|
Is there a straightforward way to tell xinit to start a new X server without explicitly telling it which number server to use? For example "xinit blah -- :1" will start blah on a X server :1 and will create server :1 if it does not already exist. But what I want is an xinit command that starts a completely new X server every time it is invoked, regardless of which or how many X servers are already running. Can I do this? Alternatively is there a command (other than parsing the output of ps) that will list the currently running X servers so that I can script such a thing?
|
#
¿
Mar 30, 2007 20:25
|
|
|
|
| # ¿ Apr 26, 2024 03:33 |
|
|
Captain Cool posted:My Ubuntu Edgy box seems to have died today. I was gone for about five hours, and when I came back the monitor wouldn't come out of powersave, like it wasn't receiving a signal. Then I noticed the hard drive light was on steady, though I couldn't hear the drive. I thought my new CPU fan had died but it was still working. No response to ssh. Caps lock didn't toggle the light. At exactly what point is your machine failing? You said "no beeps, no video" - I take that to mean that the BIOS isn't even doing it's power-on self test. If you're not even getting a POST or at least a POST beep, your problem has nothing to do with Linux and no panic keys in the world will help you. Incidentally, how are you planning to flash your BIOS if your machine does not show you the BIOS screen? If this is the case, basically there are two possibilities: your power supply is dead, or your motherboard is dead. If your CPU, RAM, or video card were dead, your mobo would be beeping its head off. If anything else were dead, you'd definitely get a POST and probably boot to Linux in some capacity. If you hear fans whirring and hard drives spinning up, it's probably your motherboard. If you hear nothing, it's probably your power supply.
|
|
#
¿
Apr 9, 2007 16:29
|
|
|
Captain Cool posted:Fans spin fine, but I can't tell if the hard drive is spinning. It's a mini-itx system, with soldered-in CPU, so I hope it's not the motherboard. I see what you're saying though. What you are doing when you take the battery out is resetting the CMOS. Most motherboards provide a jumper that you can set to do this instantaneously so that you don't have to wait for the battery to drain. There is a chance that this will fix our problem if your issue is that there is a setting in the BIOS which is causing the BIOS not to be able to load properly.
|
|
#
¿
Apr 12, 2007 05:40
|
|
|
rugbert posted:Nothing. I had to install some drivers to read NTFS drives and when I went to go look at what drives were available to mount I could only see the two other dual boot partitions (WinXP and Server 2003). The second harddrive is on the same IDE cable and the humper settings are set to slave so I dont know why fedora cant see it. The second hard drive is /dev/hdb The letter after hd identifies the physical drive, and the number identifies the partition. For example, hda3 is the third partition of the first drive, and hdb1 is the first partition of the second drive, etc. Now, /dev/hda2 is an "extended" (Ext'd) partition, which is something that windows does to compensate for hardware limitations. It's basically a container for other partitions - you can't mount it. The information that you posted shows that your first hard drive (/dev/hda) has two real windows partitions (/dev/hda1 and /dev/hda5). Look at the output of /sbin/fdisk -lu /dev/hdb, and see which partition describes itself as "HPFS/NTFS" - it's probably /dev/hdb1. That's where your music is.
|
|
#
¿
Apr 17, 2007 21:29
|
|
|
swalk posted:I have a dedicated server with 5 IPs. Whenever I SSH in under my user account, the IP that "bash" or my ssh session uses is *.54, but when another user logs in they get a different IP. This seems to be consistent because I have always gotten .54 If you have root-level access to the server, you can change the IP that the SSH daemon listens on. Edit the sshd configuration file (usually /etc/ssh/sshd_config) and add the line "ListenAddress <desired-ip>".
|
|
#
¿
Apr 17, 2007 21:36
|
|
|
Faustus D.D. posted:I need to put a path into a program manually but the path in question is not on my primary drive its on a USB external drive. In the future, the console command to show all mounted devices and mount points is 'df'. For example: code:This shows, among other things, that the USB thumbdrive that I just plugged in is mounted on /media/CARD DISK. There is probably a more user-friendly way to show this in your GUI, but that depends on which GUI apps you are using. In Konqueror, you can go to media:/, which will shwo you all mounted devices. You can right click, choose properties, and there's a "Mounted On:" line.
|
|
#
¿
Apr 22, 2007 02:49
|
|
|
rugbert posted:Anyone have anytips on how to install Nvidia GeForce fx 5200 drivers? Theyre listed in the yum but I used the GUI package updater because Im kinda nervous but got nothing but conflict errors... I wanna pretty up my desktop! For me it was as simple as installing kmod-nvidia. Can you provide a bit more information, such as which distribution you are using (guessing FC6), which repository you are getting the drivers from, and exactly what conflict messages you are receiving?
|
|
#
¿
Apr 25, 2007 05:16
|
|
|
huge rear end zit posted:Say I've got a 120gb Hard Drive, partitioned into 2, and I want to add it my linux box, more specifically Debian. How would I go about this if the 120gb is formatted in NTFS? To add to the mix, I've got a spare 250gb hard drive that's blank that can be used in this.... process. Do you want to just access it as it is (NTFS), or convert the partition to something like ext3 and preserve the data? If the former, there is now a full capability (read/write/etc) driver for NTFS on linux. It's called NTFS-3g and is available for Debian If you want to change the format and save the data, you need to use that extra drive, because you can't do a non-descructive reformat. Hook up both drives, and do this: 1. Create a 120GB or larger partition on the 250GB drive 2. Format that partition with ext3 (or your favorite FS) 3. Copy all data from the 120GB drive to the 250GB drive [note that this requires ntfs-3g!] 4. Delete all of the partitions on the 120GB drive 5. Create whatever partitioning setup you want on the 120GB drive 6. Format those partitions with ext3 (or your favorite FS) 7. Copy all data from the 250GB drive back to the 120GB drive 8. Do whatever you like with the 250GB drive In Linux, the tool to create or delete partitions is fdisk. To create a partition on the first IDE hard drive, run "fdisk /dev/hda", similarly for hdb, or sda or whatever. fdisk is pretty straightforward, once you press ? to see what the commands are. The tool to format a partition is mkfs.<filesystem-name>. So, for example, "mkfs.ext3 /dev/hda1" to format the first partition of the first IDE drive with ext3. There are lots of nifty options you can pass to mkfs.whatever, such as volume label, so read the man page if you're interested. I suggest that if you don't plan to use this drive with windows anymore that you do the second process of reformatting the disk. Running a linux system that accesses lots of files on a drive that is formatted with a *nix-unfriendly filesystem is possible, but really annoying after a while.
|
|
#
¿
Apr 25, 2007 15:52
|
|
|
huge rear end zit posted:Thanks for the advice, I didn't know there was a package available to read/write to an NTFS partition in linux! I'll most definitely give this a go once I have this assignment out of the way. Fortunately you probably don't need to upgrade your kernel to get ntfs-3g working. ntfs-3g is based on FUSE, which is a generic kernel interface for many filesystem drivers. The only thing you need to do to your kernel is to install the FUSE driver, if you don't already have it. The FUSE driver (like almost all drivers) can be compiled as a module against your existing kernel. Look at the information here: http://fuse.sourceforge.net/ You need to have the headers and sources for your existing kernel installed, and you have to run the configuration script on your kernel sources, but you do not have to actually compile or replace your kernel.
|
|
#
¿
Apr 26, 2007 15:06
|
|
|
rugbert posted:Um Im using the yum repository (I think thats right Im still new at this) with Fedora Core 6. Heres whats available through yum tho: Firstly, the repository you are using is called "livna" - yum is the program that installs software from various repositories. Livna is probably the most popular third-party repository for Fedora Core Linux. It provides software that the official repositories (Core and Extras) do not provide because of licensing issues - for example, the nvidia drivers are closed-source, and therefore not available through the official channels. The front page of the livna repository website (https://rpm.livna.org) says that this error is actually due to a bug in the FC6 installer which installed the wrong kernel on your machine by accident. The fix instructions are here: http://fedoraproject.org/wiki/Bugs/FC6Common#head-e0676100ebd965b92fbaa7111097983a3822f143 Follow those instructions and then try installing kmod-nvidia again.
|
|
#
¿
Apr 26, 2007 15:11
|
|
|
rugbert posted:OK I ran this line There's nothing wrong with having multiple kernels installed at once - the kernel packages are smart enough that they don't conflict with one another. You are still only running one kernel at once. You can figure out which kernel you are running by typing "uname -r" at a command prompt. Usually linux package managers will not automatically remove old kernel versions when you upgrade. This is for two reasons: 1. Because, while usually kernel upgrades go without issue, if there is a problem, it could prevent your system from booting entirely. That's why they keep the old kernel(s) around by default, so that you can always get your system booted to fix any problems. 2. You can't switch kernels without rebooting, so you can't delete the files for the old kernel before booting into the new kernel. Any way of automating that would be prone to a lot of problems. Once you boot into your new kernel and you know it is working properly, you can safely remove the old kernel package explicitly. Smackbilly fucked around with this message at 15:33 on Apr 27, 2007 |
|
#
¿
Apr 27, 2007 15:28
|
|
|
deathmerc posted:Upgraded my linux server from ubuntu server 6.10 to the latest and it seems to have eaten one of my drives. It shows up when the machine POSTS and in the BIOS but ubuntu doesn't want to see it anymore. sdx devices are SCSI or SATA devices. IDE devices will be called hdx. I have no idea what the third sdx device is, if you say you only have two SATA drives. I would normally suspect it might be an optical drive doing SCSI emulation, but there wouldn't be any partition devices listed for that.
|
|
#
¿
May 1, 2007 16:35
|
|
|
Lexical Unit posted:Nope. I happen to know the answer I'm looking for is 10.8.16.1, but route gives rows 10.8.0.0 and 169.254.0.0 under "Destination" and under "Gateway" I just see * for those two entries and my computer's name for the "default Destination." If that's true then there's something's configured wrongly. The last line of the route command should say "default" in the desitnation column, and the gateway column should be the name or IP of your default gateway.
|
|
#
¿
May 1, 2007 22:25
|
|
|
Lexical Unit posted:I assure you nothing is configured wrongly as this is a work computer, is one of thousands, and all of them are working perfectly well. Perhaps our network is configured in an atypical fashion. Can you post what the output of route actually is? It seems like you're describing something like this: code:
|
|
#
¿
May 1, 2007 23:13
|
|
|
Minor minicom question: I have a computer running headless that I am connecting to via a null-modem cable using minicom. The connection establishes perfectly, but this happens: code: Anyone have a clue how to fix this?
|
|
#
¿
May 10, 2007 07:31
|
|
|
teapot posted:What do you mean, does not connect at all? It will connect, but since it didn't send anything, you will see empty screen until you type something. Ah, you're right. The issue was I didn't see "login:", so I didn't think it was expecting input. I figured out that if I replace the init string with a ^L, it shows the "login:", but doesn't enter anything. Thanks.
|
|
#
¿
May 10, 2007 08:05
|
|
|
teapot posted:And what happened to ssh? My usual directory transfer procedure is Is there any advantage to this over scp with the -r and -C options?
|
|
#
¿
May 10, 2007 08:08
|
|
|
DeathChill posted:So, Compiz and Beryl, they're actually window managers and desktops and not just imaging tools (like Aqua). They're window managers only. If you want a desktop environment, you will also need to install one of those, such as KDE, Gnome, Xfce, etc.
|
|
#
¿
May 11, 2007 04:33
|
|
|
lol internet. posted:What's the best way to rename a user account? (so it changes the profiles/permissions correctly as well? That should be fine. Permissions and file ownership are done by user ID, not username, and the above lusermod command doesn't change the user ID. What went wrong when you did that?
|
|
#
¿
May 13, 2007 23:39
|
|
|
I have some files in a personal subversion repository, and I want to place them in a shared subversion repository - how can I do this while retaining the revision history for these files? I know there will be no conflicts between the files and anything that already exists in the destination repository. Is there a way to insert my files into the target repository so that if the current revision of the target repository is x, then revision y of the source repository becomes revision x+y of the source repository? I know I could script this if need be, but I'm hoping that subversion has some kind of built-in facility for this that I haven't been able to find yet.
|
|
#
¿
Jun 2, 2007 21:32
|
|
|
Jo posted:I've got ~1000 files in a folder. I was wondering if there's a way to run a command for each of the files in a folder, appending the output. (Under bash.) for i in `ls ./06CityProblems`; do ./JTSPSolver < ./06CityProblems/${i} >> ./outFile; done More than one way to skin a cat, especially in Linux. 
|
|
#
¿
Jun 7, 2007 20:08
|
|
|
Jo posted:Thanks to both of you. Works wonderfully. Yes and yes. The value that you give to nice is a value to add to your parent process' priority (if you are starting a program directly from the shell, your parent is the shell). This means that if, for example, your shell is being run with niceness 10 and you run a process with nice -n -19, the niceness of that process with actually be -9. But the answer to your question is that if you run the script nice -n -19, the inner program will by default run at a nice offset of 0 from its parent (the script), which means that it will indeed run at the same priority as its parent. If you want to change the priority of a running process, there is a command "renice".
|
|
#
¿
Jun 8, 2007 07:07
|
|
|
JoeNotCharles posted:For a home computer, I wouldn't bother. The main point of separating them would be so that if one of your partitions gets corrupted, you don't lose the other one. Obviously, this has the most effect if the other partition is actually on a separate drive, which it sounds like it won't be. It is always useful to have your OS files and your personal data on a separate partition, for any OS. In the case that you want to install a different distribution, or do a clean install of a new version of your distribution, you can simply format the / partition and start clean rather than mucking around with trying to wipe out your OS files and keep your personal files safe. Anal Wink posted:Question: I'm about to install on a 700gb drive. It'll mostly be used for a media share for my xbox, but I'm also going to play around with Xen. Most installations of Linux aren't going to need more than an 8GB or so root partition. However, given that that drive is about as big as a moon of Jupiter, you probably want to be really generous with how much space you allocate to /, just so you don't run into problems later. I'd go with maybe 50GB. I would suggest keeping Xen images either in /home or in a third partition that is separate from / and /home, so that you can still safely reformat / if necessary without having to back up anything. By the way, the easiest way to solve the "how big do I make my partitions?" question is to use Linux Volume Management, which essentially allows you to resize your partitions if you need to. If you're willing to take a few minutes to read a HOWTO on LVM, that's really the way to go.
|
|
#
¿
Jun 10, 2007 03:19
|
|
|
Professor Science posted:I've been running Fedora 7 for a few days on a workstation at work and I like it a lot. I'm tempted to try it on my desktop here and replace Ubuntu for the time being, but I'd like to do a network install. Does anyone know if the standard nForce 4 Ethernet controller is supported by the installer? Yes it is.
|
|
#
¿
Jun 11, 2007 04:15
|
|
|
Professor Science posted:<3 I don't know - I just know I happened to do a net install using an nForce4 built-in NIC last week. Most NICs should work just fine under modern Linux distributions, though. Hardware compatibility isn't nearly as big a problem as it used to be.
|
|
#
¿
Jun 11, 2007 04:45
|
|
|
Peven Stan posted:Quick question: If I plan on dual booting Linux and XP using GRUB as the bootloader, what's the order that I should install the two in? Is it possible to add a current XP install in the bootloaders as well? You can install them in either order, but if you install Linux first, XP will overwrite the boot loader and you will have to use a Rescue CD or Live CD to get back into Linux to restore grub. So it's easier (but not necessary) to install Windows first. And yes you can add existing Windows installations to the loader. See here: http://www.linuxselfhelp.com/gnu/grub/html_chapter/grub_4.html
|
|
#
¿
Jun 15, 2007 00:58
|
|
|
The Human Cow posted:
I'm assuming that 192.168.0.1 is your router. This means your router's DNS server is telling you the .36 address. If you are sure that that address is wrong, then the router must be in some sort of confused state and probably needs to be reset. One possibility might be if you just recently changed ominon from DHCP to static ip without releasing its lease, the router might still think that ominon is located at .36 because it still has a DHCP assignment to that effect.
|
|
#
¿
Jun 15, 2007 01:51
|
|
|
beuges posted:I've got a weird problem. Just set up a gentoo box - everything's been updated with the latest and greatest as of a day or two ago. The problem is, every time I restart my computer, my nameserver details in /etc/resolv.conf disappear. It's being overwritten by your dhcp client. You need to edit your dhcp client config file for the interface question and add the line "append domain-name-servers 66.18.68.1". I can't remember where this file is on Gentoo, but on Fedora it's /etc/dhclient-eth0.conf (or replace eth0 with whichever interface). If you need the "domain" line to appear as well, there are lots of other directives you can put in the dhcp client config file. man dhclient.conf
|
|
#
¿
Jun 24, 2007 19:06
|
|
|
Begby posted:Holy poo poo! That is goddam brilliant! If you like that, check out the the description of "less" given in the name section of its man page.
|
|
#
¿
Jun 28, 2007 21:46
|
|
|
How do I watched log?  I've never used logwatch before, and I'm trying to set it up. The problem is, I can't get it to give me anything but disk usage and network interface information. It doesn't tell me jack about the logs. I try: code:Naming specific services like --service secure doesn't work either. That gives me no output whatsoever. What am I missing? I'm on Fedora 8, if it matters. Edit: Nevermind. Turns out that the problem was that I didn't have a necessary perl module installed. It would've been nice if it, y'know, said something about that. Smackbilly fucked around with this message at 08:38 on Nov 26, 2007 |
|
#
¿
Nov 26, 2007 08:03
|
|
|
Sergeant Hobo posted:here's the ls -l output for the directory in question: Because your user is not the same user as "public". The permissions on that directory are read/write/exec for the owner and only read/exec for group and world. At the moment, the only user who can write to that directory is the user "public". If you want to allow everyone to write to this directory, you need to give it world-write permissions - chmod 777.
|
|
#
¿
Dec 13, 2007 06:24
|
|
|
Sergeant Hobo posted:So there's no way to give write permissions only to a user who is not the owner? Or am I going about this the wrong way entirely? Although it seems you fixed your problem, the way to give write permissions to 1 or more users who are not the owner is to create a group, put the users who should have write permissions into that group, assign group ownership of the directory to that group, and then give the directory group write permissions. If you need anything more fine-grained than that, you need to look into ACLs.
|
|
#
¿
Dec 14, 2007 14:23
|
|
|
Fortuitous Bumble posted:I think I accidentally bizarro-killed my linux This error looks like the module was compiled for a version of the kernel which had hardware random number generator support, and the kernel that you are running was not compiled with hardware random number generator support. In general, modules have to be loaded into exactly the kernel that they were compiled against - any deviation will often cause errors like this. Did you reboot after re-compiling the kernel to load the new one? Did you forget a step when re-compiling your kernel (such as make modules, or make modules install)? Did you do anything else weird?
|
|
#
¿
Dec 20, 2007 06:53
|
|
|
hooah posted:I tried to change the default OS in Grub (editing menu.lst), but the text editor said I don't have permission. Ok, pull up the terminal, sudo edit the file, and I get this response: The MS-DOS EDIT program does not exist in linux. I don't know what the "edit" executable on your system is, but it seems like it has to do with e-mail, which is not what you want. You want to use an editor such as vi, emacs, pico, nano, etc. Pico or Nano will probably be the most intuitive for a new user who just wants to quickly edit a config file. Try "sudo nano menu.lst" Edit: Or you can use whatever GUI editor you're used to, like "sudo gedit menu.lst" (or kate, or whatever you use in place of gedit). Smackbilly fucked around with this message at 00:37 on Dec 21, 2007 |
|
#
¿
Dec 21, 2007 00:35
|
|
|
Your Japanese Dad posted:When I was running the newest Ubuntu release I was having some stability issues. Random lockups where the whole system would freeze hard, and of course my logs never showed anything. However I've moved to Fedora 9 and have not had any of the problems. I really like how Ubuntu worked, and I'm sure I could just go back to 7.10, but I've gotten to the point to where it's all the same to me at this point. Oddly enough I had the exact opposite issue. I had been using Fedora since 6, and when I installed 9 (among other more minor issues) I would get random lockups, freezes, and reboots. I didn't really bother to investigate it, though. I figured it was a good excuse to try out Ubuntu, and I like it quite a bit so far - and no freezes.
|
|
#
¿
May 19, 2008 23:11
|
|
|
Does anyone else experience the problem with Amarok where the episodes of a podcast show up in pretty much random order? When I subscribe to a podcast, the episodes under the podcast heading in the playlist sidebar show up in seemingly random order, and they re-shuffle randomly every time I click the "playlists" header (the one which should alternate sorting between forward and reverse order). Does anyone know what causes this and/or how to fix it? This makes the "keep only x episodes of a podcast" option entirely useless because it will basically end up keeping x randomly selected episodes, rather than the most recent x episodes.
|
|
#
¿
May 25, 2008 18:15
|
|
|
Kim Jong-Chill posted:Im using Ubuntu 8.04 and when I delete something it will tell me that it cant be moved to the trash so do I want to delete it immediately and I click yes and then it says that it cant be deleted at all. So Im stuck with some files that I cant remove. Any help? Try deleting it from the command line (using rm <file>), which may work, or may at least give you a more informative error message.
|
|
#
¿
Nov 30, 2008 10:29
|
|
|
|
| # ¿ Apr 26, 2024 03:33 |
|
|
Onken posted:Hey is there any way I can use the command line "zip" program to batch-zip a folder worth of files into seperate zip files? Or do I need another program for that? Cheers. You could use a simple loop code:
|
|
#
¿
Jan 4, 2009 15:31
|
|