environment on my server when I connect via VNC, but I've been trying to Openbox on Crunchbang lately and I rather prefer it for being so light. I'm a relative noob and don't want to cock anything up - what's the simplest way to switch to the Openbox environment for when I connect via VNC, and once that's working, whats the safest way to remove Gnome?
On my Seagate DockStar torrent box running Debian Squeeze, I want to send email when a torrent download is done. What email client is good if I want to send e-mail via gmail server?
