Through many years and through CentOS-6.5, I would boot to runlevel 3,
do various admin stuff on my desktop and then telinit 5, which started
an X session on tty7. I would then do logout->switch user a few times to
get additional user IDs active. Many years ago I would accomplish this
with some kind of "startx ... :1', or ':2' etc. thingy, but I've
forgotten the details now.
For some reason 6.6 update makes it start on tty1. The wisdom behind
that is unknown to me. I thought I had that set back to the way it was
after I found and modified some script/conf file stuff, but no joy.
Might have been fixed but was masked because I didn't realize the
signifcance of the run level transition described below?
Anyway, I changed to boot to run level 5, start a phony session (i.e.
login with an I.D. that is resevred to do nothing at all), do
logout->switch user, which starts a session on tty7 where I login, again
for tty8 and again for tty9 which gets me operating as is my habit with
three different user-IDs on ttys 7-9.
The problem is that if I logout all those sessions, telinit 3, do my
normal backups, telinit 5, do the phony login and try to switch user I
get X aborts. What happens is the screen on tty1 blanks, I hit <ctrl>,
screen comes alive and asks for password, I enter it and I can be active
on that phony session again. But I can't do any switch user at all
because it again aborts.
I do 'shutdown -r now' and that cures the problem when it comes back in
run level 5.
As part of trying to correct, I wiped the system which had been updated
from 6.5->6.6, downloaded the install DVDs and re-installed the system,
including my changes from over the years (I need to do some learning on
the RPM scripting stuff if I have to do this very often - but I have
successfully transitioned to TDU, Typical Dumb User, since I got out of
computers professionally).
The system is pretty-much "box stock" CentOS 6.6, but for adding gpm,
some KDE components, and plugins like flash player and "official" java
for Firefox and such.
$ uname -a
Linux CentOS501.homegroannetwork 2.6.32-504.1.3.el6.x86_64 #1 SMP Tue
Nov 11 17:57:25 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux
$ cat /etc/centos-release
CentOS release 6.6 (Final)
System is 100% current AFAIK.
I tried to use the provided abort tool to report the X abort issues, but
I have no RH account.
Anybody else experience the X behavior I describe? The key seems, at my
ignorant level anyway, to be transitioning from run level 3 to run level
5. Or maybe 5 to 3 and back to 5? But early on I tried booting to 3,
telinit 5 and had similar symptioms IIRC (age and time fades the memory
and I didn't start out thinking this was a reportable thing). Booting
directly into run level 5 hasn't, so far, demonstrated the symptoms.
I figure when I get time I will duplicate the issue and open an item in
the CentOS bug reporting tool. I suppose I'll have to put the collected
trace stuff in a file, which I can do, and then upload from there.
I did a quick scan of X-related stuff in the bug tracker and didn't see
anything similar.
1. Can anybody else confirm this behavior?
2. Any known work-around?
3. While we're at it, .conf changes or such to get first X session back
to tty7?
This is the most difficult time with an upgrade for me since I started
with CentOS ... 3? 4?
TIA for any help/feedback.
Bill
_______________________________________________
CentOS mailing list
CentOS@xxxxxxxxxx
http://lists.centos.org/mailman/listinfo/centos
