|
|
|
|
|
|
|
|
|
|
xen-users
Re: [Xen-users] Monitoring software - run in Dom0 or a DomU
On Fri, 2006-11-03 at 15:16 -0500, Thomas Harold wrote:
> So if we're going to setup Nagios to monitor our network, do you think
> we should set it up in a DomU (the current plan) or put it into a Dom0?
>
> (I'm leaning towards a DomU.)
Are you talking about the ideal place to put NRPE? Or where to put the
nagios install itself?
Bad idea to run any more public services than you have to on dom-0 .. so
yes, best if Nagios resides on a dom-u.
nrpe running on dom-0 shouldn't cause any problems, though it is rather
inefficient about how it gets the server vitals. iirc, nrpe will
malloc() quite a bit of room than needed to store the output it scrapes
from system() calls to things like vmstat, rather than just going right
to /proc (and dentry with file handles) to get what it needs.
Kind of silly for the monitor to have to branch into dirty paging,
contiguous blocks of cache on dom-0 is prime real estate. SNMP is also
sort of bloated.
That's only undesirable if dom-0 doesn't have much elbow room to begin
with.
This little tidbit is a little more sane for use on dom-0:
http://dev1.netkinetics.net/xmlpulse/
Its inspired by the phpsysinfo project, but is only a 7k executable. The
XML it produces is easily parsed. I hope to bundle it with a script to
install a small version of lighttpd to make a more dom-0 friendly
resource monitor, and add some of the libxenstat functions to it so it
includes the data you'd find in xentop.
You can also just run it with conventional ssh key pairs established.
Note : I'm probably way over-paranoid about how applications preform
(especially on dom-0) .. but it does lend to better solutions in the
long run :)
Best,
-Tim
> _______________________________________________
> Xen-users mailing list
> Xen-users@xxxxxxxxxxxxxxxxxxx
> http://lists.xensource.com/xen-users
>
>
_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users
|
|
|
|
|