Hey all,
So I've been messing around with the collectd daemon, which
I think will work quite well for collecting statistics from the mesh
nodes.
Specifically, I was able to set up two copies of the daemon
and have a "client" (a PicoStation) transmit data to a "server"
(my laptop), which stores the data in RRD files (circular
databases of time-series data). From there, there appear
to be many tools that can produce pretty graphics (cacti,
cricket, drraw), but I haven't really explored this too much.
It is easy to set up and pretty lightweight and I believe it can
also be made to scale rather well. I didn't find a BATMAN module,
but there is one for OLSR which we could pillage. It uses a push
model (clients initiate the sending of data), which might be a good
fit for a decentralized network.
If we go this route we're going to need a machine on the mesh
running a web server that receives the RRD data from all the
nodes (and optionally syslog data).
Alex
On 05/05/2014 10:46 AM, Max B wrote:
Hey folks,
Last week at the meeting we talked a bunch about the necessity of
getting some basic node monitoring/error reporting up. Chris J mentioned
that NRPE w/ an Icinga (nagios fork) monitor server would be a
potentially quick and easy way to bootstrap some basic reporting. Alex
said that he was going to try to install the openWRT NRPE package on a
node and try to get a monitor server build receiving info from that test
node.
Are there any status updates on this? Alex, I realized I don't have your
email, so this is kind of my overly-public way of getting in touch. I
have some time this week to work on this, but I thought I'd check in to
see where other folks were at to prevent any duplication of efforts.
Also, I realized that if you need github write access, I can provide you
with that.
Awesome - keep on meshin!
Max
_______________________________________________
mesh mailing list
mesh(a)lists.sudoroom.org
https://lists.sudoroom.org/listinfo/mesh