Nix wrote:
> On Wed, 30 Mar 2005, Roy Schestowitz stipulated:
>>> KDE has all sorts of complicated caching which may be throwing you off:
>>> a look at the KDE log (emitted on the X server's stdout, normally) is
>>> probably necessary. Alternatively, try to start the KDE app from a
>>> shell prompt: the errors (if any) will be emitted to stdout there.
>>
>> Yes, that's what I soon realised. Some error message came up at first
>> from MATLAB, then Java refused to open windows and crashed with huge
>> error log dumps. Later on, I got a message to the effect of "maximum X
>> sessions reached".
>
> The exact error message would be awfully useful.
>
> I vaguely recall that there is a hard upper limit on the number of
> *clients* that may be attached to a given X server, but that's quite a
> different thing to the number of concurrent X servers (my reading of `X
> sessions'; X doesn't really have the concept of a `session' so I had to
> guess what you meant).
I am not comfortable with that terminology of X, so I'm sorry if I misled
you. Yes... indeed I had many clients attached to the local X server. It
wasn't the number of windows being managed. I got plenty of meaningless
error messages and windows refused to open. I couldn't afford to reboot for
days because of the remote connections.
I am glad to have discovered that you heard of a "hard upper limit on the
number of *clients* that may be attached to a given X server".
Roy
--
Roy S. Schestowitz
http://Schestowitz.com
|
|
