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

Re: Periodical server problems with sarge server.



Hi,

On Mon, 13 Jun 2005, RalfGesellensetter wrote:

> Every morning, and sometimes also in the afternoon if the server had 
> some leisure during the morning, users fail to login via samba from 
> windows clients. The message is something like 
> "domain SKOLELINUX not available".

Do you think it's the networking that's going down or just samba?  Are
there thin clients (or any other network service) in use?  If so, are they
affected?

> Alas, this only happens, when no capable admin is present (neither me)
> reboot the server. 

Always the way, isn't it?

What would be nice would be to give him a checklist of tests for next time
it happens.  If he's busy he might just do one or two tests and then solve
the problem by rebooting.  If he has a few minutes, he could run a few more
and over time we'll get a better picture of the problem.  Here are some
questions which would be worth getting answers to.

-- Simple Network issues --
1. Can the server ping the client?
2. Can the client ping the server? (cancel at login screen to try this)
3. Can the client web browser connect to http://tjener.intern/ ?
4. Can the client web browser connect to http://10.0.2.2/ ?
5. Is the web browser using a proxy server?

-- DHCP Problems --
6. Run command window in windows and type "ipconfig /all" and see does the
   output include a 10.0.[23].X or 192.168.0.X IP Address.

-- Samba Issues --
7. Is the problem fixed by 
	/etc/init.d/samba restart
8. Is the problem fixed by 
	/etc/init.d/samba stop; killall -9 nmbd; /etc/init.d/samba start
9. Are *all* machines unable to login or just one or two?

> I don't feel good with putting a "reboot" into crontab - but have to 
> keep the local admin in a good mood if you know what I mean.

If the problem is some program or driver crashing whenever it hits a
certain condition then restarting it may not help.

> What we did as possible workaround was creating a cronjob that restarts 
> all daemons in questions in the early morning (and kill nmdb zombie 
> which remains after stopping samba). This is running - but yet the 
> problem lasts on. 

Could you get him/her to run this script instead of rebooting?  That would be
quicker for him and would give you a better idea of whether it solves the
problem.  If the network is going down due to a drievr issue then

  /etc/init.d/nfs-kernel-server stop
  /etc/init.d/network stop
  rmmod [your_network_driver]
  modprobe [your_network_driver]
  /etc/init.d/network start
  /etc/init.d/nfs-kernel-server start

might give you an idea.

Gavin



Reply to: