[Planetlab-users] problem in accessing Planetlab nodes

[Kostis Kyzirakos]

There are some nodes that may have some problems with their dns servers, but
this shouldn't be the problem that you are facing. Do you use the same
gateway for all the pastry nodes? If so, the gateway node will get
overloaded while trying to serve 600 join requests. If a node is not able to
communicate with the gateway, does it try to communicate again after a
certain amount of time (like in Bamboo)?

I would also recommend using a light loaded node as a gateway (you can find
the light loaded nodes here:
http://summer.cs.princeton.edu/status/tabulator.cgi?table=table_nodeviewshort&select='resptime%20%3E%200%20&&%201minload%20%3C%205%20&&%20liveslices%20%3C=%205'&format=nameonly<http://summer.cs.princeton.edu/status/tabulator.cgi?table=table_nodeviewshort&select=%27resptime%20%3E%200%20&&%201minload%20%3C%205%20&&%20liveslices%20%3C=%205%27&format=nameonly>
)
 and use only nodes that have no problems (you can find these nodes here:
http://summer.cs.princeton.edu/status/tabulator.cgi?table=table_nodeviewshort&select='(resptime%20%3E%200%20&&%201minload%20%3C%205%20&&%20liveslices%20%3C=%205)%20&&%20((drift%20%3E%201m%20||%20(dns1udp%20%3E%2080%20&&%20dns2udp%20%3E%2080)%20||%20gbfree%20%3C%205%20||%20sshstatus%20%3E%202h)%20==%200)'&format=nameonly<http://summer.cs.princeton.edu/status/tabulator.cgi?table=table_nodeviewshort&select=%27%28resptime%20%3E%200%20&&%201minload%20%3C%205%20&&%20liveslices%20%3C=%205%29%20&&%20%28%28drift%20%3E%201m%20%7C%7C%20%28dns1udp%20%3E%2080%20&&%20dns2udp%20%3E%2080%29%20%7C%7C%20gbfree%20%3C%205%20%7C%7C%20sshstatus%20%3E%202h%29%20==%200%29%27&format=nameonly>
).

By the way we have been able to deploy our application that is built on top
of Bamboo in more than 400 nodes, so I believe that the problem you are
facing has to do with the bootstrapping phase.

Regards,
Kostis Kyzirakos


-- 
Kostis

On Sat, May 17, 2008 at 8:00 PM, Tariq Abdullah <t.abdullah at ewi.tudelft.nl>
wrote:

> Hi,
>
> I am running my experiments on Planetlab.
> My experiments involve creation of a Pastry overlay network.
> I use pssh to access Planetlab nodes. I am accessing 575 active nodes out
> of all Planetlab nodes.
> I observed from the results that very few than the required number of nodes
> is created.
> does any one has already faced the similar problem??
>
> Regards,
> Tariq Abdullah
>
> _______________________________________________
> Users mailing list: Users at lists.planet-lab.org
> https://lists.planet-lab.org/mailman/listinfo/users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.planet-lab.org/pipermail/users/attachments/20080517/72d019f7/attachment.html