Hi,
I have a problem regarding stability of UML clients while having a high
activity on it and need help to solve this.
I set up 3 UML clients with official IP addresses with tty logging and
mounted hppfs, each connected via its own tap device. Kernel is 2.4.24
for the clients and 2.6.2 for the host, which is a P4 3GHz with 1GB of
RAM. The tap devices and the outgoing interface are bridged to a
transparent bridge.
This works really great, I have captured my first uml tty logs not
created by myself if you know what I mean..., and had fun watching all
sorts of logfiles. Until someone used a command that produced a high
load, with the result that I had to reboot the machine the next day when
I saw it.
I don't know the command the user used, but I guessed it was something
with a lot output or high load. So I tested with running tcpdump, and
after some seconds, the whole host crashed. On another system (similar
setup but only one UML client, equal host machine) with a spam mail trap
I had the same problem after a spammer relayed 200000 spam mails and I
typed 'mailq', but here just UML crashed (with segfault I think).
I'm not a programmer and i have no clue how to debug this problem, so I
would appreciate help on the following things:
- Can I increase stability by tuning something on the host or the
client?
- How can I help finding this problem if it is not solved so far?
The host has no general stability problem, just for the case someone
asks.
If you need more information, don't hesitate to ask.
Thanks in advance, and thanks for the great work so far.
Regards,
Sascha
-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click
|
