Hello,
It's been a long time since I've posted to this group. I want to thank
you for all the work
that you have done.
Here is the background information. We have been using LTSP for several
years. We run DHCP
and TFTP on Windows 2003 servers. We were using a Linux server (Suse 8.2 on an
older IBM server)
to serve the NFS image for the LTSP clients. The LTSP clients lauch the ICA
client locally to log
into Citrix. We have around 150 old PC's with Intel NICs that boot to LTSP
using PXE. Our
network infrastructure is built on Cisco hardware. We recently upgraded the
firmware on our main
Router/Switch.
Last week we heard complaints from the users that they were getting the
NFS cannot mount
error. When we went to look at the server, we could move the mouse, but we
could not log into the
desktop, neither could we switch to a console view. We powered off the server
to try to reboot.
We were unable to boot the server because of a BIOS error about the MBR being
changed followed by
a dos style message saying that it could not access the MBR. I think that the
hard drive may be
going out.
Around this time we realized that our users could not access anything on
the same subnet as
our LTSP server. We could not ping servers on that subnet. Clients and
servers on another subnet
continued to work with no apparent problems. Two switches that have PCs on the
same subnet as the
LTSP server started to flash all the network activity LEDs in unison very
rapidly. The only way
we were able to calm the network down to some degree was to restart the
router/switch that the
LTSP server had been plugged into. Even then, we seemed to have much more
network traffic then we
should have (ping response rates of over 60ms over a LAN to the main ip of the
router/switch).
When we were able to boot our LTSP server the network froze up again
within a couple
minutes. I had a continuous ping up trying to ping the LTSP server and not one
ping got through,
even though the server showed that it had brought up its network interfaces.
When we saw that the
subnet that the LTSP server is on was unresponsive again, we disconnected the
network cable from
the LTSP server and reset our router/switch. Again we regained the ability to
communicate on that
subnet.
We did a lot more troubleshooting, but we still don't understand how this
could have
happened. We brought up a different Linux server to serve NFS, but as soon as
we connected it to
the network with the same IP address of the original LTSP server (which was
still disconnected),
the network again seemed to be flooded with network traffic. Now we are
running NFS on a Windows
server using Services For Unix 3.5. The traffic rates on the LTSP subnet seem
to be either normal
or very close to normal, but we are afraid that we might see a similar problem
in the future. Has
anyone seen anything similar to this?
The suspects in our eyes are NFS communication over UDP or the recent
upgrade to the Cisco
router/switch or a combination of the two. The upgrade happened 3 days before
we saw this
problem. Has anyone seen any problems in the interaction between Cisco and NFS
or LTSP? Does
anyone have any ideas? Any help or suggestions would be greatly appreciated.
Abraham Pearson
__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_____________________________________________________________________
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help, try #ltsp channel on irc.freenode.net
