[BUG] asyncio timeout error experienced when attempting to get window_status from endpoint after commissioning window expires

[j-ororke]

Reproduction steps

Test Steps:

1. TH_CR1 starts a commissioning process with DUT_CE
2. TH_CR1 reads the BasicCommissioningInfo attribute from the General Commissioning cluster and saves the MaxCumulativeFailsafeSeconds field as max_window_duration. 3a. TH_CR1 opens a commissioning window on DUT_CE using a commissioning timeout of max_window_duration using ECM 3b. DNS-SD records shows DUT_CE advertising 3c. TH_CR1 writes and reads the Basic Information Cluster's NodeLabel mandatory attribute of DUT_CE
3. TH creates a controller (TH_CR2) on a new fabric and commissions DUT_CE using that controller. TH_CR2 should commission the device using a different NodeID than TH_CR1.
4. TH_CR1 reads the Fabrics attribute from the Node Operational Credentials cluster using a fabric-filtered read
5. TH_CR2 reads the Fabrics attribute from the Node Operational Credentials cluster using a fabric-filtered read
6. TH_CR1 writes and reads the Basic Information Cluster's NodeLabel mandatory attribute of DUT_CE
7. TH_CR2 reads, writes and then reads the Basic Information Cluster's NodeLabel mandatory attribute of DUT_CE
8. TH_CR2 opens a commissioning window on DUT_CE for 180 seconds using ECM
9. Wait for the commissioning window in step 8 to timeout
10. TH_CR2 reads the window status to verify the DUT_CE window is closed

Bug prevalence

Consistent

GitHub hash of the SDK that was being used

17b1a38e909e7874593bcb87c31be03a5866f1d4

Platform

python, raspi, vscode

Platform Version(s)

No response

Anything else?

Pre-Requisites:

TH_CR1 = Test harness as Commissioner 1 TH_CR2 = Test harness as Commissioner 2 DUT = Commissionee

Expected Result:

• After test step 8, we wait 180 seconds for ECM to timeout, after timeout period ends during step 10, we should be able to check the window status and receive a kWindowNotOpen Enum in response to inform us that the window is closed

Actual Result:

• We receive a timeout response of asyncio.exceptions.TimeoutError (Please see attached log for full failure output).

Logs Provided from TH:

• test_log_debug.txt
• test_log_info.txt
• device_output2.txt

Notes:

• Issue was noticed while using TH v2.11-beta3.1+fall2024
• A possible similar issue was reportedly noticed by QA team when running the CADMIN_1_3 yaml script on TH at the following link: 135
• We created a try and except workaround currently where we are able to communicate with the endpoint after the initial attempt from test step 11.
• Same issue is noticed on TH and in WSL Linux.

[bzbarsky-apple]

@cecille This sounds like a test harness issue....

[cecille]

possibly, possibly not. This came up when automating some tests, it's unclear right now which side is losing the connection, but that seems to be what's happening. I asked Jake to open this to track and share logs.