Submitted by Anonymous on Fri, 06/07/2002 - 12:57. General Questions
I have a strange problem that seems to be related to upgrading from Intel RH7.0 to 7.2 and using X(v4.0, no upgrade) . I cannot start alternate X sessions manually to the same display at different login times.
Allow me to explain further....
I start X sessions to connect to other machines with the standard call
"X (display #) -ac -query (hostname)
While using RH 7.0., once I stop the above session (by logging out and hitting ), I was able to start another
session using the same display number. Now, using RH7.2, I can no longer use the same display number. I get the standard error that tells me that it cannot get a listener to connect to that socket because its already in use. I then went to the "/tmp" directory and sure enough there is a socket file with that display # there. But, I cannot delete it even as "root".
I sure would like to get this resolved soon as I am up to display 21 already and my "/tmp" directory is getting filled with stale socket files.
Hi,
?? are the X servers still running [defunct] ??
I would think the sockets would close when the parent app does. That`s weird. I wonder if you don`t have 20 virtual X displays open (see "man xnest").
Anyway, is killing the (gdm or xdm) on the server honoring the queries an option? It`s not a nice option but it might close the sockets.
I must also mention that a new X server fails to start with the "old" DISPLAY number across re-bootings. In other words, once I use "X 2 -ac -query ", I have been no longer able to connect to DISPLAY 2 now for months. So, somewhere, either the old links in "/tmp" are not being cleared or there is a file somewhere that is being filled that is not being reset.
Any new insights will be greatly appreciated.
Steven Orton