bug: missing until for retry in gce plugin

[cnfrancis]

Missing the until key for the wait_for module in the gce plugin. Apparently, it doesn't work when the until is missing.

[apatard]

You're patching the Linux file. What about its WINRM counterpart ? I guess it may suffer of the same kind of issue ?

[cnfrancis]

You're patching the Linux file. What about its WINRM counterpart ? I guess it may suffer of the same kind of issue ?

yes (added the same for winrm), but actually I wonder if it's because of this line in the wait_for module code, perhaps some leaky error that can't quite be captured using only ansible language as the python code isn't wrapped in a try except or using a retry.

I re-reviewed the error I shared and the failed = false, so my change although making this more robust, I believe still doesn't cover this case.

what do you think @apatard ?

edit: I think we should be good, i tested locally with this:

import socket
import threading
import struct

count=0
def malicious_server():
    server = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
    server.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)
    server.bind(("localhost", 12345))
    server.listen(1)
    while True:
        global count
        client, addr = server.accept()
        if count > 5:
            print("SSH")
            client.send("SSH\n".encode())
            client.close()
            break
        # Immediately close the connection with RST by setting SO_LINGER
        l_onoff = 1
        l_linger = 0
        client.setsockopt(
            socket.SOL_SOCKET, socket.SO_LINGER, struct.pack("ii", l_onoff, l_linger)
        )
        client.close()
        print(count)
        count+=1


# Start the server in a separate thread
server_thread = threading.Thread(target=malicious_server)
# server_thread.daemon = True
server_thread.start()

---
- name: Test waitfor
  hosts: localhost
  gather_facts: false
  remote_user: cnfrancis
  connection: local
  tasks:
    - name: Wait for SSH
      ansible.builtin.wait_for:
        port: 12345
        host: "localhost"
        search_regex: SSH
        delay: 0
      register: waitfor
      until: waitfor.failed == false
      retries: 6
      delay: 1
    - name: Print output
      ansible.builtin.debug:
        msg:
          # - "{{ waitfor }}"
          - "{{ waitfor }}"

TASK [Wait for SSH] ******************************************************************************************************
FAILED - RETRYING: [localhost]: Wait for SSH (6 retries left).
FAILED - RETRYING: [localhost]: Wait for SSH (5 retries left).
FAILED - RETRYING: [localhost]: Wait for SSH (4 retries left).
FAILED - RETRYING: [localhost]: Wait for SSH (3 retries left).
FAILED - RETRYING: [localhost]: Wait for SSH (2 retries left).
FAILED - RETRYING: [localhost]: Wait for SSH (1 retries left).
ok: [localhost]

TASK [Print output] ******************************************************************************************************
ok: [localhost] => {
    "msg": [
        {
            "attempts": 7,
            "changed": false,
            "elapsed": 0,
            "failed": false,
            "match_groupdict": {},
            "match_groups": [],
            "path": null,
            "port": 12345,
            "search_regex": "SSH",
            "state": "started"
        }
    ]
}

PLAY RECAP ***************************************************************************************************************
localhost                  : ok=2    changed=0    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0