|
See my answer in blue/italic
On Tue, 24 Oct 2006, Grames, Gernot wrote:
> I have configured a LVS/NAT with mon on 2 directors (Major and
Backup).
what is mon monitoring on the director? (I've only used mon
on the realservers to monitor services.)
The mon is running on the director and monitors the http service on each
realserver.
So it starts the http.monitor.
> Behind the director i have 4 clients running.
4 realservers?
Yes 4 realservers -> On each a webserver (jboss) is running.
> The lvs with the nat and the mon is working perfectly for some time.
> After around 5 or 6 days (i have it seen also one time after one day)
> the server (director) freezes completly.
keyboard at the console does nothing?
Nothing --> STRG+ALT+DEL is also not working
> I made some scripts (cron) to write logfiles with the info of the
> running process, memory and hard disc space.
> But there i cannot find a special behavior.
usually when something dies hard, there's nothing to write
to log files.
Yes, but i thought that i can see maybe a increased memory or some
Zombie processes, or something else!
> I can`t also say if the ipvs is the problem or the mon??
>
> Is there a possibility to debug these things?
The mon scripts are quite simple, which means that there's
not much to go wrong. The alert is a little complicated, but
there's nothing there that can cause a freeze - none of it
is kernel code. You should be able to send anything to ip_vs
vi the sysctl/ipvsadm and not freeze the machine.
What happens if you run the directors without mon for a
while, i.e. you assume that none of the services are going
to fail, and you just have one director running (at a time)?
This i can try as next testcase, but maybe it takes some days, if a next
crash (freeze) will happen.
Joe
--
Joseph Mack NA3T EME(B,D), FM05lw North Carolina
jmack (at) wm7d (dot) net - azimuthal equidistant map
generator at http://www.wm7d.net/azproj.shtml
Homepage http://www.austintek.com/ It's GNU/Linux!
|
