Fantastic. The documentation looks great, too. Is there an official
expected date for release available yet? I'm liking the software so far,
getting the reservation features working in MySQL will make it even better.
cheers,
Klaus
On Wed, Jun 22, 2016 at 8:39 AM, Marcin Siodelski
Is this the -stable release? I had issues with it not renewing / reacquiring a
lease after loading iPXE. I rolled forward to the latest version in Git which
didn't have this issue. Not sure if that's helpful.
cheers,Klaus
Sent from my tri-corder
On Thu, Jun 16, 2016 at 8:32 PM -0700,
CentOS 7 uses systemd. I'm by no means an expert on it, but you may be able
to shoehorn systemd into thinking that it's a managed service (there's
apparently legacy support for SVR4-style init scripts) using systemctl.
Again, being still new to systemd myself, I can't really offer much useful
in another
lease-file-like file.)
cheers,
Klaus
On Wed, Feb 8, 2017 at 8:43 AM, Tomek Mrugalski <tom...@isc.org> wrote:
> W dniu 08.02.2017 o 00:20, Klaus Steden pisze:
> >
> > Hi there,
> >
> > I realize I don't know what the expected behaviour is for Ke
Hi Maxime,
Have you defined the subnet as one that's owned by Kea in the config file?
Assuming you have different routers, netmasks, etc., you would have to
define a separate scope for both networks.
Presumably you've got DHCP forwarding enabled on your switches as well so
that DHCP requests get
ering with the looping boot process is one
> that affected ISC DHCP in a similar way, you just have to tune the
> conditions the DHCP server uses to distinguish the state of the booting
> client in order to direct it where you want to go. ... or flash iPXE on the
> firmware of all your serve
to direct it where you want to go. ... or flash iPXE on the
firmware of all your servers hahaha :-)
cheers,
Klaus
On Tue, Sep 13, 2016 at 9:29 AM, Christoffer Jönsson <bonc...@imap.cc>
wrote:
>
>
> On 2016-09-12 22:58, Klaus Steden wrote:
>
>
> I don't know about updating
Perfect. Thanks!
On Thu, Dec 8, 2016 at 11:48 AM, Thomas Markwalder <tm...@isc.org> wrote:
> On 12/8/16 2:44 PM, Klaus Steden wrote:
>
>
> Hi there,
>
> Is there a command line option that can be used to validate a
> configuration file without starting the daemon,
We had a similar problem, and our approach was to define a custom DHCP
attribute (ignored, but there for comment purposes) and generate the subnet
ID programmatically based on the contents of said comment using a hash
function to create a unique 32-bit integer.
e.g., "rack 41 VLAN 200" and "rack
I run all my Kea instances on Xen VMs with pretty lean provisioning, but
it's really going to depend on your use case.
A single lease transaction if you're using a SQL backend can take 2-3s to
complete, so if you're handing out thousands of leases per second, you'll
need a lot more horsepower
Hi there,
I realize that this isn't strictly speaking a Kea DHCP issue, but I thought
I'd ask anyway in hopes someone else may have seen what I'm seeing.
Basically, it looks like the onboard NIC firmware is getting confused while
talking to the tftpd server, so it abruptly stops reading the iPXE
As a footnote, I've also noticed that the machines I've tested so far do
not have any kind of GUID, e.g., it's all zeroes.
On Fri, Jun 2, 2017 at 3:11 AM, Klaus Steden <klausfi...@gmail.com> wrote:
>
> Hi there,
>
> I realize that this isn't strictly speaking a Kea DHCP issu
expanded.
> Alternatively, if the you know that the number of concurrently active
> clients is less than the addresses you have available, you may want to
> consider reducing the lease lifetime. In this way, addresses allocated
> to clients that are no longer active on the network will bec
This looks like it's got some real possibilities. Thank you, Baptiste!
cheers,
Klaus
On Thu, Aug 24, 2017 at 4:10 AM, Eron Lloyd wrote:
> Thank you for this, Baptiste! We're are preparing our Kea deployment now,
> and will test it out.
>
> On Thu, Aug 24, 2017 at 6:21
:* Sutherland, Rob <robert.b.sutherl...@windstream.com>
> *Sent:* Thursday, November 23, 2017 10:10:16 AM
> *To:* Hugh Connolly; Klaus Steden; SoupNazi izaNpuoS
> *Cc:* KEA-Users (kea-users@lists.isc.org)
> *Subject:* RE: [Kea-users] DHCP redundancy [faked-from]
>
>
> I’m using
ploy
>> my servers in the lab, it is a lot easier to deploy from a .deb package.
>>
>> Cheers,
>> Jason
>>
>> On Fri, Nov 24, 2017 at 4:52 AM, Klaus Steden <klausfi...@gmail.com>
>> wrote:
>>
>>>
>>> Owen -- I've done likew
+1 to this request.
cheers,
Klaus
On Mon, Dec 4, 2017 at 1:19 PM, Munroe Sollog wrote:
> Has there been any discussion about using GET instead of POST for the
> read-only (statistics) API? I'm working on integrating our monitoring and
> metrics with kea, and it would have
Hi Francis,
Thanks for the feedback. Your first point seems like the best practice
implementation, although in practice just using a dummy MAC for a
blocked-out address will work well enough.
We manage reservations through an external application talking to MySQL
directly, and so don't store
Hi everyone,
We had a new (for us) problem come up, and I wanted to poll the community
informally to see how everyone else has approached the issue.
We have a Windows cluster that uses L3 HA and passes a VIP back and forth
between two or more physical hosts.
To ensure that a Linux machine
I use Monit to keep tabs on mine (and also make sure they stay running).
cheers,
Klaus
On Fri, Jun 8, 2018 at 7:24 AM, Owen Dunn wrote:
> On Fri, 8 Jun 2018, itay cohen wrote:
>
> hi all
>>
>>
>> i want to check the availability of the kea-dhcp process with an external
>> program (via control
The UEFI section of my DHCP config looks like this:
{
"name": "bootstrap-efi",
"test": "substring(option[60].hex, 0, 20) == 'PXEClient:Arch:7'",
"option-data": [
{
"name": "boot-file-name",
"data": "ipxe/snponly.efi"
},
{
"name": "domain-name-servers",
This happens to us, too, using a shared MySQL backend for our
lease/reservation DBs.
I put Monit up on our Kea backends to automatically restart when it detects
that the process isn't running, and with a bit of tweaking, it's been
pretty bulletproof, with a bonus of handling other, unrelated
I've seen "unknown substitution variable ${shlibs:Depends}" many times
building Debian packages but I can't recall that it's ever caused an issue.
I think you're probably okay to ignore that one, it's a message burped by
the packaging process but shouldn't affect the binaries themselves.
cheers,
g in the design, and how I would easily stress test
> this, to see if the SQL group can handle the distributed writes from the
> Kea server.
>
> Thanks,
> Jason
>
> On Tue, Mar 13, 2018 at 8:17 PM, Klaus Steden <klausfi...@gmail.com>
> wrote:
>
>>
>&g
We have HA-MySQL behind our Kea servers, but the scopes themselves are
sharded to avoid this kind of write contention. In our case, we've got a
network partition that serves as an effective bright line between who's
responsible for what.
cheers
Klaus
On Tue, Mar 13, 2018 at 4:18 AM,
Yeah, it's lease/reservation info only. We generate our scopes
programmatically and add/remove them using configuration management.
cheers,
Klaus
On Mon, Feb 26, 2018 at 8:38 AM, xbgmsharp wrote:
> Thanks for the feedback, by using a backend i was expecting to have all
>
Hi Jordan,
You should be able to do a dump/restore from one database to the other when
you spin up the clustered ones. For separate tables, you should be able to
do this in stages, rather than migrating all your databases at once.
I would make sure your lease time is sufficiently long --
Hi there,
I've got a bit of a puzzler on my hands. We've racked some new gear and are
preparing to provision it.
Scope definitions have been added to Kea, and are active.
Hosts are on the network, and are generating DHCP requests.
Kea is responding as expected and offering leases, but none of
Hi Kristof,
We don't use Puppet, but we do use Salt to manage our Kea configurations.
Our usage is fairly tightly defined, however, as our scope definitions are
generated by an in-house IPAM application, so all the Salt state really
does is ensure the necessary software is in place and the
Vendramin
> *Date: *Friday, 15 February 2019 at 12.05
> *To: *"KEA-Users (kea-users@lists.isc.org)" ,
> Klaus Steden
> *Subject: *Re: [Kea-users] Hosts refusing lease offers?
>
>
>
> Hi Klaus,
>
>
>
> I remember a similar tricky situation. Some clien
... and it turns out our system integrator had the iDRAC interfaces set
to tag traffic, whereas our switches are configured for untagged traffic.
I'll throw myself out. :-)
thanks,
Klaus
On Fri, Feb 15, 2019 at 8:55 AM Klaus Steden wrote:
>
> Thanks for the suggestions, I checke
Hi Carlos,
If I'm understanding your question correctly, yes, you have to keep all
this information in the configuration file, it doesn't live in the database.
We had a similar problem in our environment where we allocated separate
subnets to each rack, a problem we solved by generating
You probably want the ISC DHCP mailing list, this list is for Kea DHCP,
which is a different product ...
hth,
Klaus
On Wed, Jun 19, 2019 at 3:04 PM Mayank Tiwari wrote:
> Hi,
>
> I am trying to compile ISC DHCP code as I want to use relay feature from
> it. When I compile the code from
>
On the outside chance you're using iPXE somewhere in your environment, I
had a very similar problem with some of our hardware where it would solicit
two separate leases by advertising two different hardware addresses (once
during PXE, and once at boot, if memory serves).
I was able to work around
You want something like this:
{
"Dhcp4":
{
...
# Use MySQL lease database backend to store leases in a database.
"lease-database": {
"type": "mysql",
"host": "__MYSQL_HOST__",
"name": "__MYSQL_DB__",
"user": "__MYSQL_USER__",
"password":
We have a similar concern with our data centers and ended up rolling our
own app that handles this problem.
Basically, if you use a database backend, you can use the database API (in
our case, MySQL) to create/remove reservations and manage leases (although
generally we don't do this all that
Responses inline below ...
On Mon, Aug 31, 2020 at 1:11 PM Tomek Mrugalski wrote:
> On 27/08/2020 03:58, Klaus Steden wrote:
> > I'm not sure what to do, but I'm stuck. I was trying to switch over
> > to the ISC's 1.6 packages of everything, so after installing, I ran
> >
Hi there,
I'm not sure what to do, but I'm stuck. I was trying to switch over to the
ISC's 1.6 packages of everything, so after installing, I ran through the
database upgrade procedure but it only got from 5.0 to 6.0.
For some reason, the 6.0 to 7.0 upgrade keeps failing, so I can't run the
1.6
that the link you sent discusses schema changes from v1.4 but I've
only ever used the MySQL backend here, so I'm quite stumped!
cheers,
Klaus
On Thu, Aug 27, 2020 at 11:00 AM Victoria Risk wrote:
> Klaus,
>
>
> On Aug 26, 2020, at 6:58 PM, Klaus Steden wrote:
>
>
> Hi there,
>
&
Hi there,
We use MySQL as a backend for our Kea infrastructure, and at the time we
were doing application development, the hook libraries were all licensed
products and a lot of the other nice glue just didn't exist. Granted we
don't do anything terrible sophisitcated with our setup, but within
Hi there,
Has anyone done any work with ZTP-capable switches with Kea? Specifically
Arista, although Juniper supports ZTP as well. I did some proof of concept
stuff back in 2016 with I think v0.9.0 and EOS back then, and met with some
success, but we've finally circled back to it looking to
unable to
> find. There must be some guide for developers or automation engineers which
> describes which records and columns are required, how they're to be
> formatted, etc.
>
> I'd be happy for someone to tell me a simpler way of achieving this that
> I've overlooked! :)
>
&g
A number of different HA scenarios are covered in the Kea documentation:
https://gitlab.isc.org/isc-projects/kea/-/wikis/designs/High-Availability-Design
cheers,
Klaus
On Fri, May 22, 2020 at 10:14 PM Software Info
wrote:
> Hi All
> I am moving from ISC-DHCP which I have had running now for
You can also use one of the many unused DHCP options for this. IIRC we used
option 245 in our environment (clients ignore it by default, so it's really
just for information purposes for humans).
cheers,
Klaus
On Thu, Jul 2, 2020 at 1:37 PM Francis Dupont wrote:
> > In Kea 1.6.2 how do I set a
I've got a similar workflow set up but I use VMware's *dcli* module to
create my instance, which also allows me to explicitly set the MAC address
(and thus force reservations).
Looking at the Packer docs (
https://www.packer.io/docs/builders/vsphere-iso.html) you may be able to do
this as well by
t Roland
>
> [image: www.exasoft.ch] <http://www.exasoft.ch>
> On 12/27/20 9:59 PM, Klaus Steden wrote:
>
>
> I've got a similar workflow set up but I use VMware's *dcli* module to
> create my instance, which also allows me to explicitly set the MAC address
> (and thus force
Honestly, I don't think this would be the most robust due to the way Docker
networking works. You will probably have more luck running Kea inside a
virtual machine that has raw socket access rather than a Docker container
that may or may not support raw socket passthrough.
I do not work for the
Yes: https://kea.readthedocs.io/en/kea-1.8.0/arm/agent.html
Although quickly skimming it looks like you'd probably want to bind it to
localhost and put an nginx reverse proxy in front of it to handle SSL
termination if you want to go the HTTPS route as the control agent does not
support this
My approach to this problem was to separate subnet definitions, via the
JSON file include mechanism, from the overall service configuration. The
service configuration doesn't change, so I just need to validate the
subnets file before updating, which can now be done in a hitless fashion.
So, +2
The hwaddr field in MySQL is stored as hexadecimal.
You want to use HEX/UNHEX to convert between ASCII presentation and hex
encoding.
cheers,
Klaus
On Tue, Sep 7, 2021 at 4:19 AM wrote:
> Hello.
> I have 1 question and one issue with storing leases information in mysql
> database in kea.
>
>
This looks pretty cool, but if I can offer a suggestion, I would report the
options in a slightly different format (comma-separated perhaps?) to make
it easier to distinguish from a MAC address ... I did a bit of a
double-take when I looked at the sample log message in your GitHub README.
cheers,
Our Kea solution is pretty lightweight:
2 VMs running Ubuntu 18
2 vCPUs per VM
8 GB vMEM (but using less than 4 GB)
DHCP VMs use hosted MySQL so neither VM runs a database process locally.
We're serving about 11K clients, load average hovers pretty close to zero
the whole time.
hth,
Klaus
On
FWIW we also primarily use Monit to keep tabs on the Kea daemons in our
environment. Because we're using MySQL backends, the Monit watchdog can
safely restart the daemon if it crashes or becomes unresponsive.
There is an external watchdog service we use to check the process table for
the daemon
Our largest environment serves about ~5700 clients and is an active-passive
pair of VMs with 2 vCPUs and 8 GB of vRAM using the MySQL backend.
The load average hovers around 0.00 on most days, although it's serving
predominantly bare metals in a data center and the default lease duration
is 120H.
This is what we added to our Kea configs to deal with the encapsulated
requests:
...
"option-def": [
{
"name": "link",
"code": 150,
"space": "relay-cisco",
"type": "ipv4-address",
"record-types": "",
"array": false,
ork "A" with one or more subnets and
> 192.168.120.17 might need to serve distinct network "B" with other
> subnets. These "distinct" networks do not share the same broadcast
> domain?
>
> On Wed, Feb 22, 2023 at 8:07 PM Klaus Steden wrote:
> >
>
is that they're at least using an
IPAM that has an API, and it *may* be possible to codify some kind of
association between subnets and their relays. Then it becomes less of a
management problem, and more of a programming one.
thanks!
Klaus
On Thu, Mar 2, 2023 at 8:59 AM Simon wrote:
> Klaus Ste
ince they could be all over the place?
> Hypothetical example:
>
> relay 10.1.2.1 might be a relay source for network "A"
> relay 10.1.2.2 might be a relay source for network "B"
> relay 10.1.2.3 might be a relay source for network "A"
>
> On Thu
another option would be to use shared
> > networks and add the subnet for relays along with the subnet of
> > addresses that you wish to allocate to the clients to a shared
> > network. See:
> https://kea.readthedocs.io/en/kea-2.2.0/arm/dhcp4-srv.html#shared-networks-in-d
ot;192.168.164.11-192.168.164.254"}],
"subnet": "192.168.164.0/24"
},
{
"id": ,
"option-data": [{"data": "192.168.182.1", "name": "routers"}],
"pools": [{"pool": "192.168.18
ter. It
> doesn't even have to be a real subnet. example:
>
> Network "A" is 10.1.2.0/24 and 10.1.3.0/24 and relays will be between
> 192.0.2.1 - 192.0.2.254 even though the relay subnet is actually
> 192.0.0.0/21
>
> You could add a "subnet" in your shared net
Hi there,
In some of our environments, we deal with DHCP relays, and their addresses
seem to proliferate faster than we can update our configs, which leads to
delays with DHCP service.
However, they have reserved an entire /21 for relay IPs, and ideally, I
would like to be able to add that
FWIW we've used Kea with /31s in our environment successfully (although
_why_ we used /31s is itself dumb and I can't recommend it)
cheers,
Klaus
On Thu, Jan 11, 2024 at 12:01 PM Scott Rakow wrote:
> Thanks!
>
> On Thu, Jan 11, 2024 at 7:47 PM Darren Ankney
> wrote:
>
>> Hi Scott,
>>
>> This
Hi,
We have a service that uses Kea on the backend behind an internal REST API
that performs CRUD operations against the lease and reservation tables.
We're still using v1.8, but it uses bog standard INSERT/UPDATE/DELETE
statements.
IMO, you should know why want to do it this way. In our case,
64 matches
Mail list logo