[Kea-users] Question regarding to kea configuration in memory
tachibanashita at outlook.com
tachibanashita at outlook.com
Mon May 5 16:47:50 UTC 2025
Hi Darren,
Just an update. I was not able to recreate the issue on my company lab and home lab. If I see it occurs again and I will investigate and let you know.
Regards,
> On Apr 23, 2025, at 1:26 PM, Darren Ankney <darren.ankney at gmail.com> wrote:
>
> Hi,
>
> I am still unable to replicate this behavior. It works as expected
> for me. Please can you give the exact API commands and data payload
> required to reproduce this behavior?
>
> Thank you,
> Darren Ankney
>
> On Wed, Apr 23, 2025 at 8:50 AM tachibanashita at outlook.com
> <tachibanashita at outlook.com> wrote:
>>
>> I have unicast the modified config file to your email.
>>
>>
>> Thanks,
>>
>>
>>> On Apr 22, 2025, at 10:34 AM, Darren Ankney <darren.ankney at gmail.com> wrote:
>>>
>>> Hi,
>>>
>>> I am not able to reproduce this behavior. Please can you share your
>>> entire "Before config-write" Kea configuration? Please redact any
>>> sensitive items.
>>>
>>> Thank you,
>>> Darren Ankney
>>>
>>> On Mon, Apr 21, 2025 at 9:45 AM tachibanashita at outlook.com
>>> <tachibanashita at outlook.com> wrote:
>>>>
>>>> Hi Darren,
>>>>
>>>> I am using the latest dev Kea 2.7.7 version with RHEL9, the package for dhcp6 is isc-kea-dhcp6-2.7.7-isc20250326114722.el9.x86_64.rpm. Here’re my definition on the “device” class. I only have this client class defined, don’t have the “device_class2” client class in definition. And I am using client-classes so that part is expected.
>>>>
>>>>
>>>> {
>>>> "Dhcp6": {
>>>> "allocator": "iterative",
>>>> "calculate-tee-times": true,
>>>> "client-classes": [
>>>> {
>>>> "name": “device",
>>>> "option-data": [
>>>> {
>>>> "always-send": true,
>>>> "code": 34,
>>>> "csv-format": true,
>>>> "data": “xxxxxxxx",
>>>> "name": "syslog-servers",
>>>> "never-send": false,
>>>> "space": "vendor-4491"
>>>> },
>>>> {
>>>> "always-send": true,
>>>> "code": 37,
>>>> "csv-format": true,
>>>> "data": “xxxxxxxxx",
>>>> "name": "time-servers",
>>>> "never-send": false,
>>>> "space": "vendor-4491"
>>>> },
>>>> {
>>>> "always-send": true,
>>>> "code": 38,
>>>> "csv-format": true,
>>>> "data": "0",
>>>> "name": "time-offset",
>>>> "never-send": false,
>>>> "space": "vendor-4491"
>>>> },
>>>> {
>>>> "always-send": true,
>>>> "code": 61,
>>>> "csv-format": true,
>>>> "data": “xxxxx",
>>>> "name": "ccap-cores",
>>>> "never-send": false,
>>>> "space": "vendor-4491"
>>>> },
>>>> {
>>>> "always-send": true,
>>>> "code": 17,
>>>> "csv-format": false,
>>>> "data": "",
>>>> "name": "vendor-opts",
>>>> "never-send": false,
>>>> "space": "vendor-4491"
>>>> }
>>>> ],
>>>> "test": "substring(option[17].option[2].hex,0,3) == ‘device'"
>>>> }
>>>> ],
>>>>
>>>>
>>>>
>>>>
>>>> On Apr 19, 2025, at 6:19 AM, Darren Ankney <darren.ankney at gmail.com> wrote:
>>>>
>>>> Hi,
>>>>
>>>> Please share what version of Kea you are using. Not only did the
>>>> class name change, but the statement became a list (client-class vs
>>>> client-classes). Suspect you might be using a dev version as that
>>>> feature has not been released in stable. There could be some bug
>>>> there, though I am unsure why an _class2 would be added in this case.
>>>> Do you have a "device" or "device_class2" defined in client-classes
>>>> (not in the subnet but an actual class definition)?
>>>>
>>>> Thank you,
>>>> Darren Ankney
>>>>
>>>> On Thu, Apr 17, 2025 at 5:23 PM tachibanashita at outlook.com
>>>> <tachibanashita at outlook.com> wrote:
>>>>
>>>>
>>>> Greetings,
>>>>
>>>> I have a subnet definition in my kea-dhcp6.conf, it inherited its configuration from a client class called “device”, then I gave it some customization configuration in the option-data section. After I used the config-write command to get the config from memory, I noticed the client class of this subbed is no longer “device” but “device_class2”, even I never defined this client class. My guess is because of the customization data in option-data section, so kea realized its configuration is different from the original client class “device”, and it gives a new name.
>>>>
>>>> Is this behavior expected? I think this could bring some troubles when we try to do automatization. Now after we use the subnet hook, we have to use the config-write command Darren mentioned to save the config changes from memory to harddisk. And now in the new config file, we will have those “device_class2” client class which previously didn’t exist.
>>>>
>>>>
>>>> Original config in kea-dhcp6.conf:
>>>>
>>>> {
>>>> "id": 2,
>>>> "subnet": "2100:100:10:10::1/64",
>>>> "option-data": [
>>>> {
>>>> "space": "vendor-4491",
>>>> "name": "cores",
>>>> "code": 61,
>>>> "data": "2100:100:10:10::10-2100:100:10:10::100",
>>>> "always-send": true
>>>> }
>>>> ],
>>>> "pools": [
>>>> {
>>>> "pool": "2100:100:10:10::10-2100:100:10:10::100"
>>>> }
>>>> ],
>>>> "relay": {
>>>> "ip-addresses": [ "2100:100:10:10::1" ]
>>>> },
>>>> "interface": "ens2",
>>>> "client-class": “device",
>>>> "allocator": "iterative"
>>>> },
>>>>
>>>>
>>>> After using config-write:
>>>> {
>>>> "allocator": "iterative",
>>>> "calculate-tee-times": true,
>>>> "client-classes": [ “device_class2" ],
>>>> "id": 2,
>>>> "interface": "ens2",
>>>> "max-preferred-lifetime": 907200,
>>>> "max-valid-lifetime": 1209600,
>>>> "min-preferred-lifetime": 907200,
>>>> "min-valid-lifetime": 1209600,
>>>> "option-data": [
>>>> {
>>>> "always-send": true,
>>>> "code": 61,
>>>> "csv-format": true,
>>>> "data": “2100:100:0:30:0:0:0:22",
>>>> "name": "cores",
>>>> "never-send": false,
>>>> "space": "vendor-4491"
>>>> }
>>>> ],
>>>> "pd-allocator": "iterative",
>>>> "pd-pools": [ ],
>>>> "pools": [
>>>> {
>>>> "option-data": [ ],
>>>> "pool": “2100:100:10:10::10-2100:100:10:10::100"
>>>> }
>>>> ],
>>>> "preferred-lifetime": 907200,
>>>> "rapid-commit": false,
>>>> "rebind-timer": 1058400,
>>>> "relay": {
>>>> "ip-addresses": [ "2100:100:10:10::1" ]
>>>> },
>>>> "renew-timer": 604800,
>>>> "reservations": [ ],
>>>> "store-extended-info": false,
>>>> "subnet": "2100:100:10:10::1/64",
>>>> "t1-percent": 0.5,
>>>> "t2-percent": 0.8,
>>>> "valid-lifetime": 1209600
>>>> },
>>>> --
>>>> ISC funds the development of this software with paid support subscriptions. Contact us at https://www.isc.org/contact/ for more information.
>>>>
>>>> To unsubscribe visit https://lists.isc.org/mailman/listinfo/kea-users.
>>>>
>>>> Kea-users mailing list
>>>> Kea-users at lists.isc.org
>>>> https://lists.isc.org/mailman/listinfo/kea-users
>>>>
>>>> --
>>>> ISC funds the development of this software with paid support subscriptions. Contact us at https://www.isc.org/contact/ for more information.
>>>>
>>>> To unsubscribe visit https://lists.isc.org/mailman/listinfo/kea-users.
>>>>
>>>> Kea-users mailing list
>>>> Kea-users at lists.isc.org
>>>> https://lists.isc.org/mailman/listinfo/kea-users
>>>>
>>>>
>>>> --
>>>> ISC funds the development of this software with paid support subscriptions. Contact us at https://www.isc.org/contact/ for more information.
>>>>
>>>> To unsubscribe visit https://lists.isc.org/mailman/listinfo/kea-users.
>>>>
>>>> Kea-users mailing list
>>>> Kea-users at lists.isc.org
>>>> https://lists.isc.org/mailman/listinfo/kea-users
>>> --
>>> ISC funds the development of this software with paid support subscriptions. Contact us at https://www.isc.org/contact/ for more information.
>>>
>>> To unsubscribe visit https://lists.isc.org/mailman/listinfo/kea-users.
>>>
>>> Kea-users mailing list
>>> Kea-users at lists.isc.org
>>> https://lists.isc.org/mailman/listinfo/kea-users
>>
>> --
>> ISC funds the development of this software with paid support subscriptions. Contact us at https://www.isc.org/contact/ for more information.
>>
>> To unsubscribe visit https://lists.isc.org/mailman/listinfo/kea-users.
>>
>> Kea-users mailing list
>> Kea-users at lists.isc.org
>> https://lists.isc.org/mailman/listinfo/kea-users
> --
> ISC funds the development of this software with paid support subscriptions. Contact us at https://www.isc.org/contact/ for more information.
>
> To unsubscribe visit https://lists.isc.org/mailman/listinfo/kea-users.
>
> Kea-users mailing list
> Kea-users at lists.isc.org
> https://lists.isc.org/mailman/listinfo/kea-users
More information about the Kea-users
mailing list