DNS solutions.

[Les Cargill]

(directions to a better asking-place always appreciated )

Suppose I had a pseudo random network of IP nodes. Say,
50 out of a possible 2000 were on a given network at any given time.

I want to have a DNS-like solution that uses the static hostname of
each node to populate the DNS tables.

DHCP may or may not be run on the same server. Dunno yet.

There may be DHCP for more than one Class C subnet. Dunno that yet,
either. I don't have a clear picture of how coupled DHCP and DNS
really are.

Human beings would be personally responsible for setting the hostname
per node.

Sadly, it's better if this runs on Windows, and not necessarily a
Windows server offering. It can simply be something that starts up on a
workstation.

I would not be adverse to a small-footprint Linux VM on said Windows
machine to do this. Also maybe some small ARM Linux solution perhaps.

What is this sort of thing called? I don't think the classic DNS BIND
service works for this, but I'd be eternally grateful to be proven
wrong - that's a "one book" solution.

[Char Jackson]

Post by Les Cargill
(directions to a better asking-place always appreciated )
Suppose I had a pseudo random network of IP nodes. Say,
50 out of a possible 2000 were on a given network at any given time.
I want to have a DNS-like solution that uses the static hostname of
each node to populate the DNS tables.
DHCP may or may not be run on the same server. Dunno yet.
There may be DHCP for more than one Class C subnet. Dunno that yet,
either. I don't have a clear picture of how coupled DHCP and DNS
really are.
Human beings would be personally responsible for setting the hostname
per node.
Sadly, it's better if this runs on Windows, and not necessarily a
Windows server offering. It can simply be something that starts up on a
workstation.
I would not be adverse to a small-footprint Linux VM on said Windows
machine to do this. Also maybe some small ARM Linux solution perhaps.
What is this sort of thing called?

I believe you'd call it a class assignment.

Post by Les Cargill
I don't think the classic DNS BIND
service works for this, but I'd be eternally grateful to be proven
wrong - that's a "one book" solution.

It sounds like the instructor intentionally left out key pieces of
information in an attempt to get you to connect the dots, but he/she also
included some random info to see if you would properly exclude it from your
proposed solution.

[Jorgen Grahn]

Post by Char Jackson

Post by Les Cargill
(directions to a better asking-place always appreciated )

...

Post by Char Jackson

Post by Les Cargill
I would not be adverse to a small-footprint Linux VM on said Windows
machine to do this. Also maybe some small ARM Linux solution perhaps.
What is this sort of thing called?

I believe you'd call it a class assignment.

Cargill is a Usenet regular; I suspect he's too old for classes, and
has a genuine need. (Plus it didn't sound like an assignment to me.)

/Jorgen

[Les Cargill]

Post by Char Jackson

Post by Les Cargill
(directions to a better asking-place always appreciated )

^^^^^^^^^^
Also this, Char.

Post by Char Jackson

Post by Les Cargill
Suppose I had a pseudo random network of IP nodes. Say,
50 out of a possible 2000 were on a given network at any given time.
I want to have a DNS-like solution that uses the static hostname of
each node to populate the DNS tables.
DHCP may or may not be run on the same server. Dunno yet.
There may be DHCP for more than one Class C subnet. Dunno that yet,
either. I don't have a clear picture of how coupled DHCP and DNS
really are.
Human beings would be personally responsible for setting the hostname
per node.
Sadly, it's better if this runs on Windows, and not necessarily a
Windows server offering. It can simply be something that starts up on a
workstation.
I would not be adverse to a small-footprint Linux VM on said Windows
machine to do this. Also maybe some small ARM Linux solution perhaps.
What is this sort of thing called?

I believe you'd call it a class assignment.

Heh. It's actually not. I'm just relatively clueless on the subject.

I work in the embedded space, and frequently, you'll have nodes
identified by nothing more than a static IP address. Hilarity ensues.

I'd like to find a lightweight, bolt-on solution for that. There's
often one workstation that is always on, and it is invariably a Windows
machine. If not, I can add a cheap ARM board to do this.

These networks will almost certainly not be on the larger Internet.
Totally airgapped. And "almost" really means "never" but ... things happen.

The sort of DHCP I am talking about is the sort you get included in
things like wireless routers. I understand enough to understand
that this may mean I have to use whatever is also in the DHCP server,
but I cannot say that categorically yet.

I see things like this:
https://sourceforge.net/projects/dhcp-dns-server/

Post by Char Jackson

Post by Les Cargill
I don't think the classic DNS BIND
service works for this, but I'd be eternally grateful to be proven
wrong - that's a "one book" solution.

It sounds like the instructor intentionally left out key pieces of
information in an attempt to get you to connect the dots, but he/she also
included some random info to see if you would properly exclude it from your
proposed solution.

Nope. The randomness is purely incidental or an artifact of pure ignorance.

[Char Jackson]

Post by Les Cargill
I work in the embedded space, and frequently, you'll have nodes
identified by nothing more than a static IP address. Hilarity ensues.

It's important to know how each node acquired its IP address. Is there
already a DHCP server in the picture? If so, you don't want to add a second
one. If not, and the existing IPs were all statically assigned, do you have
authority to reconfigure each node to change to DHCP? Further, who will
configure and administer the DNS and DHCP servers?

I don't know how many nodes you're talking about, or how often there are
changes, but if it's a manageable number you might consider a local
solution. For example, a simple mapping that you add to your hosts file so
that *you* get the benefit of being able to access a machine by a name that
you choose, without actually affecting the configuration of any node and
without introducing any new equipment, (DNS/DHCP), to the environment or
having to teach coworkers new procedures.

[Les Cargill]

Post by Char Jackson

Post by Les Cargill
I work in the embedded space, and frequently, you'll have nodes
identified by nothing more than a static IP address. Hilarity ensues.

It's important to know how each node acquired its IP address.

Somebody typed it in.

Post by Char Jackson
Is there
already a DHCP server in the picture?

There is a candidate, but I won't say it has to be used.

Post by Char Jackson
If so, you don't want to add a second
one.

Indeed. It has a nice off switch.

Post by Char Jackson
If not, and the existing IPs were all statically assigned, do you have
authority to reconfigure each node to change to DHCP?

Yessir.

Post by Char Jackson
Further, who will
configure and administer the DNS and DHCP servers?

That's TBD.

Post by Char Jackson
I don't know how many nodes you're talking about,

Tens, maybe a hundred if we're completely out of control.

Post by Char Jackson
or how often there are
changes, but if it's a manageable number you might consider a local
solution.

I would like that very much.

Post by Char Jackson
For example, a simple mapping that you add to your hosts file so
that *you* get the benefit of being able to access a machine by a name that
you choose, without actually affecting the configuration of any node and
without introducing any new equipment, (DNS/DHCP), to the environment or
having to teach coworkers new procedures.

I do like the /etc/hosts thing - we could simply pull the new one and
restart services.

Not a bad idea at all, but then somebody moves some equipment. I am
pretty sure I can butch up something to actually do this outside the
standard mechanisms, but I got to feeling that
using existing mechanisms might somehow be more virtuous.

[Robert Wessel]

On Wed, 23 Mar 2016 12:32:02 -0500, Les Cargill

Post by Les Cargill
(directions to a better asking-place always appreciated )
Suppose I had a pseudo random network of IP nodes. Say,
50 out of a possible 2000 were on a given network at any given time.
I want to have a DNS-like solution that uses the static hostname of
each node to populate the DNS tables.
DHCP may or may not be run on the same server. Dunno yet.
There may be DHCP for more than one Class C subnet. Dunno that yet,
either. I don't have a clear picture of how coupled DHCP and DNS
really are.
Human beings would be personally responsible for setting the hostname
per node.
Sadly, it's better if this runs on Windows, and not necessarily a
Windows server offering. It can simply be something that starts up on a
workstation.
I would not be adverse to a small-footprint Linux VM on said Windows
machine to do this. Also maybe some small ARM Linux solution perhaps.
What is this sort of thing called? I don't think the classic DNS BIND
service works for this, but I'd be eternally grateful to be proven
wrong - that's a "one book" solution.

I'm not sure I fully understand the problem you're trying to solve.
DHCP and DNS are usually fairly independent (there are some
implementations where the two are integrated), but many DHCP servers
can update a DNS server as they hand out IP address assignments. The
mechanism is usually called "Dynamic DNS" (although there can be other
sources of dynamic DNS updates, not just DHCP).

MS's DHCP server can update MS's DNS server (both on Windows Server)
fairly painlessly (although that's optional). See:

https://technet.microsoft.com/en-us/library/cc771732.aspx

You can also have Windows (DNS) client machines themselves issue the
DNS update, for example, when IP address assignments or the computer
name in the network configuration changes. MS calls that "dynamic
update", and the DNS server has to be configured to allow them:

https://technet.microsoft.com/en-us/library/cc771255.aspx

RFC2136 defines the standard Dynamic DNS protocol, but I'm not sure if
that's what MS DNS supports (although I suspect that it does, at least
in additional to the AD protocols it uses).

[Les Cargill]

Post by Robert Wessel
On Wed, 23 Mar 2016 12:32:02 -0500, Les Cargill

Post by Les Cargill
(directions to a better asking-place always appreciated )
Suppose I had a pseudo random network of IP nodes. Say,
50 out of a possible 2000 were on a given network at any given time.
I want to have a DNS-like solution that uses the static hostname of
each node to populate the DNS tables.
DHCP may or may not be run on the same server. Dunno yet.
There may be DHCP for more than one Class C subnet. Dunno that yet,
either. I don't have a clear picture of how coupled DHCP and DNS
really are.
Human beings would be personally responsible for setting the hostname
per node.
Sadly, it's better if this runs on Windows, and not necessarily a
Windows server offering. It can simply be something that starts up on a
workstation.
I would not be adverse to a small-footprint Linux VM on said Windows
machine to do this. Also maybe some small ARM Linux solution perhaps.
What is this sort of thing called? I don't think the classic DNS BIND
service works for this, but I'd be eternally grateful to be proven
wrong - that's a "one book" solution.

I'm not sure I fully understand the problem you're trying to solve.

You are excellent company. I'm not sure I do either.

In one sentence: I observe people using static IP addresses on a
semi-random network (think an engineering development/test network
without official IT sanction) and I wonder if there's some way to
organize this with a combination of DHCP and DNS in a lightweight and
inexpensive fashion.

I'd like there to be persistent state (hostname) on the nodes
that is used to influence the choice of DNS name, so that
pinging "TestTarget42" always gets the same node, but the IP address
can then be different.

Post by Robert Wessel
DHCP and DNS are usually fairly independent (there are some
implementations where the two are integrated), but many DHCP servers
can update a DNS server as they hand out IP address assignments. The
mechanism is usually called "Dynamic DNS" (although there can be other
sources of dynamic DNS updates, not just DHCP).
MS's DHCP server can update MS's DNS server (both on Windows Server)

There can be no provision for an install of Windows Server. It may
be that a Windows workstation can have a VM of a Linux server if
that is the only way.

Post by Robert Wessel
https://technet.microsoft.com/en-us/library/cc771732.aspx
You can also have Windows (DNS) client machines themselves issue the
DNS update, for example, when IP address assignments or the computer
name in the network configuration changes. MS calls that "dynamic
https://technet.microsoft.com/en-us/library/cc771255.aspx
RFC2136 defines the standard Dynamic DNS protocol, but I'm not sure if
that's what MS DNS supports (although I suspect that it does, at least
in additional to the AD protocols it uses).

[Robert Wessel]

On Thu, 24 Mar 2016 08:56:00 -0500, Les Cargill

Post by Les Cargill

Post by Robert Wessel
On Wed, 23 Mar 2016 12:32:02 -0500, Les Cargill

Post by Les Cargill
(directions to a better asking-place always appreciated )
Suppose I had a pseudo random network of IP nodes. Say,
50 out of a possible 2000 were on a given network at any given time.
I want to have a DNS-like solution that uses the static hostname of
each node to populate the DNS tables.
DHCP may or may not be run on the same server. Dunno yet.
There may be DHCP for more than one Class C subnet. Dunno that yet,
either. I don't have a clear picture of how coupled DHCP and DNS
really are.
Human beings would be personally responsible for setting the hostname
per node.
Sadly, it's better if this runs on Windows, and not necessarily a
Windows server offering. It can simply be something that starts up on a
workstation.
I would not be adverse to a small-footprint Linux VM on said Windows
machine to do this. Also maybe some small ARM Linux solution perhaps.
What is this sort of thing called? I don't think the classic DNS BIND
service works for this, but I'd be eternally grateful to be proven
wrong - that's a "one book" solution.

I'm not sure I fully understand the problem you're trying to solve.

You are excellent company. I'm not sure I do either.
In one sentence: I observe people using static IP addresses on a
semi-random network (think an engineering development/test network
without official IT sanction) and I wonder if there's some way to
organize this with a combination of DHCP and DNS in a lightweight and
inexpensive fashion.
I'd like there to be persistent state (hostname) on the nodes
that is used to influence the choice of DNS name, so that
pinging "TestTarget42" always gets the same node, but the IP address
can then be different.

How much can the nodes participate in this process? They could, for
example, issue a DDNS update to a small dedicated DNS server (which
could run on a Windows machine). As I mentioned that's basically what
Windows workstations do in dynamic update mode.

If that's too heavyweight, could the nodes send some sort of message
to a central machine, which would run a small script in response to
edit a zone file, and then trigger Bind (or whatever) to reload the
zone file? You need very little if you just want to refer to the
machines by name, just the A records (ignore reverse lookups and all
that). Just script removing any exist A record with the IP address,
and then add a new one with the IP address and name. Reserve a whole
subdomain for this (*.testnodes.mycompany.com). You'd also need to
update the zone serial number (just use the current date/time),
although if you don't have a secondary server you can omit that. Keep
the TTLs short.

If your DNS server supports something like nsupdate (Bind does, and
you can run Bind on Windows), just update the DNS server by creating
an appropriate update script executing the it with nsupdate.

If you can't do much on the nodes, perhaps you could poll the likely
IP addresses, and query the host names at each (which obviously would
require some support at each node). The do the DNS update as things
change.

In any event, then all you need is your main name server to point to
the testnodes subdomain.

If I'm understanding this correctly, it shouldn't be too difficult.

[Les Cargill]

Post by Robert Wessel
On Thu, 24 Mar 2016 08:56:00 -0500, Les Cargill

Post by Les Cargill

Post by Robert Wessel
On Wed, 23 Mar 2016 12:32:02 -0500, Les Cargill

Post by Les Cargill
(directions to a better asking-place always appreciated )
Suppose I had a pseudo random network of IP nodes. Say,
50 out of a possible 2000 were on a given network at any given time.
I want to have a DNS-like solution that uses the static hostname of
each node to populate the DNS tables.
DHCP may or may not be run on the same server. Dunno yet.
There may be DHCP for more than one Class C subnet. Dunno that yet,
either. I don't have a clear picture of how coupled DHCP and DNS
really are.
Human beings would be personally responsible for setting the hostname
per node.
Sadly, it's better if this runs on Windows, and not necessarily a
Windows server offering. It can simply be something that starts up on a
workstation.
I would not be adverse to a small-footprint Linux VM on said Windows
machine to do this. Also maybe some small ARM Linux solution perhaps.
What is this sort of thing called? I don't think the classic DNS BIND
service works for this, but I'd be eternally grateful to be proven
wrong - that's a "one book" solution.

I'm not sure I fully understand the problem you're trying to solve.

You are excellent company. I'm not sure I do either.
In one sentence: I observe people using static IP addresses on a
semi-random network (think an engineering development/test network
without official IT sanction) and I wonder if there's some way to
organize this with a combination of DHCP and DNS in a lightweight and
inexpensive fashion.
I'd like there to be persistent state (hostname) on the nodes
that is used to influence the choice of DNS name, so that
pinging "TestTarget42" always gets the same node, but the IP address
can then be different.

How much can the nodes participate in this process? They could, for
example, issue a DDNS update to a small dedicated DNS server (which
could run on a Windows machine). As I mentioned that's basically what
Windows workstations do in dynamic update mode.
If that's too heavyweight, could the nodes send some sort of message
to a central machine, which would run a small script in response to
edit a zone file, and then trigger Bind (or whatever) to reload the
zone file? You need very little if you just want to refer to the
machines by name, just the A records (ignore reverse lookups and all
that). Just script removing any exist A record with the IP address,
and then add a new one with the IP address and name. Reserve a whole
subdomain for this (*.testnodes.mycompany.com). You'd also need to
update the zone serial number (just use the current date/time),
although if you don't have a secondary server you can omit that. Keep
the TTLs short.
If your DNS server supports something like nsupdate (Bind does, and
you can run Bind on Windows), just update the DNS server by creating
an appropriate update script executing the it with nsupdate.

I rather like that. Will check. Thanks.

Post by Robert Wessel
If you can't do much on the nodes,

The nodes vary in intelligence.

Post by Robert Wessel
perhaps you could poll the likely
IP addresses, and query the host names at each (which obviously would
require some support at each node). The do the DNS update as things
change.

That's actually one thing I have considered. One node will always have
the option of a full ARP table, the rest is moving data.

I think the "nbstat" command can be used to find Windows hostnames
given the IP address.

Post by Robert Wessel
In any event, then all you need is your main name server to point to
the testnodes subdomain.
If I'm understanding this correctly, it shouldn't be too difficult.

I would not think so.

Thanks for your reply. It's kind of interesting there is no well-trod
path here*. You'd think this would be something people had done before.

*I say that based on other Googling round.