2 Mar
2011
2 Mar
'11
4:08 p.m.
On 1 March 2011 20:20, Andy Smith <andy(a)bitfolk.com> wrote:
Gotcha, thanks.
Hi Adam,
On Tue, Mar 01, 2011 at 01:42:06PM -0500, Adam Spiers wrote:
Makes sense, although I still think it would be helpful to have
DNS entries for those IPs, e.g. monitor{1,2,3}.bitfolk.com.
Are there corresponding DNS entries?
There aren't yet because I haven't actually deployed any other
monitoring nodes, and I'm not 100% sure how I'm going to do it. I
just know that we need to have more than one because:
- It's starting to get to the point where one host can't do every
check in a timely manner, and;
- If that one host is down there is no monitoring[1], which is less
than ideal
Also because at some point everything needs to renumber into
85.119.80.0/21 so might as well start with infrastructure. If it helps, the intention is for 85.119.80.0/24 to
only have
BitFolk infrastructure in it, not any customer-controlled stuff.
Yeah thanks.
And is the NTP
probe TCP or UDP?
Sorry, meant to answer that before.. AFAIK ntpd uses only 123/UDP
however /etc/services reserves both UDP and TCP for this protocol.
Personally I only open 123/UDP for this.