|
Neuntausend posted:And there's another thing: Since you are involved somewhat, do you know if AMD/nVidia are working on Wayland/Weston compatible driver blobs yet? Because in all honesty, I also can't see too many people switching to Wayland without those, and as I have come to know nVidia at least, they probably don't even care. First, I'll say that Red Hat contributes to open-source AMD and NVIDIA drivers, and for our simple desktop customers we recommend that users run those instead of proprietary drivers. For our customers for which the FOSS drivers are not good enough, I know that Red Hat does collaborate with NVIDIA, but I'm not involved in those meetings and I'm not up to date on the latest results from them. Don't quote me on this, but I think NVIDIA has said that they are working on EGL support for at least gbm. One of the things we've talked about having a very simple version of Xorg that loads NVIDIA's proprietary driver and uses it as an output layer for Wayland, but that was just one of those "it's so crazy it might just work" ideas, and there's a lot of questions about it (do we run a free EGL stack, or try to write a wrapper to convert EGL to GLX?)
|
#
?
Aug 10, 2013 19:59
|
|
|
|
| # ? May 28, 2024 23:38 |
|
|
On this topic, is there a good article explaining the different Linux environment/window manager/whatever X is stacks? It seems like there are a lot of competing implementations and I haven't seen a good explanation of what does what or what is replacing what. Wikipedia is pretty useless for this sort of thing...
|
|
#
?
Aug 11, 2013 05:00
|
|
|
I have a dedicated Centos 5 server which I can only access via SSH. I would like to install Centos 6. My research suggests that 'upgrading' my Centos 5 install to Centos 6 is a bad idea, and that I should instead cleanly install Centos 6. I will likely have to pay someone in the data center to manually install Centos 6. However, it would be very neat if I could somehow do this over ssh. (and have ssh then run automatically after installation, so I can still access it, of course) I can't seem to find any evidence that this is possible. Is there any way to do so?
|
|
#
?
Aug 11, 2013 09:05
|
|
|
I suppose you could get the installer images, stick them in /boot, update your grub to include them, with boot time options that will enable access over VNC with a pre-set password. (This also means you need to setup the network config and all that jazz in there) If you mess up your server won't boot normally anymore and you'd have to pay the datacenter operator to fix it for you but hey. e: although I think grub has an option to try to boot and then fall back to another after it fails. spankmeister fucked around with this message at 09:26 on Aug 11, 2013 |
|
#
?
Aug 11, 2013 09:11
|
|
|
evensevenone posted:On this topic, is there a good article explaining the different Linux environment/window manager/whatever X is stacks? It seems like there are a lot of competing implementations and I haven't seen a good explanation of what does what or what is replacing what. Wikipedia is pretty useless for this sort of thing... Haven't seen anyhing specifically on that topic, but you will find that info on installation guides of gentoo, arch and the like or wikis of larger distributions. In the end, it's not that difficult. Linux OSes are all about options. While there are certain quasi-standards, there are alternatives for almost everything. Only a few things (such as X) are without alternative (for now). X is your display server, it loads drivers for your graphics card, mouse, keyboard and enables other software to draw graphical stuff on your screen. As I said, right now there's no alternative, but since it's about 30 years old and somewhat messy, certain projects (Wayland, Mir) seek to create something more up to date. Window managers are just the programs that actually draw window borders (using your display server), handle the moving around, min-/maximising and such. Desktop environments are a combination of a window manager, bars, docks, tools (editor, calculator, settings manager, what have you), stuff you would expect to find on a desktop computer.
|
|
#
?
Aug 11, 2013 10:02
|
|
|
evensevenone posted:On this topic, is there a good article explaining the different Linux environment/window manager/whatever X is stacks? It seems like there are a lot of competing implementations and I haven't seen a good explanation of what does what or what is replacing what. Wikipedia is pretty useless for this sort of thing... I wrote http://blog.mecheye.net/2012/06/the-linux-graphics-stack/ a year ago for precisely this reason. I'm working on a followup, too, and if anybody has any questions, however dumb they might seem, just ask and I'll try my best to answer.
|
|
#
?
Aug 11, 2013 10:39
|
|
|
I'm trying to write a bash script for a Minecraft server that will tar three folders. This is what I have so far, but it fails when I try and use it, with the error "tar: Cowardly refusing to create an empty archive". What do I need to do to get the folders all saved? Folders I want to save, all in the same tarfile: ./world/ ./world_nether/ ./world_the_end/ code:
|
|
#
?
Aug 11, 2013 15:32
|
|
|
JerikTelorian posted:SOURCE=$PWD"/world/" $PWD"/world_nether/" $PWD"/world_the_end/" Are those supposed to be absolute paths?
|
|
#
?
Aug 11, 2013 16:33
|
|
|
JerikTelorian posted:I'm trying to write a bash script for a Minecraft server that will tar three folders. code:
|
|
#
?
Aug 11, 2013 16:46
|
|
|
Neuntausend posted:Are those supposed to be absolute paths? They are. When I first tried to write it, it didn't like me using relative paths for some reason. tonberrytoby posted:Run this: Did this, this is the output I got: code:
|
|
#
?
Aug 11, 2013 16:49
|
|
|
I guess I don't understand cpu load.%20-%20byobu_007.jpg) I can sit here and watch (h)top, and none of the cores have much load on them, and the processes all pretty much look like they do there, and yet the load average is 16.71 and the terminal over ssh is slow to respond, and the CPU is running at 2.4GHz instead of the throttled back 1.6GHz. What's going on here?
|
|
#
?
Aug 11, 2013 17:04
|
|
|
Load is the total number of processes that are running, processes that could run if something else wasn't running instead and processes that could run if they weren't waiting for IO. It looks like you have a lot of stuff waiting on IO (all those Ds in the S column).
|
|
#
?
Aug 11, 2013 17:12
|
|
|
Thermopyle posted:I guess I don't understand cpu load. Can you turn on TASK_IO_ACCOUNTING ? and then display it with atop. This should tell you. JerikTelorian posted:is the [set +x] some kind of debug mode? I was looking for that while trying to learn scripting. It looks like the source isn't being properly included, for some reason.
|
|
#
?
Aug 11, 2013 17:16
|
|
|
JerikTelorian posted:They are. When I first tried to write it, it didn't like me using relative paths for some reason. Yes, because you have SOURCE=path1 path2 path3, which will try to execute a program called path2 (with environmental variable SOURCE set and argument path3). What you want is probably SOURCE=(path1 path2 path3)
|
|
#
?
Aug 11, 2013 17:16
|
|
|
pseudorandom name posted:Load is the total number of processes that are running, processes that could run if something else wasn't running instead and processes that could run if they weren't waiting for IO. tonberrytoby posted:Load measures all resources. So if your CPUs are empty and your Memory is not totally full, then your disk must be overloaded. Thanks, that makes sense. Turns out that its mysqld, so I need to look in to that...
|
|
#
?
Aug 11, 2013 17:48
|
|
|
M31 posted:Yes, because you have SOURCE=path1 path2 path3, which will try to execute a program called path2 (with environmental variable SOURCE set and argument path3). What you want is probably I did this and it now runs, but it still only compressed the first folder (./world). This is what the code looks like now (I guess I didn't need the absolute path?) code:code:JerikTelorian fucked around with this message at 18:02 on Aug 11, 2013 |
|
#
?
Aug 11, 2013 17:57
|
|
|
Captain Pike posted:I have a dedicated Centos 5 server which I can only access via SSH. I would like to install Centos 6. My research suggests that 'upgrading' my Centos 5 install to Centos 6 is a bad idea, and that I should instead cleanly install Centos 6. Yes, this is possible. You can use the netinstall image in /boot to clean install (recommended). You can do a bunch of fuckery with Anaconda and a local repo plus GRUB2 to save everything, but you'd have just as much luck if you kickstarted the netinstall. You can upgrade 5 to 6, but it was a long development cycle. 5 can't even read RPMs from 6. You need to install newer versions of rpm, glibc, rpm-python, and a few other packages from Fedora 9 (I think, it's been a while) then do the same from F11 in order to get RPM on 5 to support the new RPM format, at which point yum installing works. Test this in a VM to make sure you get the process right, because if you don't, your system is unusable.
|
|
#
?
Aug 11, 2013 18:31
|
|
|
JerikTelorian posted:EDIT: I got it. Had to correct SOURCE to read SOURCE=("./world/ ./world_nether/ ./world_the_end/"), the internal quotes were loving it up. Thanks to everyone for the assistance. I'm sorry, I was using zsh, which has somewhat sane bahaviour. In bash you have to use "${SOURCE[@]}" (with the quotes) to output all the paths instead of just $SOURCE, which will only give you the first one. While you can use the solution you have now (in which case you don't need the parentheses), it will break if there ever is a space in one of the pathnames.
|
|
#
?
Aug 11, 2013 18:58
|
|
|
Crossposting this from support to see if anyone has an answer: I cannot preseed the portion of debian 7.1 install that deals with the wireless network name with the following and I'm wondering what I'm going wrong: d-i netcfg/wireless_show_essids manual [wireless_network_name] d-i netcfg/wireless_essid string [wireless_network_name] d-i netcfg/wireless_essid_again string [wireless_network_name] With these lines the network name is highlighted but not selected. My question is how can I actually select this network?
|
|
#
?
Aug 11, 2013 21:04
|
|
|
Thermopyle posted:I guess I don't understand cpu load.
|
|
#
?
Aug 12, 2013 03:24
|
|
|
Misogynist posted:Your 2 GB of used swap is the first place I'd look in terms of the disk I/O issues others have pointed out as the likely cause of your slowdown. If you're running sysstat, pgpin/pgpout and majflt/s should be logged in sar. You can also add IO to htop as well and then sort by something like read/write/total to find out the top disk activity.
|
|
#
?
Aug 12, 2013 05:36
|
|
|
At a user's request, I changed their home directory in LDAP to an NFS directory. This is a very common thing here. I can see the NFS mount on his machine and everywhere else, I can go to it, but it refuses to let me use it as a home directory: [root@MACHINE users]# su - USERNAME su: warning: cannot change directory to /export/nfs/users/USERNAME : No such file or directory if: Expression Syntax. [USERNAME@MACHINE users]$ echo $HOME /export/nfs/users/USERNAME [USERNAME@MACHINE users]$ cd $HOME [USERNAME@MACHINE USERNAME]$ pwd /export/nfs/users/USERNAME SElinux is set to permissive, there's no errors in /var/log/messages or /var/log/secure. Google is no help. This is in RHEL5 edit: Turns out I had a space after the home directory setting in LDAP. gently caress you, Red Hat. Adult Sword Owner fucked around with this message at 15:36 on Aug 12, 2013 |
|
#
?
Aug 12, 2013 15:31
|
|
|
Dead Inside Darwin posted:edit: Turns out I had a space after the home directory setting in LDAP. gently caress you, Red Hat. man 5 passwd This is not a Redhat problem. Wouldn't work on any distro.
|
|
#
?
Aug 12, 2013 16:34
|
|
|
Megaman posted:Crossposting this from support to see if anyone has an answer: Looking at other preseeds, do you have to specify managed or adhoc? # Choices: Infrastructure (Managed) network, Ad-hoc network (Peer to peer) d-i netcfg/wireless_adhoc_managed select Infrastructure (Managed) network
|
|
#
?
Aug 12, 2013 17:36
|
|
|
3spades posted:Looking at other preseeds, do you have to specify managed or adhoc? I don't believe so, but I'm not sure I also can't seem to get d-i partman-auto/disk string /dev/sda to work... Can anyone else verify this?
|
|
#
?
Aug 12, 2013 17:50
|
|
|
So I have a weird problem with my Chromebook (Samsung Series 5 550 Intel). I installed Ubuntu on it, and it's working great, but I'm trying to setup a swap file. Eveything I do ends up with this output: code:Ubuntu 12.04, 3.8.11 kernel.
|
|
#
?
Aug 12, 2013 23:04
|
|
|
So I have a little project going on involving ftp. Essentially I need to watch for an ftp append on a file and if it happens then I append that file to another on a remote machine. (So it would act as a buffer or ftp "cache" in case the usual connection to the remote machine goes down). My idea would be to use a bash script to check the logs for a successful append or just check the mod time on the file for a change and if so, ftp append to the remote file. Then tie the script to a cron job and run every so often or turn it into a service so it runs constantly? Not sure if that can be done or if it's even worth it. Is there a better way to do this? This seems the most straight forward unless there's a tool out there that would do it better. I'm using Ubuntu 12.04 with proftpd.
|
|
#
?
Aug 13, 2013 00:17
|
|
|
Master Stur posted:So I have a little project going on involving ftp. Essentially I need to watch for an ftp append on a file and if it happens then I append that file to another on a remote machine. (So it would act as a buffer or ftp "cache" in case the usual connection to the remote machine goes down). My idea would be to use a bash script to check the logs for a successful append or just check the mod time on the file for a change and if so, ftp append to the remote file. Then tie the script to a cron job and run every so often or turn it into a service so it runs constantly? Not sure if that can be done or if it's even worth it.
|
|
#
?
Aug 13, 2013 00:33
|
|
|
Jerk McJerkface posted:So I have a weird problem with my Chromebook (Samsung Series 5 550 Intel). I installed Ubuntu on it, and it's working great, but I'm trying to setup a swap file. It suggests 644 is insecure (allowing group/other read access) so chmod 600 /swap/swap_file and put your swap entry in /etc/fstab. Then just # swapon -a
|
|
#
?
Aug 13, 2013 00:42
|
|
|
Apparently some of the issues with gnome 3's memory usage (if not leakage) are related to the system monitor extension, particularly the graphs it makes. That's too bad, because it's pretty handy... I guess I'll try disabling it for a while to see what happens.
|
|
#
?
Aug 13, 2013 02:30
|
|
|
On https://help.ubuntu.com/community/IptablesHowTo the first example is "Allowing Established Sessions" - what exactly does that rule do?
|
|
#
?
Aug 13, 2013 06:41
|
|
|
fletcher posted:On https://help.ubuntu.com/community/IptablesHowTo the first example is "Allowing Established Sessions" - what exactly does that rule do? It means that it allows traffic from sessions that have already been established. For example if the default incoming policy is drop (it should) then any session you try to open to say, a web server at port 80 will fail. The server will receive your packets but the packets it sends back will be dropped. The established rule will track that connection and allow the traffic in. This means you don't have to open up a whole block of ports just to talk to the outside.
|
|
#
?
Aug 13, 2013 08:01
|
|
|
Misogynist posted:You can use inotifywait to watch for close_write events really easily with a shell script. The downside of the inotify approach is that if your service stops for some reason while files are still being uploaded over FTP, the changes won't be caught. Awesome thanks! I must have been using the wrong keywords because I didn't see that at all while googling. I don't expect the service to quit out on me, but it doesn't need to be up 24/7 either. The odds of enough failures happening for this thing to even get used are pretty slim.
|
|
#
?
Aug 13, 2013 13:20
|
|
|
fivre posted:There's nothing like http://www.google.com/inputtools/windows/ for Linux, is there? setxkbmap -option grp:switch,grp:alt_shift_toggle,grp_led:scroll us,ru ,phonetic Something like that. You can hard-code it as well. If you're running XFCE you can actually set this in the keyboards menu. You want the Russian-Phonetic option.
|
|
#
?
Aug 13, 2013 15:00
|
|
|
Martytoof posted:Yup. it's ridiculously easy if you're just using a default 9600,8,N,1: screen /dev/[YourDevice], then you can do all the useful screen stuff like disconnecting, and kill the session when you're done I don't know how to change parity etc but practically everything is 8n1
|
|
#
?
Aug 14, 2013 00:03
|
|
|
Is there a better way to change what part of the screen my drawing tablet maps to? The way you do it with xsetwacom is strange and confusing.
|
|
#
?
Aug 14, 2013 00:03
|
|
|
Recent versions of gnome-control-center have a custom calibration screen that works like a charm.
|
|
#
?
Aug 14, 2013 02:55
|
|
|
OWLS! posted:setxkbmap -option grp:switch,grp:alt_shift_toggle,grp_led:scroll us,ru ,phonetic I've set this up for clients in my last job, its based off the inotify facility and runs as a daemon https://code.google.com/p/lsyncd/
|
|
#
?
Aug 14, 2013 03:16
|
|
|
So my dvd drive died and I replaced it with a nice shiny blu-ray/dvd/cd etc. drive. Under VLC in Windows 7 I can play dvds without any issue but under Fedora 19.. the sound is full of static and the video is itself is really choppy. It's weird because it's not like the frames are delayed, the playback reminds me of a ruined vhs tape. What's the best way to troubleshoot this? dmesg doesn't seem to have any errors regarding cdrom or /dev/sr0. /proc/sys/dev/cdrom/info just lists the known facts about the drive. I'm at a loss and may re-install Fedora as a last ditch effort which is horrible because I can be lazy. 
|
|
#
?
Aug 14, 2013 03:49
|
|
|
|
| # ? May 28, 2024 23:38 |
|
|
I am attempting to use my phones GPS over bluetooth to connect to my system. Everything works except gpsd which keeps giving me the error that the device (/dev/rfcomm0) is already in use. Which of course it is due to it being connected via bluetooth, I at least assume bluetoothd is the process that is using it. Is there any way to force it to overlook the process that is already using it? Or any other GPS program out there that does not give a poo poo if something else is using it? I am running kali linux which is based off debian, well it basically is debian with a reskin. I also for whatever reason could not manually set up rfcomm and it just tells me that the connection is refused but blueman is able to connect it up just fine.
|
|
#
?
Aug 14, 2013 06:05
|
|
