Re: xorg: Maximum number of clients reached

[Stephen Partington]

I need to install and check.
On Jun 12, 2015 4:20 PM, "Michael Butash"  wrote:

>  Stephen, out of curiosity, what does your xrandr show as a max
> framebuffer size on your quadro?
>
> mb@host:~$ xrandr | grep maximum
> Screen 0: minimum 320 x 200, current 11520 x 1200, maximum 16384 x 16384
>
> This was a big limiter for me, in the past I couldn't figure out why my
> old ATI 5800 card with 6 ports wouldn't support a full, single framebuffer,
> but was internally limited to 8192x8192, with the 6xxx+ supporting
> 16384x16384.  Xorg wasn't too forthcoming with that info, and it was prior
> to xrandr support in their drivers, so totally left me scratching my head
> until escalating with AMD support to an engineer with a clue that told me
> that.
>
> With the advent of 4k displays, they still seem limited to that, which
> means I can only do 4x wide until vendors give to open that up.
>
> Thanks!
>
> -mb
>
>
> On 06/12/2015 03:55 PM, Michael Butash wrote:
>
> Next time I have an absolute need to upgrade hardware, I plan on avoiding
> ati/amd at all costs.  After dealing with them for a good 5 years as the
> only real viable option to run my displays, only to be wrought with
> constant disappointment, problems, and frustration.  Buying highly
> overpriced quadro cards might be money well spent at this point, but I
> still despise nvidia that they're really little other than rebranded, and
> marked-up normal video cards with driver-locked (to bios-id) features.
>
> That said, going to set up some ebay agents to look for decent quadro's to
> snipe.  I had good luck getting my last few amd cards that way on the
> cheap, gotta love jbidwatcher for cheating some other person with a
> last-second bid.
>
> Thanks as always for the input Stephen.
>
> -mb
>
>
> On 06/12/2015 03:20 PM, Stephen Partington wrote:
>
>  I have almost given up on ATI, if i want just multiple screens i would
> look into the Quadro NVS cards. Such as the NVS 510 or the K1200. They may
> be very proprietary to get running, but my success with Nvidia cards in
> both linux and windows really makes it worthwhile. These cards will only do
> a single monitor, but they are cheap enough to run 2 cards with reasonable
> usability. or one NVS and one more Gamer friendly card.
>
>
>
> ---
> PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
> To subscribe, unsubscribe, or to change your mail 
> settings:http://lists.phxlinux.org/mailman/listinfo/plug-discuss
>
>
>
> ---
> PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
> To subscribe, unsubscribe, or to change your mail settings:
> http://lists.phxlinux.org/mailman/listinfo/plug-discuss
>
---
PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
To subscribe, unsubscribe, or to change your mail settings:
http://lists.phxlinux.org/mailman/listinfo/plug-discuss

Re: xorg: Maximum number of clients reached

[Michael Butash]

  
  
Stephen, out of curiosity, what does
  your xrandr show as a max framebuffer size on your quadro?
  
  mb@host:~$ xrandr | grep maximum
  Screen 0: minimum 320 x 200, current 11520 x 1200, maximum 16384 x
  16384
  
  This was a big limiter for me, in the past I couldn't figure out
  why my old ATI 5800 card with 6 ports wouldn't support a full,
  single framebuffer, but was internally limited to 8192x8192, with
  the 6xxx+ supporting 16384x16384.  Xorg wasn't too forthcoming
  with that info, and it was prior to xrandr support in their
  drivers, so totally left me scratching my head until escalating
  with AMD support to an engineer with a clue that told me that.
  
  With the advent of 4k displays, they still seem limited to that,
  which means I can only do 4x wide until vendors give to open that
  up.
  
  Thanks!
  
  -mb
  
  
  On 06/12/2015 03:55 PM, Michael Butash wrote:


  
  Next time I have an absolute need to
upgrade hardware, I plan on avoiding ati/amd at all costs. 
After dealing with them for a good 5 years as the only real
viable option to run my displays, only to be wrought with
constant disappointment, problems, and frustration.  Buying
highly overpriced quadro cards might be money well spent at this
point, but I still despise nvidia that they're really little
other than rebranded, and marked-up normal video cards with
driver-locked (to bios-id) features.

That said, going to set up some ebay agents to look for decent
quadro's to snipe.  I had good luck getting my last few amd
cards that way on the cheap, gotta love jbidwatcher for cheating
some other person with a last-second bid.

Thanks as always for the input Stephen.

-mb


On 06/12/2015 03:20 PM, Stephen Partington wrote:
  
  

  I have almost given up on ATI, if i want just
multiple screens i would look into the Quadro NVS cards.
Such as the NVS 510 or the K1200. They may be very
proprietary to get running, but my success with Nvidia cards
in both linux and windows really makes it worthwhile. These
cards will only do a single monitor, but they are cheap
enough to run 2 cards with reasonable usability. or one NVS
and one more Gamer friendly card.
  

  
  
  
  
  ---
PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
To subscribe, unsubscribe, or to change your mail settings:
http://lists.phxlinux.org/mailman/listinfo/plug-discuss


  

---
PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
To subscribe, unsubscribe, or to change your mail settings:
http://lists.phxlinux.org/mailman/listinfo/plug-discuss

Re: xorg: Maximum number of clients reached

[Michael Butash]

  
  
Next time I have an absolute need to
  upgrade hardware, I plan on avoiding ati/amd at all costs.  After
  dealing with them for a good 5 years as the only real viable
  option to run my displays, only to be wrought with constant
  disappointment, problems, and frustration.  Buying highly
  overpriced quadro cards might be money well spent at this point,
  but I still despise nvidia that they're really little other than
  rebranded, and marked-up normal video cards with driver-locked (to
  bios-id) features.
  
  That said, going to set up some ebay agents to look for decent
  quadro's to snipe.  I had good luck getting my last few amd cards
  that way on the cheap, gotta love jbidwatcher for cheating some
  other person with a last-second bid.
  
  Thanks as always for the input Stephen.
  
  -mb
  
  
  On 06/12/2015 03:20 PM, Stephen Partington wrote:


  
I have almost given up on ATI, if i want just
  multiple screens i would look into the Quadro NVS cards. Such
  as the NVS 510 or the K1200. They may be very proprietary to
  get running, but my success with Nvidia cards in both linux
  and windows really makes it worthwhile. These cards will only
  do a single monitor, but they are cheap enough to run 2 cards
  with reasonable usability. or one NVS and one more Gamer
  friendly card.

  

  

---
PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
To subscribe, unsubscribe, or to change your mail settings:
http://lists.phxlinux.org/mailman/listinfo/plug-discuss

Re: xorg: Maximum number of clients reached

[Stephen Partington]

I have almost given up on ATI, if i want just multiple screens i would look
into the Quadro NVS cards. Such as the NVS 510 or the K1200. They may be
very proprietary to get running, but my success with Nvidia cards in both
linux and windows really makes it worthwhile. These cards will only do a
single monitor, but they are cheap enough to run 2 cards with reasonable
usability. or one NVS and one more Gamer friendly card.

On Fri, Jun 12, 2015 at 3:14 PM, Michael Butash  wrote:

> Looking at those results after responding, I noticed I only have 5 okular
> instances open, the pdf reader (sadly I stay mired in pdf documentation
> often), when it's spawning 13 xclient sessions.  I wonder if this is just
> being stupid is what's blowing it out somehow.
>
> PDF's, another bit lf legacy windoze technology I wish would die. I hate
> adobe, but it's become defacto for doc standards when libreoffice vs.
> openoffice vs. microsith office stay in a pissing match with each other.
>
> -mb
>
>
>
> On 06/12/2015 02:56 PM, Michael Butash wrote:
>
>> Thanks for the comments Matt - in line.
>>
>> On 06/12/2015 02:18 PM, Matt Graham wrote:
>>
>>> On 2015-06-11 17:20, Michael Butash wrote:
>>>
 [X reaching a maximum number of clients is a problem] in that it
 simply refuses to open new [X clients], and [I] find this happens
 more and more these days. Am I like the only actual person to use
 linux these days that this occurs with?

>>>
>>> I've never seen this happen.  What do you get when this is happening and
>>> you do "xlsclients | sort | uniq -c" ?  I currently have 64 X clients
>>> running here.  Most KDE things show up as 2 or 3 X clients.  plasma-desktop
>>> shows up as 20.  firefox shows up as 1.
>>>
>> I was actually trying to remember that command as I've seen it referenced
>> and checked before, so thanks for that.
>>
>> Every now and then problem will piss me off royally as literally my pc
>> won't be able to wake my monitors up out of dpms sleep because of this
>> (presumably), thus I have to hard reboot.  Odd part is when it freaks like
>> this, even ssh'ing in from my laptop, killing some pids (ie.
>> chrome|chromium I have an xargs script I keep to seek and destroy all), and
>> it still won't wake then.  I need add something to watch these...
>>
>> After it did this the other day upon composition of the email, I killed
>> the chrom*'s, and it's been a bit stable.  xrestop was another that was
>> recommended to watch, but by the time I hit that limit, even xrestop gives
>> me the "max number of clients reached", even though a cli application...
>>
>> Another way I know my system is almost ready to implode is waking up out
>> of monitor sleep, expecting to see the simple-locker screen, rather I see a
>> full desktop, unhidden, but I can't actually click on anything.  Yeah, so
>> much for privacy/security, but at least someone couldn't interact with it.
>> It requires me to ctrl-alt-F1, switch to a tty, and back to F7 to see the
>> locker again, unlock, and actually use my desktop.  Seem another byproduct
>> of xorg freaking out.
>>
>> At the (working) moment...
>>
>> mb@host:~$ xlsclients | sort | wc -l
>> 144
>> mb@host:~$ xlsclients | sort | uniq -c
>>   1 host  baloo_file
>>   1 host  bamfdaemon
>>   1 host  Banshee
>>   1 host  blueman-applet
>>   1 host  cairo-dock
>>   1 host  chromium-browser
>>   1 host  dolphin
>>   1 host  eom
>>   1 host  evince
>>   1 host  gcalctool
>>   1 host  gkrellm
>>   1 host  gnome-terminal
>>   1 host  ibus-ui-gtk3
>>   1 host  ibus-x11
>>   1 host  kactivitymanagerd
>>   3 host  kded4
>>   1 host  'kdeinit4: kded4 [kdeinit]'
>>   1 host  'kdeinit4: ksmserver [kdeinit]'
>>   1 host  kglobalaccel
>>   1 host  klipper
>>   1 host  kmix
>>   1 host  knotify4
>>   1 host  konsole
>>   1 host  korgac
>>   1 host  krunner
>>   2 host  ksmserver
>>   1 host  kuiserver
>>   1 host  kwalletd
>>   6 host  kwin
>>   1 host  mate-screensaver
>>   1 host  nautilus
>>   5 host  okular
>>   1 host  pavucontrol
>>   1 host  Pidgin
>>   1 host  plasma-desktop
>>   1 host  pluma
>>   1 host  polkit-kde-authentication-agent-1
>>   1 host  soffice.bin
>>   1 host  thunderbird
>>   1 host  transmission-remote-gtk
>>   2 host  /usr/bin/baloo_file
>>   3 host  /usr/bin/dolphin
>>   1 host  /usr/bin/kactivitymanagerd
>>   2 host  /usr/bin/kglobalaccel
>>   3 host  /usr/bin/klipper
>>   8 host  /usr/bin/kmix
>>   8 host  /usr/bin/konsole
>>   2 host  /usr/bin/korgac
>>   3 host  /usr/bin/krunner
>>   2 host  /usr/bin/kuiserver
>>   2 host  /usr/bin/kwalletd
>>  13 host  /usr/bin/okular
>>  29 host  /usr/bin/plasma-desktop
>>   2 host /usr/lib/kde4/libexec/polkit-kde-authentication-agent-1
>>   2 host  /usr/lib/libreoffice/program/soffice.bin
>>   1 host  /usr/l

Re: xorg: Maximum number of clients reached

[Michael Butash]

Looking at those results after responding, I noticed I only have 5 
okular instances open, the pdf reader (sadly I stay mired in pdf 
documentation often), when it's spawning 13 xclient sessions.  I wonder 
if this is just being stupid is what's blowing it out somehow.


PDF's, another bit lf legacy windoze technology I wish would die. I hate 
adobe, but it's become defacto for doc standards when libreoffice vs. 
openoffice vs. microsith office stay in a pissing match with each other.


-mb


On 06/12/2015 02:56 PM, Michael Butash wrote:

Thanks for the comments Matt - in line.

On 06/12/2015 02:18 PM, Matt Graham wrote:

On 2015-06-11 17:20, Michael Butash wrote:

[X reaching a maximum number of clients is a problem] in that it
simply refuses to open new [X clients], and [I] find this happens
more and more these days. Am I like the only actual person to use
linux these days that this occurs with?


I've never seen this happen.  What do you get when this is happening 
and you do "xlsclients | sort | uniq -c" ?  I currently have 64 X 
clients running here.  Most KDE things show up as 2 or 3 X clients.  
plasma-desktop shows up as 20.  firefox shows up as 1.
I was actually trying to remember that command as I've seen it 
referenced and checked before, so thanks for that.


Every now and then problem will piss me off royally as literally my pc 
won't be able to wake my monitors up out of dpms sleep because of this 
(presumably), thus I have to hard reboot.  Odd part is when it freaks 
like this, even ssh'ing in from my laptop, killing some pids (ie. 
chrome|chromium I have an xargs script I keep to seek and destroy 
all), and it still won't wake then.  I need add something to watch 
these...


After it did this the other day upon composition of the email, I 
killed the chrom*'s, and it's been a bit stable.  xrestop was another 
that was recommended to watch, but by the time I hit that limit, even 
xrestop gives me the "max number of clients reached", even though a 
cli application...


Another way I know my system is almost ready to implode is waking up 
out of monitor sleep, expecting to see the simple-locker screen, 
rather I see a full desktop, unhidden, but I can't actually click on 
anything.  Yeah, so much for privacy/security, but at least someone 
couldn't interact with it.  It requires me to ctrl-alt-F1, switch to a 
tty, and back to F7 to see the locker again, unlock, and actually use 
my desktop.  Seem another byproduct of xorg freaking out.


At the (working) moment...

mb@host:~$ xlsclients | sort | wc -l
144
mb@host:~$ xlsclients | sort | uniq -c
  1 host  baloo_file
  1 host  bamfdaemon
  1 host  Banshee
  1 host  blueman-applet
  1 host  cairo-dock
  1 host  chromium-browser
  1 host  dolphin
  1 host  eom
  1 host  evince
  1 host  gcalctool
  1 host  gkrellm
  1 host  gnome-terminal
  1 host  ibus-ui-gtk3
  1 host  ibus-x11
  1 host  kactivitymanagerd
  3 host  kded4
  1 host  'kdeinit4: kded4 [kdeinit]'
  1 host  'kdeinit4: ksmserver [kdeinit]'
  1 host  kglobalaccel
  1 host  klipper
  1 host  kmix
  1 host  knotify4
  1 host  konsole
  1 host  korgac
  1 host  krunner
  2 host  ksmserver
  1 host  kuiserver
  1 host  kwalletd
  6 host  kwin
  1 host  mate-screensaver
  1 host  nautilus
  5 host  okular
  1 host  pavucontrol
  1 host  Pidgin
  1 host  plasma-desktop
  1 host  pluma
  1 host  polkit-kde-authentication-agent-1
  1 host  soffice.bin
  1 host  thunderbird
  1 host  transmission-remote-gtk
  2 host  /usr/bin/baloo_file
  3 host  /usr/bin/dolphin
  1 host  /usr/bin/kactivitymanagerd
  2 host  /usr/bin/kglobalaccel
  3 host  /usr/bin/klipper
  8 host  /usr/bin/kmix
  8 host  /usr/bin/konsole
  2 host  /usr/bin/korgac
  3 host  /usr/bin/krunner
  2 host  /usr/bin/kuiserver
  2 host  /usr/bin/kwalletd
 13 host  /usr/bin/okular
 29 host  /usr/bin/plasma-desktop
  2 host /usr/lib/kde4/libexec/polkit-kde-authentication-agent-1
  2 host  /usr/lib/libreoffice/program/soffice.bin
  1 host  /usr/lib/virtualbox/VirtualBox
  2 host  /usr/lib/virtualbox/VirtualBox --comment shitxp 
--startvm 81289eb9-7de8-492c-9a4f-56977a2b8eca --no-startvm-errormsgbox

  1 host  vino-server
  2 host  VirtualBox
  1 host  vmware
  1 host  vmware-tray
  1 host  vmware-unity-helper
  1 ''  /usr/lib/libreoffice/program/soffice

I made a note on my desktop  (dry erase pen + glass tabletop == best 
whiteboard ever) to check that next time it freaks out.

I've seen reports of this, stating it's a hard-coded thing in xorg
code, which I find entirely asinine


It probably seemed like a reasonable assumption back when the X11 
protocol was designed that an X client would only make 1 connection 
to the server, and that having 256 or 512 X clients at once was 
enough.  I 

Re: xorg: Maximum number of clients reached

[Michael Butash]

  
  
I've been somewhat waiting for some
  sign from above that Wayland is actually a real, usable piece of
  software, as it's been hyped as "xorg, but doesn't suck".  My
  limiters are fglrx support (again, my damn 6-head ati card), and
  decent support for window environments, mostly kde I use on
  everything now.  
  
  At your mention, I looked at Wayland as it's been a while, and
  still seems like a tech demo more than a usable product.
  
http://blog.martin-graesslin.com/blog/2013/06/starting-a-full-kde-plasma-session-in-wayland/
http://linuxg.net/how-to-install-kde-plasma-5-on-kubuntu-14-04-kubuntu-14-10-and-linux-mint-17-kde/
  
  I did try the oss ati driver vs. fglrx, and sadly it was only
  seeing one monitor, wanting to only replicate the output across
  all 6.  Scratch.
  
  Ubuntu/compiz is just out of the question, I tried lxde, mint
  (cinnamon/mate), gnomeshell, and more or less hate them all vs.
  kde that remains the most usable/bug-free.  I might try that
  second link later and see how my mileage might vary, but I'm still
  stuck on using kde as the least sucky (usable) desktop out there
  atm.
  
  I did try lxde with marco as a compositor, but it was as buggy, or
  worse than compiz in dealing with my giant 11520x1200
  framebuffer.  So far Kwin is the least buggy at it, and I still
  can't spawn an opengl game without it flickering like an
  epilepsy-inducing apparatus by design as they all do.  
  
  Apparently no one considers someone might actually try and make 6
  displays work, aside from me of course, but then the rest of the x
  issues just compound it to make that the least of my worries.  I
  had designs to replace my 6x displays with 3x 4k res displays
  (11520x2160), but I need to see this will even work stable before
  bothering.
  
  -mb
  
  
  
  On 06/11/2015 08:09 PM, Stephen Partington wrote:


  Have you tried weyland yet to see if it is any
improvement? I think fedora has a build running weyland. (memory
is fuzzy on this one.)
  On Jun 11, 2015 5:20 PM, "Michael Butash"

wrote:
So this
  seems to be a big problem for me, in that it simply refuses to
  open new apps, and find this happens more and more these days.
  Am I like the only actual person to use linux these days that
  this occurs with?
  
  I've seen reports of this, stating it's a hard-coded thing in
  xorg code, which I find entirely asinine, but seems a reality
  when using Chrome/Chromium that launches some 300 flocks on
  various things, and blows out the 256/512 client count on
  xorg.  I find this almost stupid, and feel I'm back to the
  days of windoze me having to reboot every other day.
  
  Reality is I have 3 chrome profiles open, some pdfs,
  libreoffice, some chrome apps, some file manager windows
  (dolphin/kde), and not much else.  Sort of want to punch
  someone in the face when I see this - someone is obviously
  doing something wrong, and really can't see why.  All I can
  ever think is really, am I the only person that really "uses"
  a linux desktop to see these?
  
  Uninstalling pepperflash, or any flash vermin, seems to have
  done some good, no longer causing a persistent memory leak in
  xorg/fglrx drivers, but otherwise chrome|chromium still seems
  an absolute basketcase under linux, forgetting there are
  resource limits they should consider adhering to.
  
  -mb
  ---
  PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
  To subscribe, unsubscribe, or to change your mail settings:
  http://lists.phxlinux.org/mailman/listinfo/plug-discuss

  


  

---
PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
To subscribe, unsubscribe, or to change your mail settings:
http://lists.phxlinux.org/mailman/listinfo/plug-discuss

Re: xorg: Maximum number of clients reached

[Michael Butash]

Thanks for the comments Matt - in line.

On 06/12/2015 02:18 PM, Matt Graham wrote:

On 2015-06-11 17:20, Michael Butash wrote:

[X reaching a maximum number of clients is a problem] in that it
simply refuses to open new [X clients], and [I] find this happens
more and more these days. Am I like the only actual person to use
linux these days that this occurs with?


I've never seen this happen.  What do you get when this is happening 
and you do "xlsclients | sort | uniq -c" ?  I currently have 64 X 
clients running here.  Most KDE things show up as 2 or 3 X clients.  
plasma-desktop shows up as 20.  firefox shows up as 1.
I was actually trying to remember that command as I've seen it 
referenced and checked before, so thanks for that.


Every now and then problem will piss me off royally as literally my pc 
won't be able to wake my monitors up out of dpms sleep because of this 
(presumably), thus I have to hard reboot.  Odd part is when it freaks 
like this, even ssh'ing in from my laptop, killing some pids (ie. 
chrome|chromium I have an xargs script I keep to seek and destroy all), 
and it still won't wake then.  I need add something to watch these...


After it did this the other day upon composition of the email, I killed 
the chrom*'s, and it's been a bit stable.  xrestop was another that was 
recommended to watch, but by the time I hit that limit, even xrestop 
gives me the "max number of clients reached", even though a cli 
application...


Another way I know my system is almost ready to implode is waking up out 
of monitor sleep, expecting to see the simple-locker screen, rather I 
see a full desktop, unhidden, but I can't actually click on anything.  
Yeah, so much for privacy/security, but at least someone couldn't 
interact with it.  It requires me to ctrl-alt-F1, switch to a tty, and 
back to F7 to see the locker again, unlock, and actually use my 
desktop.  Seem another byproduct of xorg freaking out.


At the (working) moment...

mb@host:~$ xlsclients | sort | wc -l
144
mb@host:~$ xlsclients | sort | uniq -c
  1 host  baloo_file
  1 host  bamfdaemon
  1 host  Banshee
  1 host  blueman-applet
  1 host  cairo-dock
  1 host  chromium-browser
  1 host  dolphin
  1 host  eom
  1 host  evince
  1 host  gcalctool
  1 host  gkrellm
  1 host  gnome-terminal
  1 host  ibus-ui-gtk3
  1 host  ibus-x11
  1 host  kactivitymanagerd
  3 host  kded4
  1 host  'kdeinit4: kded4 [kdeinit]'
  1 host  'kdeinit4: ksmserver [kdeinit]'
  1 host  kglobalaccel
  1 host  klipper
  1 host  kmix
  1 host  knotify4
  1 host  konsole
  1 host  korgac
  1 host  krunner
  2 host  ksmserver
  1 host  kuiserver
  1 host  kwalletd
  6 host  kwin
  1 host  mate-screensaver
  1 host  nautilus
  5 host  okular
  1 host  pavucontrol
  1 host  Pidgin
  1 host  plasma-desktop
  1 host  pluma
  1 host  polkit-kde-authentication-agent-1
  1 host  soffice.bin
  1 host  thunderbird
  1 host  transmission-remote-gtk
  2 host  /usr/bin/baloo_file
  3 host  /usr/bin/dolphin
  1 host  /usr/bin/kactivitymanagerd
  2 host  /usr/bin/kglobalaccel
  3 host  /usr/bin/klipper
  8 host  /usr/bin/kmix
  8 host  /usr/bin/konsole
  2 host  /usr/bin/korgac
  3 host  /usr/bin/krunner
  2 host  /usr/bin/kuiserver
  2 host  /usr/bin/kwalletd
 13 host  /usr/bin/okular
 29 host  /usr/bin/plasma-desktop
  2 host /usr/lib/kde4/libexec/polkit-kde-authentication-agent-1
  2 host  /usr/lib/libreoffice/program/soffice.bin
  1 host  /usr/lib/virtualbox/VirtualBox
  2 host  /usr/lib/virtualbox/VirtualBox --comment shitxp --startvm 
81289eb9-7de8-492c-9a4f-56977a2b8eca --no-startvm-errormsgbox

  1 host  vino-server
  2 host  VirtualBox
  1 host  vmware
  1 host  vmware-tray
  1 host  vmware-unity-helper
  1 ''  /usr/lib/libreoffice/program/soffice

I made a note on my desktop  (dry erase pen + glass tabletop == best 
whiteboard ever) to check that next time it freaks out.

I've seen reports of this, stating it's a hard-coded thing in xorg
code, which I find entirely asinine


It probably seemed like a reasonable assumption back when the X11 
protocol was designed that an X client would only make 1 connection to 
the server, and that having 256 or 512 X clients at once was enough.  
I don't have the Xorg source here so can't find where this is set, 
either.

https://unix.stackexchange.com/questions/25273/what-is-the-max-number-of-x-clients



Chrome/Chromium that launches some 300 flocks on various things, and
blows out the 256/512 client count on xorg.


What did you mean by "flocks"?  If Chrome creates a separate X client 
for every browser tab, that'd probably cause stupidity, but I could 
see it doing that.  (Having fewer than 30 tabs open at any given time 
would fix that if it were the case.)
File locks, or rather

Re: xorg: Maximum number of clients reached

[Matt Graham]

On 2015-06-11 17:20, Michael Butash wrote:

[X reaching a maximum number of clients is a problem] in that it
simply refuses to open new [X clients], and [I] find this happens
more and more these days. Am I like the only actual person to use
linux these days that this occurs with?


I've never seen this happen.  What do you get when this is happening 
and you do "xlsclients | sort | uniq -c" ?  I currently have 64 X 
clients running here.  Most KDE things show up as 2 or 3 X clients.  
plasma-desktop shows up as 20.  firefox shows up as 1.



I've seen reports of this, stating it's a hard-coded thing in xorg
code, which I find entirely asinine


It probably seemed like a reasonable assumption back when the X11 
protocol was designed that an X client would only make 1 connection to 
the server, and that having 256 or 512 X clients at once was enough.  I 
don't have the Xorg source here so can't find where this is set, either.



Chrome/Chromium that launches some 300 flocks on various things, and
blows out the 256/512 client count on xorg.


What did you mean by "flocks"?  If Chrome creates a separate X client 
for every browser tab, that'd probably cause stupidity, but I could see 
it doing that.  (Having fewer than 30 tabs open at any given time would 
fix that if it were the case.)



I have 3 chrome profiles open, some pdfs, libreoffice,
some chrome apps, some file manager windows (dolphin/kde), and not
much else. [...]  am I the only person that really "uses" a linux
desktop to see these?


Obviously not if you found some other people complaining on a search 
engine.  First thing to do is figure out which program is causing the 
stupidity.  I was surprised to see 20 plasma-desktop clients here, 
because plasma applets are useless and I didn't think I had any of them 
running at all.


--
Crow202 Blog: http://crow202.org/wordpress
There is no Darkness in Eternity
But only Light too dim for us to see.
---
PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
To subscribe, unsubscribe, or to change your mail settings:
http://lists.phxlinux.org/mailman/listinfo/plug-discuss

Re: How RedHat Backports Vulnerability Fixes

[Michael Butash]

On 06/12/2015 11:12 AM, Keith Smith wrote:

On 2015-06-12 10:43, der.hans wrote:

Am 12. Jun, 2015 schwätzte Keith Smith so:

I do some work on a couple CentOS 6.6 servers. Payment Card Industry 
(PCI) scans seem to always see the server as vulnerable. I've have 
to submit for a review since the server is not really vulnerable.


Your auditors should understand that and be able to do proper 
verification.



You would think.


I used to think so too, before having to deal with various qsa's 
throughout the years.  Most I find to be lacking, in either real or 
practical knowledge, especially when it comes to more nebulous things 
like networks and how they play into security.  Seemingly nothing more 
than glorified tech writers pushing some automagical "scan and make 
report go" button.


Case in point, I had one tell me that trunking/802.1q was "insecure" 
(requiring huge changes from "normal" physical deployment a sane network 
guy might deploy), but hey, my MPLS network, also using dot1q, was just 
dandy.  Mostly because they didn't know what mpls presumably even did, 
which was even more extensive logical separation than even dot1q, and 
just as prone to abuse/misconfiguration should someone bleed routes 
between tables of organizations in a service provider network accidentally.


Same one also just glossed over the 50-60k firewall rules we had 
involved, more just happy we simply had one, with or without an explicit 
permit any.


Of course, inherently insecure applications or systems can always have 
"mitigating controls" documented that in my experiences equals sleight 
of hand, putting some voodoo appliance in front of it they know even 
less about, or host security software that has McAfee or Symmantec in 
the name, but as long as it's called a *security* something, it makes it 
quite ok suddenly.


Target, Home Depot, and all the others you never hear about being 
exploited for your pci/pii data are good examples of how useless the 
certification really is, other than as another profit center for firms 
selling the audit services.


-mb
---
PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
To subscribe, unsubscribe, or to change your mail settings:
http://lists.phxlinux.org/mailman/listinfo/plug-discuss

Re: How RedHat Backports Vulnerability Fixes

[der.hans]

Am 12. Jun, 2015 schwätzte Keith Smith so:


On 2015-06-12 10:43, der.hans wrote:

Am 12. Jun, 2015 schwätzte Keith Smith so:

I do some work on a couple CentOS 6.6 servers. Payment Card Industry (PCI) 
scans seem to always see the server as vulnerable. I've have to submit for 
a review since the server is not really vulnerable.


Your auditors should understand that and be able to do proper verification.



You would think.


If they don't they get gig reports from me :).

ciao,

der.hans
--
#  http://www.LuftHans.com/http://www.PhxLinux.org/
#  "Lie detector eyeglasses perfected: Civilization collapses."
#   -- Richard Powers---
PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
To subscribe, unsubscribe, or to change your mail settings:
http://lists.phxlinux.org/mailman/listinfo/plug-discuss

Re: How RedHat Backports Vulnerability Fixes

[Keith Smith]

On 2015-06-12 10:43, der.hans wrote:

Am 12. Jun, 2015 schwätzte Keith Smith so:

I do some work on a couple CentOS 6.6 servers. Payment Card Industry 
(PCI) scans seem to always see the server as vulnerable. I've have to 
submit for a review since the server is not really vulnerable.


Your auditors should understand that and be able to do proper 
verification.



You would think.




I don't think a lot of people understand how RHEL maintains it's 
packages. I know I did not for a long time.  RedHat backports 
vulnerability fixes while maintaining the original version number.


Here is a great explanation : 
https://access.redhat.com/security/updates/backporting/?sc_cid=3093


Thanks for the link! I've mostly understood it, but it's good to have a
handy official reference to point people at.

ciao,

der.hans
---
PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
To subscribe, unsubscribe, or to change your mail settings:
http://lists.phxlinux.org/mailman/listinfo/plug-discuss


--
Keith Smith
---
PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
To subscribe, unsubscribe, or to change your mail settings:
http://lists.phxlinux.org/mailman/listinfo/plug-discuss

Re: How RedHat Backports Vulnerability Fixes

[der.hans]

Am 12. Jun, 2015 schwätzte Keith Smith so:

I do some work on a couple CentOS 6.6 servers. Payment Card Industry (PCI) 
scans seem to always see the server as vulnerable. I've have to submit for a 
review since the server is not really vulnerable.


Your auditors should understand that and be able to do proper verification.

I don't think a lot of people understand how RHEL maintains it's packages. I 
know I did not for a long time.  RedHat backports vulnerability fixes while 
maintaining the original version number.


Here is a great explanation : 
https://access.redhat.com/security/updates/backporting/?sc_cid=3093


Thanks for the link! I've mostly understood it, but it's good to have a
handy official reference to point people at.

ciao,

der.hans
--
#  http://www.LuftHans.com/http://www.PhxLinux.org/
#  The Internet is the front line of the battle
#  to protect our freedom. -- Nathaniel Borenstein---
PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
To subscribe, unsubscribe, or to change your mail settings:
http://lists.phxlinux.org/mailman/listinfo/plug-discuss

Re: How RedHat Backports Vulnerability Fixes

[der.hans]

Am 12. Jun, 2015 schwätzte Keith Smith so:

I do some work on a couple CentOS 6.6 servers. Payment Card Industry (PCI) 
scans seem to always see the server as vulnerable. I've have to submit for a 
review since the server is not really vulnerable.


Your auditors should understand that and be able to do proper verification.

I don't think a lot of people understand how RHEL maintains it's packages. I 
know I did not for a long time.  RedHat backports vulnerability fixes while 
maintaining the original version number.


Here is a great explanation : 
https://access.redhat.com/security/updates/backporting/?sc_cid=3093


Thanks for the link! I've mostly understood it, but it's good to have a
handy official reference to point people at.

ciao,

der.hans
--
#  http://www.LuftHans.com/http://www.PhxLinux.org/
#  The Internet is the front line of the battle
#  to protect our freedom. -- Nathaniel Borenstein---
PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
To subscribe, unsubscribe, or to change your mail settings:
http://lists.phxlinux.org/mailman/listinfo/plug-discuss

Re: How RedHat Backports Vulnerability Fixes

[der.hans]

Am 12. Jun, 2015 schwätzte Keith Smith so:

I do some work on a couple CentOS 6.6 servers. Payment Card Industry (PCI) 
scans seem to always see the server as vulnerable. I've have to submit for a 
review since the server is not really vulnerable.


Your auditors should understand that and be able to do proper verification.

I don't think a lot of people understand how RHEL maintains it's packages. I 
know I did not for a long time.  RedHat backports vulnerability fixes while 
maintaining the original version number.


Here is a great explanation : 
https://access.redhat.com/security/updates/backporting/?sc_cid=3093


Thanks for the link! I've mostly understood it, but it's good to have a
handy official reference to point people at.

ciao,

der.hans
--
#  http://www.LuftHans.com/http://www.PhxLinux.org/
#  The Internet is the front line of the battle
#  to protect our freedom. -- Nathaniel Borenstein---
PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
To subscribe, unsubscribe, or to change your mail settings:
http://lists.phxlinux.org/mailman/listinfo/plug-discuss

How RedHat Backports Vulnerability Fixes

[Keith Smith]

I do some work on a couple CentOS 6.6 servers. Payment Card Industry 
(PCI) scans seem to always see the server as vulnerable. I've have to 
submit for a review since the server is not really vulnerable.


I don't think a lot of people understand how RHEL maintains it's 
packages. I know I did not for a long time.  RedHat backports 
vulnerability fixes while maintaining the original version number.


Here is a great explanation :  
https://access.redhat.com/security/updates/backporting/?sc_cid=3093


Keith

---
PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org
To subscribe, unsubscribe, or to change your mail settings:
http://lists.phxlinux.org/mailman/listinfo/plug-discuss