[ale] load issue

[David Corbin]

On Wednesday 21 June 2006 07:57 pm, Jim Popovitch wrote:
> David Corbin wrote:
> > I have box that is getting "loads" in the 4-5 range, but when I run top,
> > it's 97.5% idle, and there are not 5 jobs that list a %CPU > 0.0.
>
> 1m LA is not a problem, a four or five 5m LA is something to look into.
>   If you leave top running does it fluctuate up and down?

The LA was > 4 going out to 15m.  I didn't see any fluctuations during the 3-5 
minutes I was poking around the server.  The problem has been going on since 
this morning around 4AM where nagios reported it at 'WARNING level', and then 
it escalted to "CRITICAL" about 2.5 hours later.   (Sorry I don't know 
exactly what those levels are, and I'm too lazy to go to the machine and dig 
them out just now).

> > When I login to the box, it's very responsive.  But, when I ssh to it, I
> > never get any response.  It doesn't fail, just hangs.

> Could be related to the system doing a PTR lookup on the connecting IP.

I don't think so, but I don't know.

>   Does the ssh session ever connect? 

No. The last thing it does is offer my public key and it 'never' comes back or 
timesout.

>   What about firewall(s)/iptables in  between you and the host?

There are none.  Both machines are on the same LAN, and I don't think I even 
have any of the iptables stuff in the kernels (though I can't be 100% sure of 
the latter).

And, of course, things worked fine "not long ago".