• Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 537
  • Last Modified:

Auto disconnect

When I telneting to the Solaris 2.6 server and idle for few minutes, the server automatically disconnect me from the server. why ? How am I disable this setting ?
0
genius
Asked:
genius
  • 14
  • 10
  • 5
  • +2
1 Solution
 
jlevieCommented:
If the Solaris system is using standard network daemons and shells, it's likely that something else is causing the timeout as a stock system doesn't do idle timeouts. Is the Solaris system on your local network or is there a router and/or firewall in between? I you are dialing in to the Solaris system it could be your system dropping the line when idle, or it could be a terminal server or ppp daemon at the remote end.
It's hard to say without knowing more.
0
 
geniusAuthor Commented:
Feeback to jlevie:
1) Yes I am using standard Sun Solaris 2.6 network daemons and shell.

2) I'm using Windows 95 telnet utility to telnet into Sun Solaris server. Some how I leave the telnet session idle for few minutes, my connection to the server was lost. No error message appear on the screen.

3) My server is on the local network and there is no firewall install in my office. We are using Linux as the router for connecting to internet.

My question is :
1) Do there have any setting to enble/disable the timeout for idle login connection ?

Thanks
0
 
geniusAuthor Commented:
Adjusted points from 10 to 20
0
Restore individual SQL databases with ease

Veeam Explorer for Microsoft SQL Server delivers an easy-to-use, wizard-driven interface for restoring your databases from a backup. No expert SQL background required. Web interface provides a complete view of all available SQL databases to simplify the recovery of lost database

 
jlevieCommented:
Which shell do you use on Solaris? Is it using the Sun shell init scripts or locally defined scripts? Do you have any other systems on the local lan that you can open a telnet session to from the win95 box to see if that session times out also?
0
 
geniusAuthor Commented:
I am using bourne shell.
Yes, it is using Sun shell init scripts. Not runnig any local defined scripts. Yes, I have tried telneting to another sun server, it is ok. No timeout. Why ? May I know whetehr there is any parameter to enable/disable this setting ? Please let me know if you found it . Thanks
0
 
geniusAuthor Commented:
Adjusted points from 20 to 30
0
 
jlevieCommented:
The most likely culprit is that someone has changed /etc/default/login to enable idle timeouts. By default that option is commented out, but if set it will force a disconnect after the specified time.

If that's not it, then it can't be a stock Solaris box. Something else has been added or changed to cause the idle timeout. None of the other items involved (sh, telnetd, or rlogind) have provisions for idle timeouts.
0
 
geniusAuthor Commented:
I have check the /etc/default/login file and TIMEOUT setting is commented. Do you think is the Sun Solaris bugs ? Any recommended patches need to apply for ?
0
 
jlevieCommented:
In that case I suspect that an "idle time out" daemon has been installed. I'll look and see if there's anything in the patch database that addresses this, but I'm pretty sure that there won't be.
0
 
geniusAuthor Commented:
Adjusted points from 30 to 50
0
 
geniusAuthor Commented:
Jlevie, are you still there to help me ?Basically it is a normal Solaris 2.6 installation. Nothing else have been added into the server as a daemon. Just the default 2.6 installation. Please help if anyone know the answer.
0
 
jlevieCommented:
Sorry, I've had a bit of a problem with email over the last several days and couldn't see the notifications.

I didn't find anything in the patch database that would explain what you are seeing, nor have I ever experienced this with any Solaris system. So we are justing to have to figure out what is happening.

Do you see the same behaviour when telneting in to that system from any other system? Do any other user accounts experience the problem when telneting in to that system?

What kind of network do you have (hubs or switches) and what hardware platform is the 2.6 system (e.g. SparcStation 4, Ultra 5, etc.)?

Can I see the output of "ps -ef" when you telnet into the system? Also I'd like to the output of "echo $SHELL".
0
 
geniusAuthor Commented:
All the user experience the same thing as I do.
My server is Sun Enterprise Ultra 10S.
I'm afraid the output of ps -ef can't fit well in this column. Do you have a email acocunt so that I can send it to you.
The output of echo $SHELL is /sbin/sh
0
 
jlevieCommented:
Pease send the output to jlevie@bellsouth.net.

Knowing that it's an Ultra 10S is a help as is knowing that others have the same problem. For grins what are the other Solaris boxes that don't exhibit the problem.

Also I wouldn't mind seeing what "uname -a" displays on the faulty system and one of the well behaved systems.
0
 
mfatihCommented:
There may be an environment variable like AUTOLOGOUT . Set it to zero.
0
 
ishCommented:
In Solaris, there isn't ANY setting to enble/disable the timeout for idle login connections.  Unless you install a third party daemon...

Some questions to get us in the right direction?

How often do the logouts happen?

Do they ALL happen after the same amount of inactivity?
    If So, then how much?  (a regular interval might point to a proccess, an irregular one, to an instability).

You say that other users on other systems see the same?

Try this out for starts:
Let it happen, note the time.... log back in and check the /var/log/syslog    for errors at that time.

0
 
jlevieCommented:
Sorry for the delay in getting back. I've examined the ps out output and there's not anything that is obviously an idle-logout timer. Hovever, the system is far from being a stock Solaris box. There are things being run out of /usr/lib/security that I've never seen before and don't have a clue as to what they might do. Oracle, Apache and the Netscape server have all been installed on the system and there's also something running out of /opt/nas21. So while it's true that a stock Solaris box doesn't implement any idle-logout times, there's no telling what some of the other things that are running might be doing. You'd pretty much have to consult the documentation for the other stuff to see what it might be doing. I'd start by checking what the stuff running out of /usr/lib/security does, then find out what the stuff running out of /opt/nas21 does. Apache is okay as is Oracle and I think that the Netscape server should be okay also.
0
 
geniusAuthor Commented:
This server is running quite a number of applications such as Oracle enterprise server, Netscape Application server, Netscape web Server, Apache Web Server and Apache Jserv.It cannot be these application causing the probelm ! I agree with you that the stuff in /usr/lib/security/cryptorand is strange and I suspect this will the culrpit of the problem.
0
 
jlevieCommented:
That probably explains what the stuff in /opt/nas21 is (Netscape Application Server). As noted, the other server apps aren't known for introducing an idle-logout timer. So the stuff running out os /usr/lib/security (or something associated with its installation) is the most likely culprit.
0
 
ishCommented:
I'd agree, I've got Solaris boxes running 99% of that stuff... (the only one I don't have is the Apache Jserv.

None of my boxes run any Idle-Log Out stuff, nor do I remember any of the above named servers mentioning a config for such.

My too sense...
0
 
geniusAuthor Commented:
I have 2 server runnig the similar software that I have mentioned. Nothing wrong with the Jserv nor the NAS. I have kill the process related to /usr/lib/security but the problem still happen ! I can't let you guys telneting into my server cause it is our development server which only connected within the LAN.
0
 
ishCommented:
I fear I must bow out of this discussion... I've given everything I can and amdrawinga blank... I'll keep an eye on it, cause I'm interested in the resolution...
0
 
GeddyCommented:
Are the two similar boxes that we're talking about here on the same
subnet? If one box is on one subnet and the other one is on another subnet, the piece of network gear between the one still could be the problem. It's sounding to me like a network problem.
0
 
GeddyCommented:
Also, is there an error message accompanying the timeout, or does it just drop the connection. I'm assuming you're using telnet to talk to these boxes.
0
 
geniusAuthor Commented:
Yeap, there are 2 similar box and same stuff in there. One is external and the other is internal. It should be  difference subnet.
Yeap, no message, just drop the connection but the user still in the list where you check from w command. Anybody else have any idea ?
0
 
GeddyCommented:
So the ARE on two different subnets? One is an external (DMZ maybe?) box and one is internal?
0
 
GeddyCommented:
This smells like a network problem. I'd check network gear that each box connects to, and look for a timeout setting on the switches/routers that each respective box may be connected to.
0
 
ishCommented:
Has this question died?  I was hoping to read a final solution,lots of good input.  genius where do you still stand?
0
 
jlevieCommented:
Well, I don't know if he's still there or not, but I don't think that this is a software problem on that server. I've checked everything that he has installed and running and none of those things implement a idle timeout nor does it appear that there is an idle timeout process runing. I had asked, many moons ago, if the client and the server were on the same subnet and was told that they were. By the last comment, on 6 Jul, it now appears that there may by a router/firewall between the two. If so that is likely to be the cause of the problem, especially if the router/firewall is doing dynamic NAT or NPAT and configured with an idle timer on the translations (that's the default on Cisco gear).
0
 
ishCommented:
That was my feeling too, looks like Geddy was on the same track.

It just didn't make sense, and the errors and descriptions he was posting weren't always clear on that.
0
 
geniusAuthor Commented:
Well, I will try my best to let you guys have all the right info you want. Unfortunately I'm not good in networking stuff. Probably it is relating to network issue. Therefore I will like close this issue. The final thing is who should get the point ? Since jlevie is the first who responded to my question, therefore you got it. Thanks
0
 
geniusAuthor Commented:
Comment accepted as answer
0
 
geniusAuthor Commented:
Very helpful and have good knowledge on the unix especially Solaris.
0
 
jlevieCommented:
And thank you for your most generous comment.
0

Featured Post

Concerto Cloud for Software Providers & ISVs

Can Concerto Cloud Services help you focus on evolving your application offerings, while delivering the best cloud experience to your customers? From DevOps to revenue models and customer support, the answer is yes!

Learn how Concerto can help you.

  • 14
  • 10
  • 5
  • +2
Tackle projects and never again get stuck behind a technical roadblock.
Join Now