[Ltsp-discuss] performance problems

[Helmut Lichtenberg]

Hi,
we are running aroung 30-40 thin clients on a Debian Squeeze image with NBD
instead of NFS since years. The window manager is KDE and the setup is
reasonably fast.

As time came to switch to Wheezy I tested our heterogeneous hardware with the
new image (Wheezy with some backports, especially the xserver-xorg packages
and kernel 3.16.0-0.bpo.4-686-pae, also NDB, LTSP version 5.4.2-6+deb7u1).
Unfortunately some massive performance problems appeared, especially
concerning graphics throughput.

This does not mean some sophisticated 3D games or lots of HD videos running on
the clients. If you simply open a konsole on an empty desktop, grasp the
window border and move around the whole window, a long tail of window frames
is drewn behind like the tail of a comet.
While writing some text in an editor, the display of the characters lags
notably behind typing.

This happens on client hardware like:
* openthinclient 1697, Intel Atom CPU N270 1.60GHz, 2GB RAM,
Intel Mobile 945GSE Express Integrated Graphics Controller (rev 03)
* HP, AMD G-T56N Processor 1650MHz, 2GB RAM,
AMD nee ATI Wrestler [Radeon HD 6320] Graphics Controller
* terra, VIA C7-M Processor 1200MHz, 1GB RAM,
VIA CN700/P4M800 Pro/P4M800 CE/VN800 [S3 UniChrome Pro] Graphics Controller

It helps a little, when the clients are connected via gigabit ethernet, but
not really much. And it contradicts the idea of thin clients to require top
hardware and a really fast network to work smoothly.

So, what causes these problems?

In KDE I switched off all eye candy like animations.

The lts.conf defaults are:
[default]
SYSLOG_HOST = 10.1.3.33
LDM_DIRECTX = True
LDM_LIMIT_ONE_SESSION = True
LDM_LIMIT_ONE_SESSION_PROMPT = True
X_COLOR_DEPTH = 16
LOCALDEV = True
NBD_SWAP = True
X_RAMPERC = 90
LDM_LANGUAGE = de_DE.utf8
LANG = de_DE.utf8
CONSOLE_KEYMAP = de
XKBLAYOUT = de
XKBVARIANT = nodeadkeys
SOUND = True
VOLUME = 85
MIC_VOLUME = 75
TIMESERVER = time.tzv.local
RCFILE_01 = "test -x /etc/init.d/rsyslog && /etc/init.d/rsyslog restart"
RCFILE_09 = "exit 0"

The terminal server (which is separate from the NBD server) has lots of CPU
(currently 10) and RAM (24GB) for one user, me. :^)
It also runs Debian Wheezy.

I am running out of ideas, where to look and how to investigate it further.
Sadly it stops the great wheezy show.

Any help is appreciated deeply.

Regards
Helmut

--
-------------------------------------------------------------------------
Helmut Lichtenberg <***@fli.bund.de> Tel.: [phone]
Institut für Nutztiergenetik (FLI) 31535 Neustadt Germany
-------------------------------------------------------------------------

------------------------------------------------------------------------------
Dive into the World of Parallel Programming. The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials and more. Take a
look and join the conversation now. http://goparallel.sourceforge.net/
_____________________________________________________________________
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help, try #ltsp channel on irc.freenode.net

[James McQuillan]

My first guess would be the graphics drivers that have been automatically
chosen for those thin clients.

If it's using the generic driver instead of the specific driver that's best
for each chipset, you'll see graphic performance like you describe.

Perhaps you can get a shell running on the clients and look at the
/var/log/Xorg.0.log file. That'll show you
which driver was chosen.

Unfortunately, I've been out of the game too long to help much more than
that.

Jim McQuillan

On Fri, Feb 13, 2015 at 5:04 AM, Helmut Lichtenberg <***@tzv.fal.de> wrote:

> Hi,
> we are running aroung 30-40 thin clients on a Debian Squeeze image with NBD
> instead of NFS since years. The window manager is KDE and the setup is
> reasonably fast.
>
> As time came to switch to Wheezy I tested our heterogeneous hardware with
> the
> new image (Wheezy with some backports, especially the xserver-xorg packages
> and kernel 3.16.0-0.bpo.4-686-pae, also NDB, LTSP version 5.4.2-6+deb7u1).
> Unfortunately some massive performance problems appeared, especially
> concerning graphics throughput.
>
> This does not mean some sophisticated 3D games or lots of HD videos
> running on
> the clients. If you simply open a konsole on an empty desktop, grasp the
> window border and move around the whole window, a long tail of window
> frames
> is drewn behind like the tail of a comet.
> While writing some text in an editor, the display of the characters lags
> notably behind typing.
>
> This happens on client hardware like:
> * openthinclient 1697, Intel Atom CPU N270 1.60GHz, 2GB RAM,
> Intel Mobile 945GSE Express Integrated Graphics Controller (rev 03)
> * HP, AMD G-T56N Processor 1650MHz, 2GB RAM,
> AMD nee ATI Wrestler [Radeon HD 6320] Graphics Controller
> * terra, VIA C7-M Processor 1200MHz, 1GB RAM,
> VIA CN700/P4M800 Pro/P4M800 CE/VN800 [S3 UniChrome Pro] Graphics
> Controller
>
> It helps a little, when the clients are connected via gigabit ethernet, but
> not really much. And it contradicts the idea of thin clients to require top
> hardware and a really fast network to work smoothly.
>
> So, what causes these problems?
>
> In KDE I switched off all eye candy like animations.
>
> The lts.conf defaults are:
> [default]
> SYSLOG_HOST = 10.1.3.33
> LDM_DIRECTX = True
> LDM_LIMIT_ONE_SESSION = True
> LDM_LIMIT_ONE_SESSION_PROMPT = True
> X_COLOR_DEPTH = 16
> LOCALDEV = True
> NBD_SWAP = True
> X_RAMPERC = 90
> LDM_LANGUAGE = de_DE.utf8
> LANG = de_DE.utf8
> CONSOLE_KEYMAP = de
> XKBLAYOUT = de
> XKBVARIANT = nodeadkeys
> SOUND = True
> VOLUME = 85
> MIC_VOLUME = 75
> TIMESERVER = time.tzv.local
> RCFILE_01 = "test -x /etc/init.d/rsyslog && /etc/init.d/rsyslog
> restart"
> RCFILE_09 = "exit 0"
>
> The terminal server (which is separate from the NBD server) has lots of CPU
> (currently 10) and RAM (24GB) for one user, me. :^)
> It also runs Debian Wheezy.
>
> I am running out of ideas, where to look and how to investigate it further.
> Sadly it stops the great wheezy show.
>
> Any help is appreciated deeply.
>
> Regards
> Helmut
>
> --
> -------------------------------------------------------------------------
> Helmut Lichtenberg <***@fli.bund.de> Tel.: [phone]
> Institut fÃŒr Nutztiergenetik (FLI) 31535 Neustadt Germany
> -------------------------------------------------------------------------
>
>
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming. The Go Parallel Website,
> sponsored by Intel and developed in partnership with Slashdot Media, is
> your
> hub for all things parallel software development, from weekly thought
> leadership blogs to news, videos, case studies, tutorials and more. Take a
> look and join the conversation now. http://goparallel.sourceforge.net/
> _____________________________________________________________________
> Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
> https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
> For additional LTSP help, try #ltsp channel on irc.freenode.net
>

[Karl Duc - Vormingplus regio Brugge]

------------------------------------------------------------------------------
Dive into the World of Parallel Programming. The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials and more. Take a
look and join the conversation now. http://goparallel.sourceforge.net/

[Helmut Lichtenberg]

James, Karl,
thanks for your replies.

James McQuillan schrieb am 13.02.2015 14:53:
> My first guess would be the graphics drivers that have been automatically
> chosen for those thin clients.
>
> If it's using the generic driver instead of the specific driver that's best
> for each chipset, you'll see graphic performance like you describe.
>
> Perhaps you can get a shell running on the clients and look at the
> /var/log/Xorg.0.log file. That'll show you
> which driver was chosen.

I think you are right that it's some generic problem as it spreads across
several distinct hardware.
I have access to the /var/log/Xorg.7.log on the clients (via ssh). X uses the
specific drivers (OpenChrome for VIA-Chips, Intel, etc.).

But looking into the logs I found messages like this:
[ 347.509] (II) LoadModule: "glx"
[ 347.509] (II) Loading /usr/lib/xorg/modules/linux/libglx.so
[ 347.553] (II) Module glx: vendor="NVIDIA Corporation"
[ 347.553] (II) NVIDIA GLX Module 340.65 Tue Dec 2 09:02:32 PST 2014
[ 347.553] Loading extension GLX
[ 347.712] (EE) Failed to initialize GLX extension (Compatible NVIDIA X
driver not found)

This happened on an Intel board.
I never dived deeply into this graphics stuff (and never wanted to) and
therefore have to learn some basics like:
- OpenGL is a hardware- and platformindependant Interface for 3D-Graphics.
- GLX ist the layer between OpenGL and X-Windows
- DRI provides hardware acceleration for OpenGL.

On Debian, there are currently 3 GLX providers: mesa, nvidia, and fglrx. Only
mesa and nvidia have been installed in the image. I added the fglrx (for AMD
boards) and changed the glx-diversions from nvidia (hence the above error
message) to mesa with

update-alternatives --config glx

At least on an Intel-based client (openthinclient, I didn't yet test others),
the Xorg.7.log file now show this:

(LTSP)***@otc5:/~# grep -i glx /var/log/Xorg.7.log
[ 2187.600] (II) LoadModule: "glx"
[ 2187.601] (II) Loading /usr/lib/xorg/modules/extensions/libglx.so
[ 2187.602] (II) Module glx: vendor="X.Org Foundation"
[ 2187.602] (==) AIGLX enabled
[ 2187.602] Loading extension GLX
[ 2188.837] (II) AIGLX: enabled GLX_MESA_copy_sub_buffer
[ 2188.837] (II) AIGLX: enabled GLX_INTEL_swap_event
[ 2188.837] (II) AIGLX: enabled GLX_ARB_create_context
[ 2188.837] (II) AIGLX: enabled GLX_ARB_create_context_profile
[ 2188.837] (II) AIGLX: enabled GLX_EXT_create_context_es2_profile
[ 2188.837] (II) AIGLX: enabled GLX_SGI_swap_control and GLX_MESA_swap_control
[ 2188.837] (II) AIGLX: GLX_EXT_texture_from_pixmap backed by buffer objects
[ 2188.837] (II) AIGLX: Loaded and initialized i915
[ 2188.837] (II) GLX: Initialized DRI2 GL provider for screen 0

Seems to look better, but the windows behave as sluggish as before.

Karl Duc - Vormingplus regio Brugge schrieb am 13.02.2015 16:41:
> could it be a QT-problem?
> check https://en.opensuse.org/SDB:KIWI-LTSP_troubleshooting, at the bottom of the
> page:
> "In case of slow KDE applications on the thinclients (Qt apps like Dolphin,
> Kate, etc...), you have to modify the rendering of the Qt graphics system.
> One way to do this is create a file /etc/profile.local with following in it
>
> export QT_GRAPHICSSYSTEM=native
>
> Restart the thinclient and enjoy a fast ride"
> this solved a similar speed-problem for us, a year ago, using opensuse and kde

When I tried to implement your hint, I found that I've already done it in Nov.
2014. Seems you already wrote it last year on the list. :^)
But this also means, it doesn't help that much.

Again I'm running out of ideas.

Thanks for you help
Helmut

I'll add some system information and the Xorg.7.log:
==============================================================================
(LTSP)***@otc5:/~# lspci |grep -i vga
00:02.0 VGA compatible controller: Intel Corporation Mobile 945GSE Express Integrated Graphics Controller (rev 03)
==============================================================================
(LTSP)***@otc5:/~# lsof |grep glx
atievents 3276 root mem REG 0,17 99748 1097 /usr/lib/i386-linux-gnu/libxcb-glx.so.0.0.0
Xorg 11196 root mem REG 0,17 474292 1356 /usr/lib/xorg/modules/extensions/libglx.so
==============================================================================
(LTSP)***@otc5:/~# lsof |grep dri
Xorg 11196 root mem CHR 226,0 8755 /dev/dri/card0
Xorg 11196 root mem REG 0,17 3841816 1379 /usr/lib/i386-linux-gnu/dri/i915_dri.so
Xorg 11196 root mem REG 0,17 1340920 1335 /usr/lib/xorg/modules/drivers/intel_drv.so
Xorg 11196 root 13u CHR 226,0 0t0 8755 /dev/dri/card0
==============================================================================
(LTSP)***@otc5:/~# cat /var/log/Xorg.7.log
[ 2187.589]
X.Org X Server 1.14.3
Release Date: 2013-09-12
[ 2187.589] X Protocol Version 11, Revision 0
[ 2187.589] Build Operating System: Linux 3.2.0-4-amd64 i686 Debian
[ 2187.589] Current Operating System: Linux otc5 3.16.0-0.bpo.4-686-pae #1 SMP Debian 3.16.7-ckt2-1~bpo70+1 (2014-12-08) i686
[ 2187.589] Kernel command line: ro initrd=initrd.img init=/sbin/init-ltsp root=/dev/nbd0 ksdevice=bootif BOOT_IMAGE=vmlinuz BOOTIF=01-00-e0-c5-45-2b-75
[ 2187.590] Build Date: 08 October 2013 09:42:53PM
[ 2187.590] xorg-server 2:1.14.3-4 (Julien Cristau <***@debian.org>)
[ 2187.590] Current version of pixman: 0.30.2
[ 2187.590] Before reporting problems, check http://wiki.x.org
to make sure that you have the latest version.
[ 2187.590] Markers: (--) probed, (**) from config file, (==) default setting,
(++) from command line, (!!) notice, (II) informational,
(WW) warning, (EE) error, (NI) not implemented, (??) unknown.
[ 2187.590] (==) Log file: "/var/log/Xorg.7.log", Time: Mon Feb 16 11:47:01 2015
[ 2187.591] (++) Using config file: "/var/run/ltsp-xorg.conf"
[ 2187.591] (==) Using system config directory "/usr/share/X11/xorg.conf.d"
[ 2187.592] (==) No Layout section. Using the first Screen section.
[ 2187.592] (**) |-->Screen "Screen0" (0)
[ 2187.592] (**) | |-->Monitor "<default monitor>"
[ 2187.593] (==) No monitor specified for screen "Screen0".
Using a default monitor configuration.
[ 2187.593] (==) Automatically adding devices
[ 2187.593] (==) Automatically enabling devices
[ 2187.593] (==) Automatically adding GPU devices
[ 2187.593] (WW) The directory "/usr/share/fonts/X11/cyrillic" does not exist.
[ 2187.594] Entry deleted from font path.
[ 2187.594] (==) FontPath set to:
/usr/share/fonts/X11/misc,
/usr/share/fonts/X11/100dpi/:unscaled,
/usr/share/fonts/X11/75dpi/:unscaled,
/usr/share/fonts/X11/Type1,
/usr/share/fonts/X11/100dpi,
/usr/share/fonts/X11/75dpi,
built-ins
[ 2187.594] (==) ModulePath set to "/usr/lib/xorg/modules"
[ 2187.594] (II) The server relies on udev to provide the list of input devices.
If no devices become available, reconfigure udev or disable AutoAddDevices.
[ 2187.594] (II) Loader magic: 0xb77956a0
[ 2187.594] (II) Module ABI versions:
[ 2187.594] X.Org ANSI C Emulation: 0.4
[ 2187.594] X.Org Video Driver: 14.1
[ 2187.594] X.Org XInput driver : 19.1
[ 2187.594] X.Org Server Extension : 7.0
[ 2187.595] (II) xfree86: Adding drm device (/dev/dri/card0)
[ 2187.598] (--) PCI:*(0:0:2:0) 8086:27ae:8086:27ae rev 3, Mem @ 0xfde80000/524288, 0xd0000000/268435456, 0xfdf80000/262144, I/O @ 0x0000ff00/8
[ 2187.599] (--) PCI: (0:0:2:1) 8086:27a6:8086:27ae rev 3, Mem @ 0xfdf00000/524288
[ 2187.599] (WW) Open ACPI failed (/var/run/acpid.socket) (No such file or directory)
[ 2187.599] Initializing built-in extension Generic Event Extension
[ 2187.599] Initializing built-in extension SHAPE
[ 2187.599] Initializing built-in extension MIT-SHM
[ 2187.599] Initializing built-in extension XInputExtension
[ 2187.599] Initializing built-in extension XTEST
[ 2187.599] Initializing built-in extension BIG-REQUESTS
[ 2187.599] Initializing built-in extension SYNC
[ 2187.599] Initializing built-in extension XKEYBOARD
[ 2187.599] Initializing built-in extension XC-MISC
[ 2187.599] Initializing built-in extension SECURITY
[ 2187.599] Initializing built-in extension XINERAMA
[ 2187.599] Initializing built-in extension XFIXES
[ 2187.599] Initializing built-in extension RENDER
[ 2187.599] Initializing built-in extension RANDR
[ 2187.599] Initializing built-in extension COMPOSITE
[ 2187.599] Initializing built-in extension DAMAGE
[ 2187.599] Initializing built-in extension MIT-SCREEN-SAVER
[ 2187.599] Initializing built-in extension DOUBLE-BUFFER
[ 2187.599] Initializing built-in extension RECORD
[ 2187.599] Initializing built-in extension DPMS
[ 2187.599] Initializing built-in extension X-Resource
[ 2187.599] Initializing built-in extension XVideo
[ 2187.599] Initializing built-in extension XVideo-MotionCompensation
[ 2187.599] Initializing built-in extension SELinux
[ 2187.600] Initializing built-in extension XFree86-VidModeExtension
[ 2187.600] Initializing built-in extension XFree86-DGA
[ 2187.600] Initializing built-in extension XFree86-DRI
[ 2187.600] Initializing built-in extension DRI2
[ 2187.600] (II) LoadModule: "glx"
[ 2187.601] (II) Loading /usr/lib/xorg/modules/extensions/libglx.so
[ 2187.602] (II) Module glx: vendor="X.Org Foundation"
[ 2187.602] compiled for 1.14.3, module version = 1.0.0
[ 2187.602] ABI class: X.Org Server Extension, version 7.0
[ 2187.602] (==) AIGLX enabled
[ 2187.602] Loading extension GLX
[ 2187.602] (==) Matched intel as autoconfigured driver 0
[ 2187.602] (==) Matched intel as autoconfigured driver 1
[ 2187.602] (==) Matched vesa as autoconfigured driver 2
[ 2187.602] (==) Matched modesetting as autoconfigured driver 3
[ 2187.602] (==) Matched fbdev as autoconfigured driver 4
[ 2187.602] (==) Assigned the driver to the xf86ConfigLayout
[ 2187.602] (II) LoadModule: "intel"
[ 2187.602] (II) Loading /usr/lib/xorg/modules/drivers/intel_drv.so
[ 2187.603] (II) Module intel: vendor="X.Org Foundation"
[ 2187.603] compiled for 1.14.3, module version = 2.21.15
[ 2187.603] Module class: X.Org Video Driver
[ 2187.603] ABI class: X.Org Video Driver, version 14.1
[ 2187.603] (II) LoadModule: "vesa"
[ 2187.604] (II) Loading /usr/lib/xorg/modules/drivers/vesa_drv.so
[ 2187.604] (II) Module vesa: vendor="X.Org Foundation"
[ 2187.604] compiled for 1.14.3, module version = 2.3.3
[ 2187.604] Module class: X.Org Video Driver
[ 2187.604] ABI class: X.Org Video Driver, version 14.1
[ 2187.604] (II) LoadModule: "modesetting"
[ 2187.605] (II) Loading /usr/lib/xorg/modules/drivers/modesetting_drv.so
[ 2187.605] (II) Module modesetting: vendor="X.Org Foundation"
[ 2187.605] compiled for 1.14.3, module version = 0.8.0
[ 2187.605] Module class: X.Org Video Driver
[ 2187.605] ABI class: X.Org Video Driver, version 14.1
[ 2187.605] (II) LoadModule: "fbdev"
[ 2187.606] (II) Loading /usr/lib/xorg/modules/drivers/fbdev_drv.so
[ 2187.606] (II) Module fbdev: vendor="X.Org Foundation"
[ 2187.606] compiled for 1.14.3, module version = 0.4.4
[ 2187.606] Module class: X.Org Video Driver
[ 2187.606] ABI class: X.Org Video Driver, version 14.1
[ 2187.606] (II) intel: Driver for Intel(R) Integrated Graphics Chipsets:
i810, i810-dc100, i810e, i815, i830M, 845G, 854, 852GM/855GM, 865G,
915G, E7221 (i915), 915GM, 945G, 945GM, 945GME, Pineview GM,
Pineview G, 965G, G35, 965Q, 946GZ, 965GM, 965GME/GLE, G33, Q35, Q33,
GM45, 4 Series, G45/G43, Q45/Q43, G41, B43, HD Graphics,
HD Graphics 2000, HD Graphics 3000, HD Graphics 2500,
HD Graphics 4000, HD Graphics P4000, HD Graphics 4600,
HD Graphics 5000, HD Graphics P4600/P4700, Iris(TM) Graphics 5100,
HD Graphics 4400, HD Graphics 4200, Iris(TM) Pro Graphics 5200
[ 2187.607] (II) VESA: driver for VESA chipsets: vesa
[ 2187.607] (II) modesetting: Driver for Modesetting Kernel Drivers: kms
[ 2187.607] (II) FBDEV: driver for framebuffer: fbdev
[ 2187.607] (++) using VT number 7

[ 2187.608] (WW) Falling back to old probe method for vesa
[ 2187.608] (WW) Falling back to old probe method for modesetting
[ 2187.608] (WW) Falling back to old probe method for fbdev
[ 2187.608] (II) Loading sub module "fbdevhw"
[ 2187.608] (II) LoadModule: "fbdevhw"
[ 2187.609] (II) Loading /usr/lib/xorg/modules/libfbdevhw.so
[ 2187.609] (II) Module fbdevhw: vendor="X.Org Foundation"
[ 2187.609] compiled for 1.14.3, module version = 0.0.2
[ 2187.610] ABI class: X.Org Video Driver, version 14.1
[ 2187.610] (++) intel(0): Depth 16, (--) framebuffer bpp 16
[ 2187.610] (==) intel(0): RGB weight 565
[ 2187.610] (==) intel(0): Default visual is TrueColor
[ 2187.610] (--) intel(0): Integrated Graphics Chipset: Intel(R) 945GME
[ 2187.610] (**) intel(0): Relaxed fencing disabled
[ 2187.610] (**) intel(0): Wait on SwapBuffers? enabled
[ 2187.610] (**) intel(0): Triple buffering? enabled
[ 2187.610] (**) intel(0): Framebuffer tiled
[ 2187.610] (**) intel(0): Pixmaps tiled
[ 2187.610] (**) intel(0): 3D buffers tiled
[ 2187.611] (**) intel(0): SwapBuffers wait enabled
[ 2187.611] (==) intel(0): video overlay key set to 0x83e
[ 2187.648] (II) intel(0): Output VGA1 has no monitor section
[ 2187.724] (II) intel(0): Output DVI1 has no monitor section
[ 2188.118] (II) intel(0): Output TV1 has no monitor section
[ 2188.156] (II) intel(0): EDID for output VGA1
[ 2188.235] (II) intel(0): EDID for output DVI1
[ 2188.235] (II) intel(0): Manufacturer: DEL Model: 405e Serial#: 808858700
[ 2188.235] (II) intel(0): Year: 2010 Week: 26
[ 2188.235] (II) intel(0): EDID Version: 1.3
[ 2188.235] (II) intel(0): Digital Display Input
[ 2188.235] (II) intel(0): Max Image Size [cm]: horiz.: 48 vert.: 27
[ 2188.235] (II) intel(0): Gamma: 2.20
[ 2188.235] (II) intel(0): DPMS capabilities: StandBy Suspend Off
[ 2188.235] (II) intel(0): Supported color encodings: RGB 4:4:4 YCrCb 4:4:4
[ 2188.235] (II) intel(0): Default color space is primary color space
[ 2188.235] (II) intel(0): First detailed timing is preferred mode
[ 2188.236] (II) intel(0): redX: 0.640 redY: 0.330 greenX: 0.300 greenY: 0.600
[ 2188.236] (II) intel(0): blueX: 0.150 blueY: 0.060 whiteX: 0.313 whiteY: 0.329
[ 2188.236] (II) intel(0): Supported established timings:
[ 2188.236] (II) intel(0): ***@70Hz
[ 2188.236] (II) intel(0): ***@60Hz
[ 2188.236] (II) intel(0): ***@75Hz
[ 2188.236] (II) intel(0): ***@60Hz
[ 2188.236] (II) intel(0): ***@75Hz
[ 2188.236] (II) intel(0): ***@60Hz
[ 2188.236] (II) intel(0): ***@75Hz
[ 2188.236] (II) intel(0): ***@75Hz
[ 2188.236] (II) intel(0): Manufacturer's mask: 0
[ 2188.236] (II) intel(0): Supported standard timings:
[ 2188.236] (II) intel(0): #0: hsize: 1152 vsize 864 refresh: 75 vid: 20337
[ 2188.236] (II) intel(0): #1: hsize: 1280 vsize 1024 refresh: 60 vid: 32897
[ 2188.236] (II) intel(0): #2: hsize: 1920 vsize 1080 refresh: 60 vid: 49361
[ 2188.236] (II) intel(0): Supported detailed timing:
[ 2188.236] (II) intel(0): clock: 148.5 MHz Image Size: 475 x 267 mm
[ 2188.236] (II) intel(0): h_active: 1920 h_sync: 2008 h_sync_end 2052 h_blank_end 2200 h_border: 0
[ 2188.236] (II) intel(0): v_active: 1080 v_sync: 1084 v_sync_end 1089 v_blanking: 1125 v_border: 0
[ 2188.236] (II) intel(0): Serial No: D1TPK06U064L
[ 2188.236] (II) intel(0): Monitor name: DELL U2211H
[ 2188.236] (II) intel(0): Ranges: V min: 56 V max: 76 Hz, H min: 30 H max: 83 kHz, PixClock max 175 MHz
[ 2188.236] (II) intel(0): EDID (in hex):
[ 2188.236] (II) intel(0): 00ffffffffffff0010ac5e404c343630
[ 2188.236] (II) intel(0): 1a14010380301b78eeee95a3544c9926
[ 2188.236] (II) intel(0): 0f5054a54b00714f8180d1c001010101
[ 2188.236] (II) intel(0): 010101010101023a801871382d40582c
[ 2188.236] (II) intel(0): 4500db0b1100001e000000ff00443154
[ 2188.237] (II) intel(0): 504b3036553036344c0a000000fc0044
[ 2188.237] (II) intel(0): 454c4c205532323131480a20000000fd
[ 2188.237] (II) intel(0): 00384c1e5311000a2020202020200072
[ 2188.237] (II) intel(0): Printing probed modes for output DVI1
[ 2188.237] (II) intel(0): Modeline "1920x1080"x60.0 148.50 1920 2008 2052 2200 1080 1084 1089 1125 +hsync +vsync (67.5 kHz eP)
[ 2188.237] (II) intel(0): Modeline "1280x1024"x75.0 135.00 1280 1296 1440 1688 1024 1025 1028 1066 +hsync +vsync (80.0 kHz e)
[ 2188.237] (II) intel(0): Modeline "1280x1024"x60.0 108.00 1280 1328 1440 1688 1024 1025 1028 1066 +hsync +vsync (64.0 kHz e)
[ 2188.237] (II) intel(0): Modeline "1152x864"x75.0 108.00 1152 1216 1344 1600 864 865 868 900 +hsync +vsync (67.5 kHz e)
[ 2188.237] (II) intel(0): Modeline "1024x768"x75.1 78.80 1024 1040 1136 1312 768 769 772 800 +hsync +vsync (60.1 kHz e)
[ 2188.237] (II) intel(0): Modeline "1024x768"x60.0 65.00 1024 1048 1184 1344 768 771 777 806 -hsync -vsync (48.4 kHz e)
[ 2188.237] (II) intel(0): Modeline "800x600"x75.0 49.50 800 816 896 1056 600 601 604 625 +hsync +vsync (46.9 kHz e)
[ 2188.237] (II) intel(0): Modeline "800x600"x60.3 40.00 800 840 968 1056 600 601 605 628 +hsync +vsync (37.9 kHz e)
[ 2188.237] (II) intel(0): Modeline "640x480"x75.0 31.50 640 656 720 840 480 481 484 500 -hsync -vsync (37.5 kHz e)
[ 2188.237] (II) intel(0): Modeline "640x480"x60.0 25.20 640 656 752 800 480 490 492 525 -hsync -vsync (31.5 kHz e)
[ 2188.237] (II) intel(0): Modeline "720x400"x70.1 28.32 720 738 846 900 400 412 414 449 -hsync +vsync (31.5 kHz e)
[ 2188.634] (II) intel(0): EDID for output TV1
[ 2188.634] (II) intel(0): Output VGA1 disconnected
[ 2188.634] (II) intel(0): Output DVI1 connected
[ 2188.634] (II) intel(0): Output TV1 disconnected
[ 2188.634] (II) intel(0): Using exact sizes for initial modes
[ 2188.634] (II) intel(0): Output DVI1 using initial mode 1920x1080
[ 2188.634] (II) intel(0): Using default gamma of (1.0, 1.0, 1.0) unless otherwise stated.
[ 2188.634] (II) intel(0): Kernel page flipping support detected, enabling
[ 2188.634] (==) intel(0): DPI set to (96, 96)
[ 2188.634] (II) Loading sub module "fb"
[ 2188.634] (II) LoadModule: "fb"
[ 2188.635] (II) Loading /usr/lib/xorg/modules/libfb.so
[ 2188.636] (II) Module fb: vendor="X.Org Foundation"
[ 2188.636] compiled for 1.14.3, module version = 1.0.0
[ 2188.636] ABI class: X.Org ANSI C Emulation, version 0.4
[ 2188.636] (II) Loading sub module "dri2"
[ 2188.636] (II) LoadModule: "dri2"
[ 2188.636] (II) Module "dri2" already built-in
[ 2188.636] (II) UnloadModule: "vesa"
[ 2188.636] (II) Unloading vesa
[ 2188.636] (II) UnloadModule: "modesetting"
[ 2188.636] (II) Unloading modesetting
[ 2188.636] (II) UnloadModule: "fbdev"
[ 2188.636] (II) Unloading fbdev
[ 2188.636] (II) UnloadSubModule: "fbdevhw"
[ 2188.636] (II) Unloading fbdevhw
[ 2188.637] (II) intel(0): [DRI2] Setup complete
[ 2188.637] (II) intel(0): [DRI2] DRI driver: i915
[ 2188.637] (II) intel(0): Allocated new frame buffer 1920x1080 stride 4096, tiled
[ 2188.637] (II) UXA(0): Driver registered support for the following operations:
[ 2188.637] (II) solid
[ 2188.637] (II) copy
[ 2188.637] (II) composite (RENDER acceleration)
[ 2188.637] (II) put_image
[ 2188.637] (II) get_image
[ 2188.637] (==) intel(0): Backing store disabled
[ 2188.637] (==) intel(0): Silken mouse enabled
[ 2188.637] (II) intel(0): Initializing HW Cursor
[ 2188.637] (II) intel(0): RandR 1.2 enabled, ignore the following RandR disabled message.
[ 2188.638] (==) intel(0): DPMS enabled
[ 2188.638] (==) intel(0): Intel XvMC decoder disabled
[ 2188.638] (II) intel(0): Set up textured video
[ 2188.638] (II) intel(0): Set up overlay video
[ 2188.638] (II) intel(0): direct rendering: DRI2 Enabled
[ 2188.638] (==) intel(0): hotplug detection: "enabled"
[ 2188.802] (--) RandR disabled
[ 2188.826] (II) SELinux: Disabled on system
[ 2188.837] (II) AIGLX: enabled GLX_MESA_copy_sub_buffer
[ 2188.837] (II) AIGLX: enabled GLX_INTEL_swap_event
[ 2188.837] (II) AIGLX: enabled GLX_ARB_create_context
[ 2188.837] (II) AIGLX: enabled GLX_ARB_create_context_profile
[ 2188.837] (II) AIGLX: enabled GLX_EXT_create_context_es2_profile
[ 2188.837] (II) AIGLX: enabled GLX_SGI_swap_control and GLX_MESA_swap_control
[ 2188.837] (II) AIGLX: GLX_EXT_texture_from_pixmap backed by buffer objects
[ 2188.837] (II) AIGLX: Loaded and initialized i915
[ 2188.837] (II) GLX: Initialized DRI2 GL provider for screen 0
[ 2188.839] (II) intel(0): Setting screen physical size to 507 x 285
[ 2188.985] (II) config/udev: Adding input device Power Button (/dev/input/event3)
[ 2188.985] (**) Power Button: Applying InputClass "evdev keyboard catchall"
[ 2188.986] (II) LoadModule: "evdev"
[ 2188.986] (II) Loading /usr/lib/xorg/modules/input/evdev_drv.so
[ 2188.987] (II) Module evdev: vendor="X.Org Foundation"
[ 2188.987] compiled for 1.14.3, module version = 2.8.2
[ 2188.987] Module class: X.Org XInput Driver
[ 2188.987] ABI class: X.Org XInput driver, version 19.1
[ 2188.987] (II) Using input driver 'evdev' for 'Power Button'
[ 2188.987] (**) Power Button: always reports core events
[ 2188.987] (**) evdev: Power Button: Device: "/dev/input/event3"
[ 2188.987] (--) evdev: Power Button: Vendor 0 Product 0x1
[ 2188.987] (--) evdev: Power Button: Found keys
[ 2188.987] (II) evdev: Power Button: Configuring as keyboard
[ 2188.987] (**) Option "config_info" "udev:/sys/devices/LNXSYSTM:00/LNXPWRBN:00/input/input4/event3"
[ 2188.987] (II) XINPUT: Adding extended input device "Power Button" (type: KEYBOARD, id 6)
[ 2188.987] (**) Option "xkb_rules" "evdev"
[ 2188.987] (**) Option "xkb_model" "pc105"
[ 2188.988] (**) Option "xkb_layout" "de"
[ 2188.988] (**) Option "xkb_variant" "nodeadkeys"
[ 2189.052] (II) config/udev: Adding input device Power Button (/dev/input/event2)
[ 2189.052] (**) Power Button: Applying InputClass "evdev keyboard catchall"
[ 2189.052] (II) Using input driver 'evdev' for 'Power Button'
[ 2189.052] (**) Power Button: always reports core events
[ 2189.052] (**) evdev: Power Button: Device: "/dev/input/event2"
[ 2189.052] (--) evdev: Power Button: Vendor 0 Product 0x1
[ 2189.052] (--) evdev: Power Button: Found keys
[ 2189.052] (II) evdev: Power Button: Configuring as keyboard
[ 2189.052] (**) Option "config_info" "udev:/sys/devices/LNXSYSTM:00/LNXSYBUS:00/PNP0C0C:00/input/input3/event2"
[ 2189.052] (II) XINPUT: Adding extended input device "Power Button" (type: KEYBOARD, id 7)
[ 2189.052] (**) Option "xkb_rules" "evdev"
[ 2189.052] (**) Option "xkb_model" "pc105"
[ 2189.052] (**) Option "xkb_layout" "de"
[ 2189.052] (**) Option "xkb_variant" "nodeadkeys"
[ 2189.053] (II) config/udev: Adding drm device (/dev/dri/card0)
[ 2189.055] (II) config/udev: Adding input device HDA Intel Line Out Surround (/dev/input/event9)
[ 2189.055] (II) No input driver specified, ignoring this device.
[ 2189.055] (II) This device may have been added with another device file.
[ 2189.055] (II) config/udev: Adding input device HDA Intel Line Out CLFE (/dev/input/event10)
[ 2189.055] (II) No input driver specified, ignoring this device.
[ 2189.055] (II) This device may have been added with another device file.
[ 2189.056] (II) config/udev: Adding input device HDA Intel Line Out Side (/dev/input/event11)
[ 2189.056] (II) No input driver specified, ignoring this device.
[ 2189.056] (II) This device may have been added with another device file.
[ 2189.057] (II) config/udev: Adding input device HDA Intel Front Headphone (/dev/input/event12)
[ 2189.057] (II) No input driver specified, ignoring this device.
[ 2189.057] (II) This device may have been added with another device file.
[ 2189.058] (II) config/udev: Adding input device HDA Intel Rear Mic (/dev/input/event5)
[ 2189.058] (II) No input driver specified, ignoring this device.
[ 2189.058] (II) This device may have been added with another device file.
[ 2189.059] (II) config/udev: Adding input device HDA Intel Front Mic (/dev/input/event6)
[ 2189.059] (II) No input driver specified, ignoring this device.
[ 2189.059] (II) This device may have been added with another device file.
[ 2189.060] (II) config/udev: Adding input device HDA Intel Line (/dev/input/event7)
[ 2189.060] (II) No input driver specified, ignoring this device.
[ 2189.060] (II) This device may have been added with another device file.
[ 2189.061] (II) config/udev: Adding input device HDA Intel Line Out Front (/dev/input/event8)
[ 2189.061] (II) No input driver specified, ignoring this device.
[ 2189.061] (II) This device may have been added with another device file.
[ 2189.062] (II) config/udev: Adding input device CHICONY HP Basic USB Keyboard (/dev/input/event0)
[ 2189.062] (**) CHICONY HP Basic USB Keyboard: Applying InputClass "evdev keyboard catchall"
[ 2189.062] (II) Using input driver 'evdev' for 'CHICONY HP Basic USB Keyboard'
[ 2189.062] (**) CHICONY HP Basic USB Keyboard: always reports core events
[ 2189.062] (**) evdev: CHICONY HP Basic USB Keyboard: Device: "/dev/input/event0"
[ 2189.062] (--) evdev: CHICONY HP Basic USB Keyboard: Vendor 0x3f0 Product 0x24
[ 2189.062] (--) evdev: CHICONY HP Basic USB Keyboard: Found keys
[ 2189.062] (II) evdev: CHICONY HP Basic USB Keyboard: Configuring as keyboard
[ 2189.062] (**) Option "config_info" "udev:/sys/devices/pci0000:00/0000:00:1d.3/usb5/5-1/5-1:1.0/0003:03F0:0024.0001/input/input1/event0"
[ 2189.062] (II) XINPUT: Adding extended input device "CHICONY HP Basic USB Keyboard" (type: KEYBOARD, id 8)
[ 2189.062] (**) Option "xkb_rules" "evdev"
[ 2189.062] (**) Option "xkb_model" "pc105"
[ 2189.062] (**) Option "xkb_layout" "de"
[ 2189.062] (**) Option "xkb_variant" "nodeadkeys"
[ 2189.064] (II) config/udev: Adding input device Logitech USB-PS/2 Optical Mouse (/dev/input/event1)
[ 2189.064] (**) Logitech USB-PS/2 Optical Mouse: Applying InputClass "evdev pointer catchall"
[ 2189.064] (II) Using input driver 'evdev' for 'Logitech USB-PS/2 Optical Mouse'
[ 2189.064] (**) Logitech USB-PS/2 Optical Mouse: always reports core events
[ 2189.064] (**) evdev: Logitech USB-PS/2 Optical Mouse: Device: "/dev/input/event1"
[ 2189.064] (--) evdev: Logitech USB-PS/2 Optical Mouse: Vendor 0x46d Product 0xc03d
[ 2189.065] (--) evdev: Logitech USB-PS/2 Optical Mouse: Found 3 mouse buttons
[ 2189.065] (--) evdev: Logitech USB-PS/2 Optical Mouse: Found scroll wheel(s)
[ 2189.065] (--) evdev: Logitech USB-PS/2 Optical Mouse: Found relative axes
[ 2189.065] (--) evdev: Logitech USB-PS/2 Optical Mouse: Found x and y relative axes
[ 2189.065] (II) evdev: Logitech USB-PS/2 Optical Mouse: Configuring as mouse
[ 2189.065] (II) evdev: Logitech USB-PS/2 Optical Mouse: Adding scrollwheel support
[ 2189.065] (**) evdev: Logitech USB-PS/2 Optical Mouse: YAxisMapping: buttons 4 and 5
[ 2189.065] (**) evdev: Logitech USB-PS/2 Optical Mouse: EmulateWheelButton: 4, EmulateWheelInertia: 10, EmulateWheelTimeout: 200
[ 2189.065] (**) Option "config_info" "udev:/sys/devices/pci0000:00/0000:00:1d.3/usb5/5-2/5-2:1.0/0003:046D:C03D.0002/input/input2/event1"
[ 2189.065] (II) XINPUT: Adding extended input device "Logitech USB-PS/2 Optical Mouse" (type: MOUSE, id 9)
[ 2189.065] (II) evdev: Logitech USB-PS/2 Optical Mouse: initialized for relative axes.
[ 2189.065] (**) Logitech USB-PS/2 Optical Mouse: (accel) keeping acceleration scheme 1
[ 2189.066] (**) Logitech USB-PS/2 Optical Mouse: (accel) acceleration profile 0
[ 2189.066] (**) Logitech USB-PS/2 Optical Mouse: (accel) acceleration factor: 2.000
[ 2189.066] (**) Logitech USB-PS/2 Optical Mouse: (accel) acceleration threshold: 4
[ 2189.066] (II) config/udev: Adding input device Logitech USB-PS/2 Optical Mouse (/dev/input/mouse0)
[ 2189.067] (II) No input driver specified, ignoring this device.
[ 2189.067] (II) This device may have been added with another device file.
[ 2189.067] (II) config/udev: Adding input device PC Speaker (/dev/input/event4)
[ 2189.067] (II) No input driver specified, ignoring this device.
[ 2189.067] (II) This device may have been added with another device file.
[ 2189.069] (II) config/udev: Adding input device ACPI Virtual Keyboard Device (/dev/input/event13)
[ 2189.069] (**) ACPI Virtual Keyboard Device: Applying InputClass "evdev keyboard catchall"
[ 2189.069] (II) Using input driver 'evdev' for 'ACPI Virtual Keyboard Device'
[ 2189.069] (**) ACPI Virtual Keyboard Device: always reports core events
[ 2189.069] (**) evdev: ACPI Virtual Keyboard Device: Device: "/dev/input/event13"
[ 2189.069] (--) evdev: ACPI Virtual Keyboard Device: Vendor 0 Product 0
[ 2189.069] (--) evdev: ACPI Virtual Keyboard Device: Found keys
[ 2189.069] (II) evdev: ACPI Virtual Keyboard Device: Configuring as keyboard
[ 2189.070] (**) Option "config_info" "udev:/sys/devices/virtual/input/input14/event13"
[ 2189.070] (II) XINPUT: Adding extended input device "ACPI Virtual Keyboard Device" (type: KEYBOARD, id 10)
[ 2189.070] (**) Option "xkb_rules" "evdev"
[ 2189.070] (**) Option "xkb_model" "pc105"
[ 2189.070] (**) Option "xkb_layout" "de"
[ 2189.070] (**) Option "xkb_variant" "nodeadkeys"
[ 2189.306] (II) intel(0): EDID vendor "DEL", prod id 16478
[ 2189.307] (II) intel(0): Using EDID range info for horizontal sync
[ 2189.307] (II) intel(0): Using EDID range info for vertical refresh
[ 2189.307] (II) intel(0): Printing DDC gathered Modelines:
[ 2189.307] (II) intel(0): Modeline "1920x1080"x0.0 148.50 1920 2008 2052 2200 1080 1084 1089 1125 +hsync +vsync (67.5 kHz eP)
[ 2189.307] (II) intel(0): Modeline "800x600"x0.0 40.00 800 840 968 1056 600 601 605 628 +hsync +vsync (37.9 kHz e)
[ 2189.307] (II) intel(0): Modeline "640x480"x0.0 31.50 640 656 720 840 480 481 484 500 -hsync -vsync (37.5 kHz e)
[ 2189.307] (II) intel(0): Modeline "640x480"x0.0 25.18 640 656 752 800 480 490 492 525 -hsync -vsync (31.5 kHz e)
[ 2189.307] (II) intel(0): Modeline "720x400"x0.0 28.32 720 738 846 900 400 412 414 449 -hsync +vsync (31.5 kHz e)
[ 2189.307] (II) intel(0): Modeline "1280x1024"x0.0 135.00 1280 1296 1440 1688 1024 1025 1028 1066 +hsync +vsync (80.0 kHz e)
[ 2189.307] (II) intel(0): Modeline "1024x768"x0.0 78.75 1024 1040 1136 1312 768 769 772 800 +hsync +vsync (60.0 kHz e)
[ 2189.307] (II) intel(0): Modeline "1024x768"x0.0 65.00 1024 1048 1184 1344 768 771 777 806 -hsync -vsync (48.4 kHz e)
[ 2189.307] (II) intel(0): Modeline "800x600"x0.0 49.50 800 816 896 1056 600 601 604 625 +hsync +vsync (46.9 kHz e)
[ 2189.307] (II) intel(0): Modeline "1152x864"x0.0 108.00 1152 1216 1344 1600 864 865 868 900 +hsync +vsync (67.5 kHz e)
[ 2189.307] (II) intel(0): Modeline "1280x1024"x0.0 108.00 1280 1328 1440 1688 1024 1025 1028 1066 +hsync +vsync (64.0 kHz e)
[ 2189.307] (II) intel(0): Modeline "1920x1080"x60.0 172.80 1920 2040 2248 2576 1080 1081 1084 1118 -hsync +vsync (67.1 kHz e)
[ 2189.826] (II) intel(0): EDID vendor "DEL", prod id 16478
[ 2189.827] (II) intel(0): Using hsync ranges from config file
[ 2189.827] (II) intel(0): Using vrefresh ranges from config file
[ 2189.827] (II) intel(0): Printing DDC gathered Modelines:
[ 2189.827] (II) intel(0): Modeline "1920x1080"x0.0 148.50 1920 2008 2052 2200 1080 1084 1089 1125 +hsync +vsync (67.5 kHz eP)
[ 2189.827] (II) intel(0): Modeline "800x600"x0.0 40.00 800 840 968 1056 600 601 605 628 +hsync +vsync (37.9 kHz e)
[ 2189.827] (II) intel(0): Modeline "640x480"x0.0 31.50 640 656 720 840 480 481 484 500 -hsync -vsync (37.5 kHz e)
[ 2189.827] (II) intel(0): Modeline "640x480"x0.0 25.18 640 656 752 800 480 490 492 525 -hsync -vsync (31.5 kHz e)
[ 2189.827] (II) intel(0): Modeline "720x400"x0.0 28.32 720 738 846 900 400 412 414 449 -hsync +vsync (31.5 kHz e)
[ 2189.827] (II) intel(0): Modeline "1280x1024"x0.0 135.00 1280 1296 1440 1688 1024 1025 1028 1066 +hsync +vsync (80.0 kHz e)
[ 2189.827] (II) intel(0): Modeline "1024x768"x0.0 78.75 1024 1040 1136 1312 768 769 772 800 +hsync +vsync (60.0 kHz e)
[ 2189.827] (II) intel(0): Modeline "1024x768"x0.0 65.00 1024 1048 1184 1344 768 771 777 806 -hsync -vsync (48.4 kHz e)
[ 2189.827] (II) intel(0): Modeline "800x600"x0.0 49.50 800 816 896 1056 600 601 604 625 +hsync +vsync (46.9 kHz e)
[ 2189.827] (II) intel(0): Modeline "1152x864"x0.0 108.00 1152 1216 1344 1600 864 865 868 900 +hsync +vsync (67.5 kHz e)
[ 2189.827] (II) intel(0): Modeline "1280x1024"x0.0 108.00 1280 1328 1440 1688 1024 1025 1028 1066 +hsync +vsync (64.0 kHz e)
[ 2189.827] (II) intel(0): Modeline "1920x1080"x60.0 172.80 1920 2040 2248 2576 1080 1081 1084 1118 -hsync +vsync (67.1 kHz e)
[ 2225.292] (II) intel(0): EDID vendor "DEL", prod id 16478
[ 2225.293] (II) intel(0): Using hsync ranges from config file
[ 2225.293] (II) intel(0): Using vrefresh ranges from config file
[ 2225.293] (II) intel(0): Printing DDC gathered Modelines:
[ 2225.293] (II) intel(0): Modeline "1920x1080"x0.0 148.50 1920 2008 2052 2200 1080 1084 1089 1125 +hsync +vsync (67.5 kHz eP)
[ 2225.293] (II) intel(0): Modeline "800x600"x0.0 40.00 800 840 968 1056 600 601 605 628 +hsync +vsync (37.9 kHz e)
[ 2225.293] (II) intel(0): Modeline "640x480"x0.0 31.50 640 656 720 840 480 481 484 500 -hsync -vsync (37.5 kHz e)
[ 2225.293] (II) intel(0): Modeline "640x480"x0.0 25.18 640 656 752 800 480 490 492 525 -hsync -vsync (31.5 kHz e)
[ 2225.293] (II) intel(0): Modeline "720x400"x0.0 28.32 720 738 846 900 400 412 414 449 -hsync +vsync (31.5 kHz e)
[ 2225.293] (II) intel(0): Modeline "1280x1024"x0.0 135.00 1280 1296 1440 1688 1024 1025 1028 1066 +hsync +vsync (80.0 kHz e)
[ 2225.293] (II) intel(0): Modeline "1024x768"x0.0 78.75 1024 1040 1136 1312 768 769 772 800 +hsync +vsync (60.0 kHz e)
[ 2225.293] (II) intel(0): Modeline "1024x768"x0.0 65.00 1024 1048 1184 1344 768 771 777 806 -hsync -vsync (48.4 kHz e)
[ 2225.293] (II) intel(0): Modeline "800x600"x0.0 49.50 800 816 896 1056 600 601 604 625 +hsync +vsync (46.9 kHz e)
[ 2225.293] (II) intel(0): Modeline "1152x864"x0.0 108.00 1152 1216 1344 1600 864 865 868 900 +hsync +vsync (67.5 kHz e)
[ 2225.293] (II) intel(0): Modeline "1280x1024"x0.0 108.00 1280 1328 1440 1688 1024 1025 1028 1066 +hsync +vsync (64.0 kHz e)
[ 2225.293] (II) intel(0): Modeline "1920x1080"x60.0 172.80 1920 2040 2248 2576 1080 1081 1084 1118 -hsync +vsync (67.1 kHz e)
==============================================================================

--
-------------------------------------------------------------------------
Helmut Lichtenberg <***@fli.bund.de> Tel.: [phone]
Institut für Nutztiergenetik (FLI) 31535 Neustadt Germany
-------------------------------------------------------------------------

------------------------------------------------------------------------------
Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
from Actuate! Instantly Supercharge Your Business Reports and Dashboards
with Interactivity, Sharing, Native Excel Exports, App Integration & more
Get technology previously reserved for billion-dollar corporations, FREE
http://pubads.g.doubleclick.net/gampad/clk?id=190641631&iu=/4140/ostg.clktrk
_____________________________________________________________________
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help, try #ltsp channel on irc.freenode.net

[Michal Rybárik]

Hi Helmut,

we have same performance problems as you. Everything was working well on
older distributions for long years (we used OpenSuse 11.3 past years).
But when installing newer distributions (Kubuntu 14.04, CentOS 6),
graphics is very very slow, on the same hardware and same network as
before. Very simple graphics changes now uses 100% CPU on 1GHz thin
client, and load goes to 30.00 and more in a short while. Thin client
with dualcore i5 CPU is now slower, than 350 MHz PII was before.

There are some simple graphics operations that kills performance very
drastically - for example in OpenOffice Calc if user selects more cells
and press Ctrl-C (copy), moving dashes are displayed around selection,
and these moving dashes makes thin client almost unresponsive (because
of very high CPU usage).

Setting QT_GRAPHICSSYSTEM=native partially fixes problem for QT
applications. Fullscreen Konsole window is redrawn in ~5 seconds without
this setting, and in <1 second with it. But we are unable to fix other
operations (for example OpenOffice selection mentioned above).

I can say that this appies to
- all graphics chipsets that we have in thinclients (SiS, Radeon 280,
intel HD 4000), no difference between chipsets
- both LDM and XDMCP method of communication
- both Kubuntu 14.04 and Centos 6

We are trying to find solution for more than 4 months, but still no
luck. All suggestions are welcome...

--
Michal Rybarik



Dňa 02/16/2015 o 01:27 PM Helmut Lichtenberg napísal(a):
> James, Karl,
> thanks for your replies.
>
> James McQuillan schrieb am 13.02.2015 14:53:
>> My first guess would be the graphics drivers that have been automatically
>> chosen for those thin clients.
>>
>> If it's using the generic driver instead of the specific driver that's best
>> for each chipset, you'll see graphic performance like you describe.
>>
>> Perhaps you can get a shell running on the clients and look at the
>> /var/log/Xorg.0.log file. That'll show you
>> which driver was chosen.
> I think you are right that it's some generic problem as it spreads across
> several distinct hardware.
> I have access to the /var/log/Xorg.7.log on the clients (via ssh). X uses the
> specific drivers (OpenChrome for VIA-Chips, Intel, etc.).
>
> But looking into the logs I found messages like this:
> [ 347.509] (II) LoadModule: "glx"
> [ 347.509] (II) Loading /usr/lib/xorg/modules/linux/libglx.so
> [ 347.553] (II) Module glx: vendor="NVIDIA Corporation"
> [ 347.553] (II) NVIDIA GLX Module 340.65 Tue Dec 2 09:02:32 PST 2014
> [ 347.553] Loading extension GLX
> [ 347.712] (EE) Failed to initialize GLX extension (Compatible NVIDIA X
> driver not found)
>
> This happened on an Intel board.
> I never dived deeply into this graphics stuff (and never wanted to) and
> therefore have to learn some basics like:
> - OpenGL is a hardware- and platformindependant Interface for 3D-Graphics.
> - GLX ist the layer between OpenGL and X-Windows
> - DRI provides hardware acceleration for OpenGL.
>
> On Debian, there are currently 3 GLX providers: mesa, nvidia, and fglrx. Only
> mesa and nvidia have been installed in the image. I added the fglrx (for AMD
> boards) and changed the glx-diversions from nvidia (hence the above error
> message) to mesa with
>
> update-alternatives --config glx
>
> At least on an Intel-based client (openthinclient, I didn't yet test others),
> the Xorg.7.log file now show this:
>
> (LTSP)***@otc5:/~# grep -i glx /var/log/Xorg.7.log
> [ 2187.600] (II) LoadModule: "glx"
> [ 2187.601] (II) Loading /usr/lib/xorg/modules/extensions/libglx.so
> [ 2187.602] (II) Module glx: vendor="X.Org Foundation"
> [ 2187.602] (==) AIGLX enabled
> [ 2187.602] Loading extension GLX
> [ 2188.837] (II) AIGLX: enabled GLX_MESA_copy_sub_buffer
> [ 2188.837] (II) AIGLX: enabled GLX_INTEL_swap_event
> [ 2188.837] (II) AIGLX: enabled GLX_ARB_create_context
> [ 2188.837] (II) AIGLX: enabled GLX_ARB_create_context_profile
> [ 2188.837] (II) AIGLX: enabled GLX_EXT_create_context_es2_profile
> [ 2188.837] (II) AIGLX: enabled GLX_SGI_swap_control and GLX_MESA_swap_control
> [ 2188.837] (II) AIGLX: GLX_EXT_texture_from_pixmap backed by buffer objects
> [ 2188.837] (II) AIGLX: Loaded and initialized i915
> [ 2188.837] (II) GLX: Initialized DRI2 GL provider for screen 0
>
> Seems to look better, but the windows behave as sluggish as before.
>
> Karl Duc - Vormingplus regio Brugge schrieb am 13.02.2015 16:41:
>> could it be a QT-problem?
>> check https://en.opensuse.org/SDB:KIWI-LTSP_troubleshooting, at the bottom of the
>> page:
>> "In case of slow KDE applications on the thinclients (Qt apps like Dolphin,
>> Kate, etc...), you have to modify the rendering of the Qt graphics system.
>> One way to do this is create a file /etc/profile.local with following in it
>>
>> export QT_GRAPHICSSYSTEM=native
>>
>> Restart the thinclient and enjoy a fast ride"
>> this solved a similar speed-problem for us, a year ago, using opensuse and kde
> When I tried to implement your hint, I found that I've already done it in Nov.
> 2014. Seems you already wrote it last year on the list. :^)
> But this also means, it doesn't help that much.
>
> Again I'm running out of ideas.
>
> Thanks for you help
> Helmut
>


------------------------------------------------------------------------------
Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
from Actuate! Instantly Supercharge Your Business Reports and Dashboards
with Interactivity, Sharing, Native Excel Exports, App Integration & more
Get technology previously reserved for billion-dollar corporations, FREE
http://pubads.g.doubleclick.net/gampad/clk?id=190641631&iu=/4140/ostg.clktrk
_____________________________________________________________________
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help, try #ltsp channel on irc.freeno

[Alkis Georgopoulos]

Maybe you're using a compositing window manager?
Everything is still ok here with Ubuntu 14.04 and the gnome-flashback
session (metacity window manager without effects).

If you think that graphics in general are slower, and it's not a
window-manager/compositing issue, then try x11perf, gtkperf or something
similar to compare the results with the older installation.


On 17/02/2015 03:54 μμ, Michal Rybárik wrote:
> Hi Helmut,
>
> we have same performance problems as you. Everything was working well on
> older distributions for long years (we used OpenSuse 11.3 past years).
> But when installing newer distributions (Kubuntu 14.04, CentOS 6),
> graphics is very very slow, on the same hardware and same network as
> before. Very simple graphics changes now uses 100% CPU on 1GHz thin
> client, and load goes to 30.00 and more in a short while. Thin client
> with dualcore i5 CPU is now slower, than 350 MHz PII was before.
>
> There are some simple graphics operations that kills performance very
> drastically - for example in OpenOffice Calc if user selects more cells
> and press Ctrl-C (copy), moving dashes are displayed around selection,
> and these moving dashes makes thin client almost unresponsive (because
> of very high CPU usage).
>
> Setting QT_GRAPHICSSYSTEM=native partially fixes problem for QT
> applications. Fullscreen Konsole window is redrawn in ~5 seconds without
> this setting, and in <1 second with it. But we are unable to fix other
> operations (for example OpenOffice selection mentioned above).
>
> I can say that this appies to
> - all graphics chipsets that we have in thinclients (SiS, Radeon 280,
> intel HD 4000), no difference between chipsets
> - both LDM and XDMCP method of communication
> - both Kubuntu 14.04 and Centos 6
>
> We are trying to find solution for more than 4 months, but still no
> luck. All suggestions are welcome...
>
> --
> Michal Rybarik
>

------------------------------------------------------------------------------
Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
from Actuate! Instantly Supercharge Your Business Reports and Dashboards
with Interactivity, Sharing, Native Excel Exports, App Integration & more
Get technology previously reserved for billion-dollar corporations, FREE
http://pubads.g.doubleclick.net/gampad/clk?id=190641631&iu=/4140/ostg.clktrk
_____________________________________________________________________
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help, try #ltsp channel on irc.freenode.ne

[Helmut Lichtenberg]

Hi Alkis,

Alkis Georgopoulos schrieb am 17.02.2015 16:38:
> Maybe you're using a compositing window manager?
> Everything is still ok here with Ubuntu 14.04 and the gnome-flashback
> session (metacity window manager without effects).

it's good to hear that Gnome (e.g.) is running fast. So it's possible.

But we made a decision to use KDE in 2000! At this time, KDE was the most
complete window manager and this was an important reason to migrate common
users to a Linux desktop. Since then we have a userbase of 40-60 colleagues
(scientists, technicians, administration) using a KDE-Linuxdesktop on thin
clients. Around the middle of the 90s we switched to LTSP.

It is no option to switch from KDE to Gnome, xfce or any other window manager
only because a problem occurs. We have to solve it. Most of our users only
want to manage their daily work efficiently. They don't have lots of computer
skills. Rolling out a new window manager means a lot of work and likely yields
new/other problems.

> If you think that graphics in general are slower, and it's not a
> window-manager/compositing issue, then try x11perf, gtkperf or something
> similar to compare the results with the older installation.

How can I know if it is an "window-manager/compositing issue"?
The KDE window manager kwin uses either OpenGL or XRender. I found no
significant differences between those.
In my opinion this must be handled automatically. When I sometimes work with
my same account on AMD-, Intel-, or VIA-based clients I (the users) don't want
to always change some rather internal options like the composite type.

I only can imagine some generic options either on the server or on the client
side about things like compression, encryption, or general thin client
options.

I really would appreciate some hints, how to systematically get closer to the
problem.

Thanks
Helmut

--
-------------------------------------------------------------------------
Helmut Lichtenberg <***@fli.bund.de> Tel.: [phone]
Institut für Nutztiergenetik (FLI) 31535 Neustadt Germany
-------------------------------------------------------------------------

------------------------------------------------------------------------------
Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
from Actuate! Instantly Supercharge Your Business Reports and Dashboards
with Interactivity, Sharing, Native Excel Exports, App Integration & more
Get technology previously reserved for billion-dollar corporations, FREE
http://pubads.g.doubleclick.net/gampad/clk?id=190641631&iu=/4140/ostg.clktrk
_____________________________________________________________________
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help, try #ltsp channel on irc.freenode.net

[Alkis Georgopoulos]

On 18/02/2015 11:16 πμ, Helmut Lichtenberg wrote:
> I really would appreciate some hints, how to systematically get closer to the
> problem.

Well one way would be to try with gnome-flashback or LXDE or XFCE etc,
using the same kernel/xorg that you're currently using with KDE.

That would determine if it's a KDE-specific problem or not.
That's why I mentioned it...

If you see that the performance is better there, then you can start
searching about KDE, Qt, compositing managers, buffering etc.

If you see that the performance is the same, then you'll start checking
if it's a regression in Xorg etc (that's why I mentioned x11perf there).

------------------------------------------------------------------------------
Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
from Actuate! Instantly Supercharge Your Business Reports and Dashboards
with Interactivity, Sharing, Native Excel Exports, App Integration & more
Get technology previously reserved for billion-dollar corporations, FREE
http://pubads.g.doubleclick.net/gampad/clk?id=190641631&iu=/4140/ostg.clktrk
_____________________________________________________________________
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional

[John Hupp]

I want to write a script that, for my thin clients, will run a different
screen locker if the script detects that it is in an LTSP session.

The biggest problem I'm running into is testing for LTSP. I thought one
simple test would be to return the hostname and match it against ltsp*.
But "ltsp-localapps hostname" has no output. I'm running ltsp-pnp and
/bin/hostname is certainly in the image. And localapps is enabled (I'm
using that for another purpose already). In fact, I can run
"ltsp-localapps xterm" and then run hostname from there and get output
(clearly not useful for a script).

What am I missing? Does ltsp-localapps not work with apps from /bin?
If that is the case, can someone think of another solution for my purpose?

------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/
_____________________________________________________________________
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help, try #ltsp channel on irc.freenode.net

[Vagrant Cascadian]

On 2015-03-16, John Hupp wrote:
> I want to write a script that, for my thin clients, will run a different
> screen locker if the script detects that it is in an LTSP session.
>
> The biggest problem I'm running into is testing for LTSP.

Check for the LTSP_CLIENT environment variable; it should be set to True
when running in an LTSP environment.


> I thought one simple test would be to return the hostname and match it
> against ltsp*. But "ltsp-localapps hostname" has no output. I'm
> running ltsp-pnp and /bin/hostname is certainly in the image. And
> localapps is enabled (I'm using that for another purpose already). In
> fact, I can run "ltsp-localapps xterm" and then run hostname from
> there and get output (clearly not useful for a script).
>
> What am I missing? Does ltsp-localapps not work with apps from /bin?
> If that is the case, can someone think of another solution for my purpose?

That's because ltsp-localapps (on the server) essentially sets an X
property (a.k.a. xatom) that ltsp-localappsd (on the client) reads from
the running X server (on the client) and executes executes the
application specified in the X properties (on the client, displaying to
the client's X server). There's no way to read input or output from
ltsp-localapps.


live well,
vagrant

[John Hupp]

On 3/16/2015 4:34 PM, Vagrant Cascadian wrote:
> On 2015-03-16, John Hupp wrote:
>> I want to write a script that, for my thin clients, will run a different
>> screen locker if the script detects that it is in an LTSP session.
>>
>> The biggest problem I'm running into is testing for LTSP.
> Check for the LTSP_CLIENT environment variable; it should be set to True
> when running in an LTSP environment.
>
>
>> I thought one simple test would be to return the hostname and match it
>> against ltsp*. But "ltsp-localapps hostname" has no output. I'm
>> running ltsp-pnp and /bin/hostname is certainly in the image. And
>> localapps is enabled (I'm using that for another purpose already). In
>> fact, I can run "ltsp-localapps xterm" and then run hostname from
>> there and get output (clearly not useful for a script).
>>
>> What am I missing? Does ltsp-localapps not work with apps from /bin?
>> If that is the case, can someone think of another solution for my purpose?
> That's because ltsp-localapps (on the server) essentially sets an X
> property (a.k.a. xatom) that ltsp-localappsd (on the client) reads from
> the running X server (on the client) and executes executes the
> application specified in the X properties (on the client, displaying to
> the client's X server). There's no way to read input or output from
> ltsp-localapps.
>
>
> live well,
> vagrant

Thanks, Vagrant, on both counts.

------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/
_____________________________________________________________________
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help, try #ltsp channel on irc.freenode.net

[John Hupp]

With a Lubuntu 14.04 i386 and ltsp-pnp setup, and the default
network-manager used to configure networking, thin client system trays
report "Networking disabled" and "NetworkManager is not running."

In truth, network and Internet connectivity are OK, but I have been
trying to figure out how to make that disconcerting notification go away.

The system tray provided by lxpanel is not configurable (except that I
could remove it altogether, which I don't want to do). I can't set it
to not report this status.

Nm-applet is the app that provides the networking systray icon, but that
is not a service that lts.conf RM_SYSTEM_SERVICES will remove.

I'm imagining that the disabled/not-running status is reported due to
the way that network services are distributed across the LTSP local
environment and the user session on the server.

But it seems like this is probably a behavior seen on other distros
besides Lubuntu. True? If so, how are people dealing with it?

------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/
_____________________________________________________________________
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help, try #ltsp channel on irc.freenode.net

[Michal Rybárik]

Dear Helmut & everybody,

I found solution (better say workaround) for graphics performance
problems under new LTSP, which were reported few months ago by me, and
by Helmut few weeks later. It i now clear that it is not problem in LTSP
itself, neither X window manager, but somenthing is broken probably
inside X. I was doing these tests with Radeon driver and also SIS
driver, and i've got same results so this is probably nor related to
graphics chipset/driver.

I have found that performance problem depends on X server which running
inside thin client, and it does not depend on version of LTSP/XDMCP
server. If I use any of current distributions (*ubuntu 14, debian 7,
opensuse 13, ...) on thin client, then graphics performance is very
poor. If I use much older distribution on X client (while connecting to
the same LTSP/XDMCP server with current distribution), performance is
much better. Again - to get very good performance, I can run new
distribution on terminal server, but I have to run very old distribution
on thin clients.

To compare results, I did two tests with apps, where we could clearly
see the difference:

(1) oocalc test - start oocalc, select maximum possible cells on the
screen, press ctrl-c to draw selection background and dashed line around
it, and then press cursor down key 10-times, as fast as possible, and
count how long does it take to redraw selected area (deselect it) and
move cursor from 1st cell to 10th, while dashed line is still
continously (re)drawn. Test has been repeated and final values are averages:
openSUSE 11.0 (xorg-server 1.4.0, radeon 4.2.0) => *1.7* second
openSUSE 11.1 (xorg-server 1.5.2, radeon 4.3.0) => *2.3* second
openSUSE 11.2 (xorg-server 1.6.5, radeon 6.12.4) => *1.7* second
openSUSE 11.3 (xorg-server 1.8.0, radeon 6.13.0) => *2.6* seconds !!
openSUSE 11.4 (xorg-server 1.9.3, radeon 6.14.0) => *3.3* seconds !!!
openSUSE 12.1 (xorg-server crash, radeon unknown) => unknown
openSUSE 12.2 (xorg-server 1.12.3, radeon 6.14.4) => *5.5* seconds !!!!!
openSUSE 12.3 (xorg-server 1.13.2, radeon 7.0.0) => *11.3* seconds
!!!!!!!!!!!


(2) rdesktop test - connect to WinXP machine running TSreader
application, which is redrawing its window contents every 1 second, and
look how long does it take to redraw the window on thin client.
openSUSE 11.0 (xorg-server 1.4.0, radeon 4.2.0) => *superfast *(few
mililseconds)
openSUSE 11.1 (xorg-server 1.5.2, radeon 4.3.0) => *superfast *(few
mililseconds)
openSUSE 11.2 (xorg-server 1.6.5, radeon 6.12.4) => *fast* (little
flicking is visible)
openSUSE 11.3 (xorg-server 1.8.0, radeon 6.13.0) => *fast* (some
flicking is visible)
openSUSE 11.4 (xorg-server 1.9.3, radeon 6.14.0) => *slow* (flicking is
very visible)
openSUSE 12.1 (xorg-server crash, radeon unknown) => unknown
openSUSE 12.2 (xorg-server 1.12.3, radeon 6.14.4) => *slow* (flicking is
very visible)
openSUSE 12.3 (xorg-server 1.13.2, radeon 7.0.0) => *extemely slow*
(client is not able to draw whole window in 1 second intervals)

I did similar tests with some older *ubintu, Debian and CentOS
distributions, and similar xorg-server version gave me similar results.

Unfortunately I don't know how to fix it with current distributions, but
I'm happy that If I run ~7 years old openSUSE 11.2 on thin client,
performance problems are gone.... :o)

Regards,
Michal


Dňa 02/18/2015 o 10:16 AM Helmut Lichtenberg napísal(a):
> Hi Alkis,
>
> Alkis Georgopoulos schrieb am 17.02.2015 16:38:
>> Maybe you're using a compositing window manager?
>> Everything is still ok here with Ubuntu 14.04 and the gnome-flashback
>> session (metacity window manager without effects).
> it's good to hear that Gnome (e.g.) is running fast. So it's possible.
>
> But we made a decision to use KDE in 2000! At this time, KDE was the most
> complete window manager and this was an important reason to migrate common
> users to a Linux desktop. Since then we have a userbase of 40-60 colleagues
> (scientists, technicians, administration) using a KDE-Linuxdesktop on thin
> clients. Around the middle of the 90s we switched to LTSP.
>
> It is no option to switch from KDE to Gnome, xfce or any other window manager
> only because a problem occurs. We have to solve it. Most of our users only
> want to manage their daily work efficiently. They don't have lots of computer
> skills. Rolling out a new window manager means a lot of work and likely yields
> new/other problems.
>
>> If you think that graphics in general are slower, and it's not a
>> window-manager/compositing issue, then try x11perf, gtkperf or something
>> similar to compare the results with the older installation.
> How can I know if it is an "window-manager/compositing issue"?
> The KDE window manager kwin uses either OpenGL or XRender. I found no
> significant differences between those.
> In my opinion this must be handled automatically. When I sometimes work with
> my same account on AMD-, Intel-, or VIA-based clients I (the users) don't want
> to always change some rather internal options like the composite type.
>
> I only can imagine some generic options either on the server or on the client
> side about things like compression, encryption, or general thin client
> options.
>
> I really would appreciate some hints, how to systematically get closer to the
> problem.
>
> Thanks
> Helmut
>

[Helmut Lichtenberg]

Hi Michal,

Michal Rybárik schrieb am 28.04.2015 20:07:
> I found solution (better say workaround) for graphics performance
> problems under new LTSP, which were reported few months ago by me,
> and by Helmut few weeks later. It i now clear that it is not problem
> in LTSP itself, neither X window manager, but somenthing is broken
> probably inside X. I was doing these tests with Radeon driver and
> also SIS driver, and i've got same results so this is probably nor
> related to graphics chipset/driver.
>
> I have found that performance problem depends on X server which
> running inside thin client, and it does not depend on version of
> LTSP/XDMCP server. If I use any of current distributions (*ubuntu
> 14, debian 7, opensuse 13, ...) on thin client, then graphics
> performance is very poor. If I use much older distribution on X
> client (while connecting to the same LTSP/XDMCP server with current
> distribution), performance is much better. Again - to get very good
> performance, I can run new distribution on terminal server, but I
> have to run very old distribution on thin clients.
>
> To compare results, I did two tests with apps, where we could
> clearly see the difference:

you really did a lot of work and I greatly appreciate it that you took the
time to run all these test. Very interesting!
Now we are faced with the problem, what to do with the new information, you
researched. Has anybody on this list a clue, how to handle this, even a
solution?

> Unfortunately I don't know how to fix it with current distributions,
> but I'm happy that If I run ~7 years old openSUSE 11.2 on thin
> client, performance problems are gone.... :o)

Do the old versions of the client-xserver cooperate well with current xservers
on the server? Aren't there lots of missing features? And I would guess, that
recent chipsets are not supported (VIA, Openchrome, Intel Atom, etc.).

Thanks Michal and best regards
Helmut

--
-------------------------------------------------------------------------
Helmut Lichtenberg <***@fli.bund.de> Tel.: [phone]
Institut für Nutztiergenetik (FLI) 31535 Neustadt Germany
-------------------------------------------------------------------------

------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_____________________________________________________________________
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help, try #ltsp channel on irc.freenode.net

[Michal Rybarik]

Hi Helmut,

Dňa 29. 4. 2015 o 8:08 Helmut Lichtenberg napísal(a):
>
>> Unfortunately I don't know how to fix it with current distributions,
>> but I'm happy that If I run ~7 years old openSUSE 11.2 on thin
>> client, performance problems are gone.... :o)
> Do the old versions of the client-xserver cooperate well with current xservers
> on the server? Aren't there lots of missing features? And I would guess, that
> recent chipsets are not supported (VIA, Openchrome, Intel Atom, etc.).

I was expecting some compatibility problems, but it works very nice, I
don't see any issues. Screen looks exactly the same, it doesn't matter
which version I use at thin client.

You can try it easily if it will work for you - download Live CD of
distribution you wish to try, boot it on thin client, run terminal, exec
"/etc/init.d/xdm stop" (or kdm instead of xdm, depending on
distribution), X will be shutted down. Then login to liveCD and start "X
-query 1.2.3.4" where 1.2.3.4 is IP of LTSP terminal server. You should
get login screen of LTSP server.

--
Regards,
Michal Rybarik


------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_____________________________________________________________________
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help, try #ltsp channel on irc.freenode.ne

[Helmut Lichtenberg]

Michal Rybarik schrieb am 29.04.2015 13:14:
> You can try it easily if it will work for you - download Live CD of
> distribution you wish to try, boot it on thin client, run terminal, exec
> "/etc/init.d/xdm stop" (or kdm instead of xdm, depending on
> distribution), X will be shutted down. Then login to liveCD and start "X
> -query 1.2.3.4" where 1.2.3.4 is IP of LTSP terminal server. You should
> get login screen of LTSP server.

ok, I will try it next week.

Thanks a lot
Helmut

--
-------------------------------------------------------------------------
Helmut Lichtenberg <***@fli.bund.de> Tel.: [phone]
Institut für Nutztiergenetik (FLI) 31535 Neustadt Germany
-------------------------------------------------------------------------

------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_____________________________________________________________________
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help, try #ltsp channel on irc.freenode.net

[John Hupp]

Running LTSP-PNP with Lubuntu 14.04.2, I have thin clients running smoothly.

But with a fat client, Simple Scan reports "No scanners detected." And
System Tools: Printers reports "Printing service not available. Start
the service on this computer or connect to another server.”

In lts.conf, I have:
SCANNER=True
LDM_PRINTER_DEFAULT="EPSON-Stylus-C120"

On the server, System Tools: Printers reports that the Epson is Shared.

The logged in user has permission to use scanners.

The cupsd process is running on the client.

Do fat clients need avahi-daemon to detect these peripherals? Or does
anyone otherwise recognize the issue?

------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_____________________________________________________________________
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help, try #ltsp channel on irc.freenode.net

[Johan Kragsterman]

Hi!


-----John Hupp <***@prpcompany.com> skrev: -----
Till: ltsp-***@lists.sourceforge.net
Från: John Hupp <***@prpcompany.com>
Datum: 2015-04-30 23:52
Ärende: [Ltsp-discuss] Fat clients can't use printer/scanner connected to server

Running LTSP-PNP with Lubuntu 14.04.2, I have thin clients running smoothly.

But with a fat client, Simple Scan reports "No scanners detected." And
System Tools: Printers reports "Printing service not available. Start
the service on this computer or connect to another server.&#8221;

In lts.conf, I have:
SCANNER=True
LDM_PRINTER_DEFAULT="EPSON-Stylus-C120"

On the server, System Tools: Printers reports that the Epson is Shared.

The logged in user has permission to use scanners.

The cupsd process is running on the client.

Do fat clients need avahi-daemon to detect these peripherals?  Or does
anyone otherwise recognize the issue?






I run edubuntu 14.4 with 4 different chroot's. I don't know about scanners, but printers I configura with editing cups-browsed.conf and adding your own printers.conf. I never touch lts.conf.


Rgrds Johan



------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_____________________________________________________________________
Ltsp-discuss mailing list.   To un-subscribe, or change prefs, goto:
      https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help,   try #ltsp channel on irc.freenode.net




------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_____________________________________________________________________
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help, try #ltsp channel on irc.freenode.net

[John Hupp]

> -----John Hupp <***@prpcompany.com> skrev: -----
> Till: ltsp-***@lists.sourceforge.net
> Från: John Hupp <***@prpcompany.com>
> Datum: 2015-04-30 23:52
> Ärende: [Ltsp-discuss] Fat clients can't use printer/scanner connected to server
>
> Running LTSP-PNP with Lubuntu 14.04.2, I have thin clients running smoothly.
>
> But with a fat client, Simple Scan reports "No scanners detected." And
> System Tools: Printers reports "Printing service not available. Start
> the service on this computer or connect to another server."
>
> In lts.conf, I have:
> SCANNER=True
> LDM_PRINTER_DEFAULT="EPSON-Stylus-C120"
>
> On the server, System Tools: Printers reports that the Epson is Shared.
>
> The logged in user has permission to use scanners.
>
> The cupsd process is running on the client.
>
> Do fat clients need avahi-daemon to detect these peripherals? Or does
> anyone otherwise recognize the issue?
>
>
>
>
> On 5/1/2015 2:23 AM, Johan Kragsterman wrote:
>
> I run edubuntu 14.4 with 4 different chroot's. I don't know about scanners, but printers I configura with editing cups-browsed.conf and adding your own printers.conf. I never touch lts.conf.
>
>
> Rgrds Johan
>

Continuing for the moment to work just on the printing question, the
"Printing service not available" message went away after I ran (on the
server) system-config-printer and under Server: Settings checked
"Publish shared printers connected to this system" and then updated the
image.

But there are still no printers showing up on the fat client when I run
system-config-printer there, and I see that in the image printers.conf
has no content except 3 lines of comments. The content is apparently
stripped during an image update. So even if I were to set up a network
printer on the server (for the printer connected to the server), that
configuration would be stripped during an image update and so not
propagated to the fat client.

@ Johan: If the above is true, how do you then add your own printers.conf?

Neither the server nor the image has an existing cups-browsed.conf,
though from the manpage it seems that it may not be necessary. If I
understand it correctly, there are defaults for the values that probably
allow things to just work in many/most circumstances.

Is there a how-to for setting up server-connected printers for fat
clients, or can someone explain how?



------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_____________________________________________________________________
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help

[Johan Kragsterman]

Hej!


-----John Hupp <***@prpcompany.com> skrev: -----
Till: ltsp-***@lists.sourceforge.net
Från: John Hupp <***@prpcompany.com>
Datum: 2015-05-02 20:40
Ärende: Re: [Ltsp-discuss] Fat clients can't use printer/scanner connected to server

> -----John Hupp <***@prpcompany.com> skrev: -----
> Till: ltsp-***@lists.sourceforge.net
> Från: John Hupp <***@prpcompany.com>
> Datum: 2015-04-30 23:52
> Ärende: [Ltsp-discuss] Fat clients can't use printer/scanner connected to server
>
> Running LTSP-PNP with Lubuntu 14.04.2, I have thin clients running smoothly.
>
> But with a fat client, Simple Scan reports "No scanners detected." And
> System Tools: Printers reports "Printing service not available. Start
> the service on this computer or connect to another server."
>
> In lts.conf, I have:
> SCANNER=True
> LDM_PRINTER_DEFAULT="EPSON-Stylus-C120"
>
> On the server, System Tools: Printers reports that the Epson is Shared.
>
> The logged in user has permission to use scanners.
>
> The cupsd process is running on the client.
>
> Do fat clients need avahi-daemon to detect these peripherals?  Or does
> anyone otherwise recognize the issue?
>
>
>
>
> On 5/1/2015 2:23 AM, Johan Kragsterman wrote:
>
> I run edubuntu 14.4 with 4 different chroot's. I don't know about scanners, but printers I configura with editing cups-browsed.conf and adding your own printers.conf. I never touch lts.conf.
>
>
> Rgrds Johan
>

Continuing for the moment to work just on the printing question, the
"Printing service not available" message went away after I ran (on the
server) system-config-printer and under Server: Settings checked
"Publish shared printers connected to this system" and then updated the
image.

But there are still no printers showing up on the fat client when I run
system-config-printer there, and I see that in the image printers.conf
has no content except 3 lines of comments.  The content is apparently
stripped during an image update.  So even if I were to set up a network
printer on the server (for the printer connected to the server), that
configuration would be stripped during an image update and so not
propagated to the fat client.

@ Johan: If the above is true, how do you then add your own printers.conf?





Here's how it looks like in my fat clients filesystem:

***@ws0:/etc/cups$ ls
cups-browsed.conf ppd raw.types subscriptions.conf.O
cupsd.conf printers.conf snmp.conf
cups-files.conf printers.conf.O ssl
interfaces raw.convs subscriptions.conf
***@ws0:/etc/cups$


As you can see, I got all the normal cups files there, I don't know if there is a different with the ltsp-pnp. I didn't build my chroot's with any special options.

As you also can see, I got both printers.conf and printers.conf.0. The latter is the one that cups is building automatically, and it is named like that, because the printers.conf already exists.

Here's my cups-browsed.conf:

***@ws0:/etc/cups$ cat cups-browsed.conf
# Which protocols will we use to discover printers on the network?
# Can use DNSSD and/or CUPS, or 'none' for neither.
BrowseRemoteProtocols dnssd cups

# Which protocols will we use to broadcast shared local printers to the network?
# Can use DNSSD and/or CUPS, or 'none' for neither.
# Only CUPS is actually supported, as DNSSD is done by CUPS itself (we ignore
# DNSSD in this directive).
# BrowseLocalProtocols none

# Settings of this directive apply to both BrowseRemoteProtocols and
# BrowseLocalProtocols.
# Can use DNSSD and/or CUPS, or 'none' for neither.
BrowseProtocols dnssd cups

# Only browse remote printers from selected servers
# BrowseAllow cups.example.com
BrowseAllow 192.168.10.250
BrowseAllow 192.168.10.0/24
BrowseAllow 192.168.10.0/255.255.255.0

# Use BrowsePoll to poll a particular CUPS server
BrowsePoll 192.168.10.250:631
BrowsePoll 192.168.10.250
# BrowsePoll cups.example.com:631/version=1.1

# Use DomainSocket to access the local CUPS daemon via another than the
# default domain socket
# DomainSocket /var/run/cups/cups.sock

# Set CreateIPPPrinterQueues to "Yes" to let cups-browsed discover IPP
# network printers (native printers, not CUPS queues) with known page
# description languages (PWG Raster, PDF, PostScript, PCL XL, PCL
# 5c/e) in the local network and auto-create PPD-less print queues for
# them (using a System V interface script to control the filter
# chain). Clients have to IPP-poll the capabilities of the printer and
# send option settings as standard IPP attributes. We do not poll the
# capabilities by ourselves to not wake up the printer from
# power-saving mode when creating the queues. Jobs have to be sent in
# PDF format. Other formats are not accepted. This functionality is
# primarily for mobile devices running CUPS to not need a printer
# setup tool nor a collection of printer drivers and PPDs.

CreateIPPPrinterQueues Yes

# The AutoShutdown directive specifies whether cups-browsed should
# automatically terminate when it has no local raw queues set up
# pointing to any discovered remote printers (auto shutdown
# mode). Setting it to "On" activates the auto-shutdown mode, setting
# it to "Off" deactiivates it (the default). The special mode "avahi"
# turns auto shutdown off while avahi-daemon is running and on when
# avahi-daemon stops. This allows running cups-browsed on-demand when
# avahi-daemon is run on-demand.

# AutoShutdown Off
# AutoShutdown On
# AutoShutdown avahi

# The AutoShutdownTimeout directive specifies after how many seconds
# without local raw queues set up pointing to any discovered remote
# printers cups-browsed should actually shut down in auto shutdown
# mode. Default is 30 seconds, 0 means immediate shutdown.

# AutoShutdownTimeout 30

# Unknown directives are ignored, also unknown values.
***@ws0:/etc/cups$




And here's my printers.conf:

***@ws0:/etc/cups$ sudo cat printers.conf
[sudo] password for admin:
# Printer configuration file for CUPS v1.7.2
# Written by cupsd
# DO NOT EDIT THIS FILE WHEN CUPSD IS RUNNING
<Printer HP-Color-LaserJet-CP4520-Series>
UUID urn:uuid:21be3120-0f61-3fa8-7d7d-7129f1350fab
Info HP Color LaserJet CP4520 Series [90D6CD]
Location print
DeviceURI ipp://print.local:631/print
PPDTimeStamp *
State Idle
StateTime 1430565618
Type 4
Accepting Yes
Shared No
ColorManaged Yes
JobSheets none none
QuotaPeriod 0
PageLimit 0
KLimit 0
OpPolicy default
ErrorPolicy retry-job
Option cups-browsed true
</Printer>
***@ws0:/etc/cups$



I got this file through first editing cups-browsed, so that I could configure the exact printer I wanted, and then copy the file from my running fat client, and putting it into the chroot(and of coarse rebuilding the image). You will also need a root user in the fat client image.


Regards Johan


Neither the server nor the image has an existing cups-browsed.conf,
though from the manpage it seems that it may not be necessary.  If I
understand it correctly, there are defaults for the values that probably
allow things to just work in many/most circumstances.

Is there a how-to for setting up server-connected printers for fat
clients, or can someone explain how?



------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_____________________________________________________________________
Ltsp-discuss mailing list.   To un-subscribe, or change prefs, goto:
      https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help,   try #ltsp channel on irc.freenode.net



------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_____________________________________________________________________
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help, try #ltsp channel on irc.freenode.net

[Alkis Georgopoulos]

Hi guys, sorry I don't have time for a long answer, but a few quick hints,

by default the clients have CUPS_SERVER=server, so they see whatever
printers the server has or sees,
but, since a few cups versions ago, cups does a DNS check for the
ServerName and that means that the ltsp server needs to actually be
called "server" in order for that to work.
Or one could specify CUPS_SERVER in lts.conf and have DNS working
correctly in both the server and the clients.
Or one could check "share printers" and "see printers from other
systems" on the server and the chroot (or ltsp-pnp), and set
CUPS_SERVER=localhost for fat clients.

That's all from me, sorry again, no time for more detailed feedback
after those hints! :)

------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_____________________________________________________________________
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help, try #ltsp channel on irc.freenode.net

[John Hupp]

On 5/2/2015 4:13 PM, Alkis Georgopoulos wrote:
> Hi guys, sorry I don't have time for a long answer, but a few quick hints,
>
> by default the clients have CUPS_SERVER=server, so they see whatever
> printers the server has or sees,
> but, since a few cups versions ago, cups does a DNS check for the
> ServerName and that means that the ltsp server needs to actually be
> called "server" in order for that to work.
> Or one could specify CUPS_SERVER in lts.conf and have DNS working
> correctly in both the server and the clients.
> Or one could check "share printers" and "see printers from other
> systems" on the server and the chroot (or ltsp-pnp), and set
> CUPS_SERVER=localhost for fat clients.
>
> That's all from me, sorry again, no time for more detailed feedback
> after those hints! :)
>
> ------------------------------------------------------------------------------
> One dashboard for servers and applications across Physical-Virtual-Cloud
> Widest out-of-the-box monitoring support with 50+ applications
> Performance metrics, stats and reports that give you Actionable Insights
> Deep dive visibility with transaction tracing using APM Insight.
> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
> _____________________________________________________________________
> Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
> https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
> For additional LTSP help, try #ltsp channel on irc.freenode.net

Thanks, Alkis, I'll look this over. (I just saw this after I finished
my previous post.)

------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_____________________________________________________________________
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help, try #ltsp channel on irc.freenode.net

[John Hupp]

On 5/2/2015 4:13 PM, Alkis Georgopoulos wrote:
> Hi guys, sorry I don't have time for a long answer, but a few quick hints,
>
> by default the clients have CUPS_SERVER=server, so they see whatever
> printers the server has or sees,
> but, since a few cups versions ago, cups does a DNS check for the
> ServerName and that means that the ltsp server needs to actually be
> called "server" in order for that to work.
> Or one could specify CUPS_SERVER in lts.conf and have DNS working
> correctly in both the server and the clients.
> Or one could check "share printers" and "see printers from other
> systems" on the server and the chroot (or ltsp-pnp), and set
> CUPS_SERVER=localhost for fat clients.
>
> That's all from me, sorry again, no time for more detailed feedback
> after those hints! :)

Thanks again, Alkis.

I subsequently found
https://help.ubuntu.com/community/UbuntuLTSP/localprinter (updated
2015-04-08), but it turns out that even its specified configuration is
not needed.

I was tripping on the fact that trying to run system-config-printer on
the client results in a BAD REQUEST error, but this seems due to regular
users not having privileges to configure printers.

All that was needed was to run system-config-printer on the server, and
under Server: Settings check "Publish shared printers connected to this
system." I may have also needed to update the client image, but
probably not. I am not using /etc/cups/client.conf or ~/.cups/client.conf.

After that, the server-connected printer automatically shows up in any
app's Print dialog.

Various references also say to check "Show printers shared by other
systems," but that setting is no longer presented under the current
version of system-config-printer.

Interestingly, I got no discernable results from setting
CUPS_SERVER=<ltsp-server> or KEEP_SYSTEM_SERVICES=cups in lts.conf. On
the fat client, cups was running even without the KEEP_SYSTEM_SERVICES
line. And setting CUPS_SERVER in lts.conf did not set up this
environment variable on the fat client.

------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_____________________________________________________________________
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help, try #ltsp channel on irc.freenode.net

[John Hupp]

[snip]

On 5/2/2015 3:22 PM, Johan Kragsterman wrote:
> Here's how it looks like in my fat clients filesystem:
>
> ***@ws0:/etc/cups$ ls
> cups-browsed.conf ppd raw.types subscriptions.conf.O
> cupsd.conf printers.conf snmp.conf
> cups-files.conf printers.conf.O ssl
> interfaces raw.convs subscriptions.conf
> ***@ws0:/etc/cups$
>
>
> As you can see, I got all the normal cups files there, I don't know if there is a different with the ltsp-pnp. I didn't build my chroot's with any special options.
>
> As you also can see, I got both printers.conf and printers.conf.0. The latter is the one that cups is building automatically, and it is named like that, because the printers.conf already exists.
>
> Here's my cups-browsed.conf:
>
> ***@ws0:/etc/cups$ cat cups-browsed.conf
> # Which protocols will we use to discover printers on the network?
> # Can use DNSSD and/or CUPS, or 'none' for neither.
> BrowseRemoteProtocols dnssd cups
>
> # Which protocols will we use to broadcast shared local printers to the network?
> # Can use DNSSD and/or CUPS, or 'none' for neither.
> # Only CUPS is actually supported, as DNSSD is done by CUPS itself (we ignore
> # DNSSD in this directive).
> # BrowseLocalProtocols none
>
> # Settings of this directive apply to both BrowseRemoteProtocols and
> # BrowseLocalProtocols.
> # Can use DNSSD and/or CUPS, or 'none' for neither.
> BrowseProtocols dnssd cups
>
> # Only browse remote printers from selected servers
> # BrowseAllow cups.example.com
> BrowseAllow 192.168.10.250
> BrowseAllow 192.168.10.0/24
> BrowseAllow 192.168.10.0/255.255.255.0
>
> # Use BrowsePoll to poll a particular CUPS server
> BrowsePoll 192.168.10.250:631
> BrowsePoll 192.168.10.250
> # BrowsePoll cups.example.com:631/version=1.1
>
> # Use DomainSocket to access the local CUPS daemon via another than the
> # default domain socket
> # DomainSocket /var/run/cups/cups.sock
>
> # Set CreateIPPPrinterQueues to "Yes" to let cups-browsed discover IPP
> # network printers (native printers, not CUPS queues) with known page
> # description languages (PWG Raster, PDF, PostScript, PCL XL, PCL
> # 5c/e) in the local network and auto-create PPD-less print queues for
> # them (using a System V interface script to control the filter
> # chain). Clients have to IPP-poll the capabilities of the printer and
> # send option settings as standard IPP attributes. We do not poll the
> # capabilities by ourselves to not wake up the printer from
> # power-saving mode when creating the queues. Jobs have to be sent in
> # PDF format. Other formats are not accepted. This functionality is
> # primarily for mobile devices running CUPS to not need a printer
> # setup tool nor a collection of printer drivers and PPDs.
>
> CreateIPPPrinterQueues Yes
>
> # The AutoShutdown directive specifies whether cups-browsed should
> # automatically terminate when it has no local raw queues set up
> # pointing to any discovered remote printers (auto shutdown
> # mode). Setting it to "On" activates the auto-shutdown mode, setting
> # it to "Off" deactiivates it (the default). The special mode "avahi"
> # turns auto shutdown off while avahi-daemon is running and on when
> # avahi-daemon stops. This allows running cups-browsed on-demand when
> # avahi-daemon is run on-demand.
>
> # AutoShutdown Off
> # AutoShutdown On
> # AutoShutdown avahi
>
> # The AutoShutdownTimeout directive specifies after how many seconds
> # without local raw queues set up pointing to any discovered remote
> # printers cups-browsed should actually shut down in auto shutdown
> # mode. Default is 30 seconds, 0 means immediate shutdown.
>
> # AutoShutdownTimeout 30
>
> # Unknown directives are ignored, also unknown values.
> ***@ws0:/etc/cups$
>
>
>
>
> And here's my printers.conf:
>
> ***@ws0:/etc/cups$ sudo cat printers.conf
> [sudo] password for admin:
> # Printer configuration file for CUPS v1.7.2
> # Written by cupsd
> # DO NOT EDIT THIS FILE WHEN CUPSD IS RUNNING
> <Printer HP-Color-LaserJet-CP4520-Series>
> UUID urn:uuid:21be3120-0f61-3fa8-7d7d-7129f1350fab
> Info HP Color LaserJet CP4520 Series [90D6CD]
> Location print
> DeviceURI ipp://print.local:631/print
> PPDTimeStamp *
> State Idle
> StateTime 1430565618
> Type 4
> Accepting Yes
> Shared No
> ColorManaged Yes
> JobSheets none none
> QuotaPeriod 0
> PageLimit 0
> KLimit 0
> OpPolicy default
> ErrorPolicy retry-job
> Option cups-browsed true
> </Printer>
> ***@ws0:/etc/cups$
>
>
>
> I got this file through first editing cups-browsed, so that I could configure the exact printer I wanted, and then copy the file from my running fat client, and putting it into the chroot(and of coarse rebuilding the image). You will also need a root user in the fat client image.
>
>
> Regards Johan
>
>
> Neither the server nor the image has an existing cups-browsed.conf,
> though from the manpage it seems that it may not be necessary. If I
> understand it correctly, there are defaults for the values that probably
> allow things to just work in many/most circumstances.
>
> Is there a how-to for setting up server-connected printers for fat
> clients, or can someone explain how?
>
>

@ Johan: Thanks for detailing that, but we may be dealing with a
difference between ltsp-pnp and the ltsp non-pnp that you use. I'm
finding that ltsp-pnp strips the configuration from printers.conf with
every image update, so I don't see how your method will work for me.

But I did some further reading, including
https://sourceforge.net/p/ltsp/mailman/ltsp-discuss/thread/***@alkis/
https://www.cups.org/documentation.php/doc-1.7/sharing.html?VERSION=1.7&Q=
https://www.cups.org/documentation.php/doc-1.7/man-client.conf.html?VERSION=1.7&Q=

For some reason I can create an /etc/cups/client.conf that is included
in the image (I double-checked by mounting the image) but is not in the
filesystem loaded on the fat client.

Leaving that mystery for later research, I worked around it by creating
~/.cups/client.conf, and if I add this sole line to it (plus an EOL
character), the server-connected printer is automatically available for
the fat client:
ServerName 192.168.1.8

Great! But I'm using DHCP for the server IP address for the sake of
easy deployment, so I would like to use the hostname option for
ServerName instead of the IP address option. So I tried (to provide a
specific example):
ServerName Dell-Lubuntu

This fails. It seems that by "hostname" they mean fully-qualified
domain name, as in the examples from the older documentation:
https://www.cups.org/documentation.php/doc-1.5/ref-client-conf.html

On the LAN, does anyone know how a hostname can be specified as a FQDN?

(I'll also note that in one of the above links, Alkis suggested trying
just "ServerName server" but that failed for me.)


------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_____________________________________________________________________
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help, try #ltsp channel on irc.freenode.net

[John Hupp]

I'm casting around for ideas on how to best detect that the current
environment is on a fat client (not on a thin client or on the server
itself), and then run a startup script to modify pcmanfm's volume
management preferences. The script should run whether the user is
logging into the server, a thin client or a fat client, so I would
probably not want to run it via an RCFILE entry in lts.conf. Maybe an
upstart job (I'm on Lubuntu 14.04).

(I'm tweaking the behavior of what happens when a USB flash drive, audio
CD or data disc is inserted. My first approach was to turn off
pcmanfm's "Mount removable media automatically when they are inserted"
setting for all users except user1. I arrived at an acceptable solution
in a thin-only environment that involved ltspfs notifications, but the
same thing doesn't work for fat clients.)

I know that pcmanfm's preferences are in
~/.config/pcmanfm/lubuntu/pcmanfm.conf, but don't have any slick ideas
beyond that. My first idea is to make a script that looks at whether
the output of hostname contains "ltsp" or not.

Other ideas?

------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_____________________________________________________________________
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help, try #ltsp channel on irc.freenode.net

[John Hupp]

On 5/8/2015 6:46 PM, John Hupp wrote:
> I'm casting around for ideas on how to best detect that the current
> environment is on a fat client (not on a thin client or on the server
> itself), and then run a startup script to modify pcmanfm's volume
> management preferences. The script should run whether the user is
> logging into the server, a thin client or a fat client, so I would
> probably not want to run it via an RCFILE entry in lts.conf. Maybe an
> upstart job (I'm on Lubuntu 14.04).
>
> (I'm tweaking the behavior of what happens when a USB flash drive,
> audio CD or data disc is inserted. My first approach was to turn off
> pcmanfm's "Mount removable media automatically when they are inserted"
> setting for all users except user1. I arrived at an acceptable
> solution in a thin-only environment that involved ltspfs
> notifications, but the same thing doesn't work for fat clients.)
>
> I know that pcmanfm's preferences are in
> ~/.config/pcmanfm/lubuntu/pcmanfm.conf, but don't have any slick ideas
> beyond that. My first idea is to make a script that looks at whether
> the output of hostname contains "ltsp" or not.
>
> Other ideas?

Here is the script code from an upstart job I'm trying to get working
(but which doesn't work yet):

# test for a fat client:
if [[ $(hostname | grep 'ltsp') == ltsp* ]]; then
sed 's/mount_removable=0/mount_removable=1/'
<~/.config/pcmanfm/lubuntu/pcmanfm.conf
>~/.config/pcmanfm/lubuntu/pcmanfm.conf
# test for a thin client:
elif [ -n "$LTSP_CLIENT" ]; then
sed 's/mount_removable=1/mount_removable=0/'
<~/.config/pcmanfm/lubuntu/pcmanfm.conf
>~/.config/pcmanfm/lubuntu/pcmanfm.conf
# machine must be the server
else
sed 's/mount_removable=0/mount_removable=1/'
<~/.config/pcmanfm/lubuntu/pcmanfm.conf
>~/.config/pcmanfm/lubuntu/pcmanfm.conf
fi

Assuming that it was tripping on ~, I also tried:

# test for a fat client:
if [[ $(hostname | grep 'ltsp') == ltsp* ]]; then
sed 's/mount_removable=0/mount_removable=1/'
<$HOME/.config/pcmanfm/lubuntu/pcmanfm.conf
>$HOME/.config/pcmanfm/lubuntu/pcmanfm.conf
# test for a thin client:
elif [ -n "$LTSP_CLIENT" ]; then
sed 's/mount_removable=1/mount_removable=0/'
<$HOME/.config/pcmanfm/lubuntu/pcmanfm.conf
>$HOME/.config/pcmanfm/lubuntu/pcmanfm.conf
# machine must be the server
else
sed 's/mount_removable=0/mount_removable=1/'
<$HOME/.config/pcmanfm/lubuntu/pcmanfm.conf
>$HOME/.config/pcmanfm/lubuntu/pcmanfm.conf
fi

I'm guessing that my file identification strings need fixing up, but I'm
such a poor scripter that this could only be one of several mistakes.

Check my homework?

------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_____________________________________________________________________
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help, try #ltsp channel on irc.freenode.net

[John Hupp]

On 5/9/2015 11:18 AM, John Hupp wrote:
> On 5/8/2015 6:46 PM, John Hupp wrote:
>> I'm casting around for ideas on how to best detect that the current
>> environment is on a fat client (not on a thin client or on the server
>> itself), and then run a startup script to modify pcmanfm's volume
>> management preferences. The script should run whether the user is
>> logging into the server, a thin client or a fat client, so I would
>> probably not want to run it via an RCFILE entry in lts.conf. Maybe an
>> upstart job (I'm on Lubuntu 14.04).
>>
>> (I'm tweaking the behavior of what happens when a USB flash drive,
>> audio CD or data disc is inserted. My first approach was to turn off
>> pcmanfm's "Mount removable media automatically when they are
>> inserted" setting for all users except user1. I arrived at an
>> acceptable solution in a thin-only environment that involved ltspfs
>> notifications, but the same thing doesn't work for fat clients.)
>>
>> I know that pcmanfm's preferences are in
>> ~/.config/pcmanfm/lubuntu/pcmanfm.conf, but don't have any slick
>> ideas beyond that. My first idea is to make a script that looks at
>> whether the output of hostname contains "ltsp" or not.
>>
>> Other ideas?
> [snip]

Here is an improved upstart job (/etc/init/pcmanfm-ltsp.conf). The
script code is tested as working when run manually as a script, but it
was not effective as an upstart job. Perhaps environment variables like
$HOME are not available to upstart jobs even though this one doesn't run
until after the desktop session has started and $HOME should be defined
at that point.

But given that the code works as a script, I can probably get what I
want by setting that up with a desktop shortcut in an autostarting location.

# Edit pcmanfm Volume Management preferences at startup
#
# Set prefs based on whether the machine is a thin client or a
server/fat client

description "Edit pcmanfm Volume Management preferences based on machine
type"

start on desktop-session-start

script
# test for a fat client:
if [[ $(hostname | grep 'ltsp') == ltsp* ]]; then
sed -i 's/mount_removable=0/mount_removable=1/'
"$HOME/.config/pcmanfm/lubuntu/pcmanfm.conf"
# test for a thin client:
elif [ -n "$LTSP_CLIENT" ]; then
sed -i 's/mount_removable=0/mount_removable=0/'
"$HOME/.config/pcmanfm/lubuntu/pcmanfm.conf"
# machine must be the server
else
sed -i 's/mount_removable=0/mount_removable=1/'
"$HOME/.config/pcmanfm/lubuntu/pcmanfm.conf"
fi
end script


------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_____________________________________________________________________
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help, try #ltsp channel on irc.freenode.net