[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: [K12OSN] users kicked out on 4.2.1

Hash: SHA1

> Julius Szelagiewicz wrote:
>> On Thu, 14 Apr 2005, Eric Harrison wrote:
>>>Julius Szelagiewicz wrote:
>>>>Apr 13 10:12:45 solinden gdm[31911]: Ping to ws078.turtle.com:0 failed,
>>>>whacking display!
>>>>The workstation where user jszoke was forcibly logged out is ws078.
>>>There is the guilty party right there. If the server cannot ping a
>>>terminal, after a period of time it assumes the terminal is dead and
>>>kills off the session. The default for this is 15 seconds.
>>>You can change this value in /etc/X11/gdm/gdm.conf:
>> This forces me to ask some questions: 1. why it can not be pinged - when
>> I
>> ping it it is always fine. 2. why when this happens -nobody- can log in.
>> Of course the answer to those question is "network problem", but I had
>> no
>> changes to the network from before the upgrade. julius
> At least it helps narrow down the problem set.
> The first thing that jumps to my mind is an ethernet driver problem. Or
> maybe the RJ45 cable got kinked as you were moving the system around.
> Etc, etc.
> I'd run dmesg, ifconfig, and scan /var/log/messages looking for anything
> related to the ethernet. If it is connected into a managed switch, I'd
> check the logs there & stats on the port. Etc, etc.
> I actually had this happen to me about an hour ago! I have an old junk
> Netgear hub shoved under my desk that started spontaneously rebooting. I
> fixed mine by plugging the terminal directly into the wall ;-)
> -Eric

I upped the ping interval to 30 seconds, to no avail. one station
spontaneously lost display twice, but over the course of the week it
happened to all the stations. Maybe switches - relatively new HP Procurve,
limited management capability. I am now running continous ping to the
station that had most problems. I wonder if I'll see any packet loss. My
gut feeling (and I do have a considerable gut) is that this is -not- a
hardware issue. I'll concentrate on watching one station and I'll get all
the data about it from the /var/log/messages at the end of the day. Maybe
it will point us to something. The ping now running will tell us if there
is a real networking issue, or is gdm imagining something.
thanks, julius

Version: GnuPG v1.2.3 (GNU/Linux)


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]