[Planetlab-users] cpu freezes question

[Roger Pack]

Thank you for the replies!  Yeah it turns out to definitely be somehow
related to network latency, not cpu (though Disk IO is also not fairly
processed, as some of you pointed out).  Anybody ever had situations of
network latency (socket opens/sends) taking (always) multiples of 10s? (like
10s, 20s, or 30s)?  I think it may be a problem with Ruby's generic sockets,
but, if anybody's ever seen anything like it... :)

> on a large number of PL nodes I get an SSH warning that the host
> identification has changed and thus all connections are refused. So I
> downloaded the latest known_hosts file from PLC, but the problem still
> persists.
>
> Does anyone know how to solve this?


Also, as a note for the host key changing warning, one thing that
(apparently) seems to works is to put
StrictHostKeyChecking no
in ~/.ssh/config

Thanks again.
-Roger
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.planet-lab.org/pipermail/users/attachments/20070827/b80f7d52/attachment.html