DNAC - static port assignment with python

[Rajesh Kongath]

Greetings

dnac version 2.2.2.5

we have huge site where we have to assign pools statically, obviously its a cumbersome task due to slow GUI and many pools. Hence we thought of using " /dna/intent/api/v1/business/sda/hostonboarding/user-device". Unfortunately we are ending up with an error while trying more than one port at a time. I am really a newbie in python, appreciate your support.

I tried to follow @Aninda Chatterjee 's code achieve the ports assignment, however that ending with another error mentioning "Device not found" - however i really like his logic, appreciate the efforts.

Every time when we run the code, a dnac event generated as follows " Service request processing has failed with error: This intent (with name 'edge-1.company.com') is outdated as the controller has a newer intent. Resubmit the intent based on the current state of controller.

my code attached below for your reference

#Import required libraries
import requests
import json
from requests.auth import HTTPBasicAuth

# GET the AUTH TOKEN
token = requests.post("https://<DNAC-IP>/dna/system/api/v1/auth/token",auth=HTTPBasicAuth(username="<username>",password="<password>”),
                headers={'content-type': 'application/json'},verify=False,)
data = token.json()
# Assign Token to a variable
token=data['Token']

ports=['4','5']
print  (ports)

for i in ports:
    payload={
    "siteNameHierarchy": "<SiteHierarchy>",
    "deviceManagementIpAddress": "<deviceIP>",
    "interfaceName": "GigabitEthernet1/0/"+i,
    "dataIpAddressPoolName": "IP_Pool",
    "authenticateTemplateName": "Closed Authentication"
    }
    print (payload)
    response =  requests.post ('https://<DNAC-IP>/dna/intent/api/v1/business/sda/hostonboarding/user-device',json=payload,headers={
                    'X-Auth-Token': '{}'.format(token),
                    'Content-type': 'application/json',
                },
                
                verify=False
            )
    print (response)

Thanks in advance. 

[Rajesh Kongath · Accepted solution]

Adding a delay solved the issue..

however, adding 10 second for each port for a stack with 5 members is a joke. looking for any alternative, thanks.

import time
 
# Within the loop
time.sleep(10)
 

[Aninda Chatterjee]

Honestly, the code I wrote is pretty bad and was an extremely quick and dirty way of testing something. As you've tested, a manual delay is potentially the only way for now. The reason that this issue happens is that a second port assignment is sent to DNAC before the first one is fully processed (newer intent vs older intent).

Typically, there would be ways to make it faster - using python threading or async code. However, as you'll quickly realize (when you try to implement async code for this use case), this API has a concurrency of 1. Which means you cannot call this API multiple times concurrently (I've added my refactored code but it will not work for the reasons above - what you can try and do is remove the threading portion of it and just add delays).

We've roadmapped bulk port assignment via the API but it is not implemented yet. I can check and find out when it's coming.

[Rajesh Kongath]

Thank You Aninda for the clarification. I can see improvements in the API, for example until 2.2.2.5 we were not able to apply interface description but in 2.2.2.6, interface description also working. I hope bulk port assignment will be done soon!

Allow me to ask a newbie API question. as we have 3 nodes in the cluster, can we improve the performance of the script by sending each post request to different node?   

[Johannes_Grimm1]

Hi Rajesh and Aninda,

I just came across the thread and don't know if the answer is still relevant.

Working with static timers is always dangerous. My approach would be to filter the TaskID from the response and query the endTime parameter for completion. If the endTime is not yet present, the check is repeated with an increased wait interval. A function could look like the following (using dnacentersdk for simplicity):

def intent_execution_details(taskId):
    n = 1
    status = None
    while (status == None):
        time.sleep(n)
        status = dnac.task.get_task_by_id(taskId).response.endTime
        n = n * 2
        if n > 64:
            raise Exception("Call with " + taskId + " takes too much time.")

[AxelFox1982]

Another "bump" of this topic more than a year later.
I'm facing the same issue. While my code works perfectly, it takes an eternity to provision multiple ports.
My code picks up the task ID returned in the API response and then queries the status of the task every 10 seconds.
Only when the task is complete I continue with the next port.
The problem is ... each task to provision a single port triggers a re-provision of the switch, taking over 1 minute each time. So to provision 100 ports on the same switch/stack, you're looking at about 2 hours.

There does appear to be something available (or perhaps still in development) to do it in bulk for ports that have the same parameters, since I see the following in the API documentation (on DNAC version 2.3.5.3) :

But I can't get it to work, so I think it's not implemented yet or the documentation is not correct.
The API call doesn't accept a JSON object with the "interfaceNames" array and without the "interfaceName" string, since the "interfaceName" string is mandatory.
When I put the first interface in "interfaceName" and additional interfaces in the "interfaceNames" array, the API call accepts the object and succeeds, but still only provisions the first interface, so the "interfaceNames" array is ignored/not used.

[Preston Chilcote]

If you are following the documentation (which I believe you are), please open a TAC case so they can work with Engineering to figure out if this is a documentation error or supported feature.