[BUG] Discovery got same problem with body variable

[elwood218]

Describe the bug

There are 2 errors but one is related to the module and the other is imho a problem of the API.
The error with the module is again that "body" is used but not defined. (If I put that in a loop over "play_hosts" then it works)
The general problem is that first hosts get "service discovered" and then there is a 500 error or so..

TASK [cmk_host_registration : Add/update/remove host] ****************************************************************
changed: [Host1 -> localhost]
changed: [Host2 -> localhost]
changed: [Host3 -> localhost]
changed: [Host4 -> localhost]
changed: [Host5 -> localhost]

RUNNING HANDLER [cmk_host_registration : service discovery] **********************************************************
changed: [Host4 -> localhost]
changed: [Host1 -> localhost]
changed: [Host3 -> localhost]
An exception occurred during task execution. To see the full traceback, use -vvv. The error was: KeyError: 'body'
fatal: [Host5 -> localhost]: FAILED! => {"changed": false, "module_stderr": "Traceback (most recent call last):\n  File \"/Users/mathias.buresch/.ansible/tmp/ansible-tmp-1653903276.5778098-98682-264119765984817/AnsiballZ_cmk_discovery.py\", line 107, in <module>\n    _ansiballz_main()\n  File \"/Users/mathias.buresch/.ansible/tmp/ansible-tmp-1653903276.5778098-98682-264119765984817/AnsiballZ_cmk_discovery.py\", line 99, in _ansiballz_main\n    invoke_module(zipped_mod, temp_path, ANSIBALLZ_PARAMS)\n  File \"/Users/mathias.buresch/.ansible/tmp/ansible-tmp-1653903276.5778098-98682-264119765984817/AnsiballZ_cmk_discovery.py\", line 47, in invoke_module\n    runpy.run_module(mod_name='ansible.modules.cmk_discovery', init_globals=dict(_module_fqn='ansible.modules.cmk_discovery', _modlib_path=modlib_path),\n  File \"/usr/local/Cellar/python@3.10/3.10.4/Frameworks/Python.framework/Versions/3.10/lib/python3.10/runpy.py\", line 209, in run_module\n    return _run_module_code(code, init_globals, run_name, mod_spec)\n  File \"/usr/local/Cellar/python@3.10/3.10.4/Frameworks/Python.framework/Versions/3.10/lib/python3.10/runpy.py\", line 96, in _run_module_code\n    _run_code(code, mod_globals, init_globals,\n  File \"/usr/local/Cellar/python@3.10/3.10.4/Frameworks/Python.framework/Versions/3.10/lib/python3.10/runpy.py\", line 86, in _run_code\n    exec(code, run_globals)\n  File \"/var/folders/6_/qjhhb7fn13g6r839l49tjg58s_y621/T/ansible_cmk_discovery_payload_vulpzc1w/ansible_cmk_discovery_payload.zip/ansible/modules/cmk_discovery.py\", line 170, in <module>\n  File \"/var/folders/6_/qjhhb7fn13g6r839l49tjg58s_y621/T/ansible_cmk_discovery_payload_vulpzc1w/ansible_cmk_discovery_payload.zip/ansible/modules/cmk_discovery.py\", line 166, in main\n  File \"/var/folders/6_/qjhhb7fn13g6r839l49tjg58s_y621/T/ansible_cmk_discovery_payload_vulpzc1w/ansible_cmk_discovery_payload.zip/ansible/modules/cmk_discovery.py\", line 147, in run_module\nKeyError: 'body'\n", "module_stdout": "", "msg": "MODULE FAILURE\nSee stdout/stderr for the exact error", "rc": 1}
An exception occurred during task execution. To see the full traceback, use -vvv. The error was: KeyError: 'body'

Component Name

discovery

Ansible Version

$ ansible --version
ansible [core 2.12.5]

Checkmk Version

2.1.0 (CEE)

Collection Version

$ ansible-galaxy collection list
Tested with 0.3.2 and "devel"

Environment

To Reproduce
Steps to reproduce the behavior:

- name: "Add/update/remove host"
#  tribe29.checkmk.host:
  cmk_host:
    server_url: "https://{{ cmk_central }}/"
    site: "{{ cmk_central_site }}"
    automation_user: "{{ cmk_site_user }}"
    automation_secret: "{{ cmk_site_password }}"
    host_name: "{{ cmk_host_name | default(host_name) }}"
    attributes:
      "{{ cmk_host_attributes | default(omit) }}"
    folder: "{{ host_folder }}"
    state: "{{ cmk_host_state | default('present') }}"
  delegate_to: localhost
  become: no
  notify:
    - service discovery
    - activate changes

- name: service discovery
#  tribe29.checkmk.discovery:
  cmk_discovery:
    server_url: "https://{{ cmk_central }}/"
    site: "{{ cmk_central_site }}"
    automation_user: "{{ cmk_site_user }}"
    automation_secret: "{{ cmk_site_password }}"
    host_name: "{{ host_name }}"
#    host_name: "{{ hostvars[item]['cmk_host_name'] | default(hostvars[item]['host_name']) }}"
#    state: "fix_all"
    state: "new"
  delegate_to: localhost
#  loop: "{{ play_hosts }}"
##  loop_control:
##   pause: 3
#  run_once: true
  become: no

Expected behavior

1. There should be a correct error message
2. API should not throw an error

Actual behavior

Screenshots

Additional context

[lgetwan]

Hi @elwood218,

I fixed the discovery module in the devel branch. The error message still looks a bit ugly, but at least it's no longer an "Exception", and it contains the actual error message from the API.
Can you please test it?

Best regards
Lars

[elwood218]

@lgetwan This is the output now.

TASK [cmk_host_registration : service discovery] *********************************************************************
Friday 03 June 2022  11:09:39 +0200 (0:00:17.068)       0:01:15.787 ***********
Friday 03 June 2022  11:09:39 +0200 (0:00:17.068)       0:01:15.787 ***********
changed: [pgdb03 -> localhost]
changed: [pgdb02 -> localhost]
changed: [db01 -> localhost]
fatal: [pgdb01 -> localhost]: FAILED! => changed=false
  http_code: -1
  msg: 'Error calling API. HTTP Return Code is -1 Details: N/A'
fatal: [db02 -> localhost]: FAILED! => changed=false
  http_code: -1
  msg: 'Error calling API. HTTP Return Code is -1 Details: N/A'
fatal: [app01 -> localhost]: FAILED! => changed=false
  http_code: -1
  msg: 'Error calling API. HTTP Return Code is -1 Details: N/A'

[lgetwan]

Hi @elwood218,

I definitely didn't expect a "-1". :-o
But when checking the sources of fetch_url, I found some further information about that case. I improved the error message. Can you please check again?
Thanks so far for your help!

Best regards
Lars

[elwood218]

Hi, if that git pull are your changes then the output is same:

From https://github.com/tribe29/ansible-collection-tribe29.checkmk
   3c52245..32ca914  devel                    -> origin/devel
   58aba7f..8315f77  fix-51-exceptionhandling -> origin/fix-51-exceptionhandling
Updating 3c52245..32ca914

 fatal: [pgdb03 -> localhost]: FAILED! => changed=false
  http_code: -1
  msg: 'Error calling API. HTTP Return Code is -1 Details: N/A'

[lgetwan]

Hi,
for some reason, my fix wasn't merged into the devel branch. I did that again and verified it. Can you please try again?
Best regards
Lars

[elwood218]

Hi, now it looks like that:

fatal: [pgdb03 -> localhost]: FAILED! => changed=false
  http_code: -1
  msg: 'Error calling API. HTTP Return Code is -1 Details: Connection failure: The read operation timed out'

Btw. before when I was getting a 500 from the API sometimes a link to a crashreport was directly shown.. but it seems that this is not working anymore. Just wanted to mention it - don't know if it is related.

[lgetwan]

Hi,
for some reason, your API is not responding in time.
Probably you should try to do the call with curl for testing:

out=$(curl -v \
    --request POST \
    --write-out "\nxxx-status_code=%{http_code}\n" \
    --header "Authorization: Bearer <your_user> <your_password>" \
    --header "Accept: application/json" \
    --header "Content-Type: application/json" \
    --data '{ "alias": "SLES12_cgroup2", "name": "SLES12_cgroup2" }' \
    "http://<your_cmk_server>/<your_cmk_site>/check_mk/api/1.0/objects/host/<your_host>/actions/discover_services/invoke")

echo "$out"
resp=$( echo "${out}" | grep -v "xxx-status_code" )
code=$( echo "${out}" | awk -F"=" '/^xxx-status_code/ {print $2}')
echo "$resp" | jq
if [[ $code -lt 400 ]]; then
     echo "OK"
else
     echo "Request error"
fi

How long does this call take approximately?

[elwood218]

Hi, I don't think that this is the problem.

xxx-status_code=302
OK

The problem only occurs after the first 2 or 3 hosts and only when called like I already described in #33.

But you can close the this issue if you want because it was more or less only about the body variable. At least we got now a "better" error message.
The issue with the API I don't really know how to debug and for now my workaround is at least working.

Thank you for your help.

[elwood218]

I just recognized that the curl was missing a "follow" so I put https manually. Now this is the output:

{"title": "Bad Request", "status": 400, "detail": "These fields have problems: name, alias", "fields": {"name": ["Unknown field."], "alias": ["Unknown field."]}}
xxx-status_code=400
{
  "title": "Bad Request",
  "status": 400,
  "detail": "These fields have problems: name, alias",
  "fields": {
    "name": [
      "Unknown field."
    ],
    "alias": [
      "Unknown field."
    ]
  }
}
Request error

[elwood218]

It must look like that I guess:

   --data '{
          "mode": "new"
        }' \

[elwood218]

           http_code:  200
     time_namelookup:  0.072466s
        time_connect:  0.087947s
     time_appconnect:  0.250204s
    time_pretransfer:  0.250424s
       time_redirect:  0.000000s
  time_starttransfer:  0.250430s
                     ----------
          time_total:  4.760772s

But service discovery could take longer.. and like I said it is working when I "loop" over it in Ansible but when I run it for each inventory host (which could be parallel) then it fails.

[elwood218]

I have just tested it on shell.

discover-all-hosts.sh -> this script does a "for" loop - so sequential and it takes about 4 secs for each host like in my last post.

But if I do that and run all those 4 scripts in parallel:

discover-host1.sh
discover-host2.sh
discover-host3.sh
discover-host4.sh

Then these are the times of each host:

      time_total:  4.544909s
      time_total:  9.610563s
      time_total:  13.358466s
      time_total:  17.429323s

[robin-checkmk]

If I understand this right, we are chasing another issue here, so @elwood218 would you be so kind to close this issue and move your research about the timing issue into a new issue? There we can then research, whether it is a collection issue or an API issue.

Thanks!

[elwood218]

@robin-tribe29 It is an issue of the API because I have tested it on shell. So I write you an email at feedback if you can't put a ticket yourself. This ticket can be closed.

[robin-checkmk]

@elwood218 Alright, then please send an email, so all relevant information is in there. Thanks!

[elwood218]

But maybe the module could use bulk discovery but I don't know how to solve that in Ansible.. (and the discovery and "accepting" is not working anyway - already opened a FEED for that)