On 4/1/20 2:23 PM, Martin Necas wrote:
It's possible that the issue was introduced in the patch [1], but as Arthurs logs showed properly formatted ovirt_ca_cert, so not sure with it. 
Arthur/Marcin could you please check command in ovirt-engine/share/ovirt-engine/ansible-runner-service-project/artifacts you should see there variables with which the ansible-playbook is executed. 
It should be same as you linked but still want to make sure that there isn't some issue. Also you can check stdout file if there is some issue. 

I tried changing the host deployment playbook to inject a debug message:

  - name: Add vdsm cacert files
    copy:
      content: "{{ ovirt_ca_cert }}"
      dest: "{{ filedest }}"
      owner: 'root'
      group: 'kvm'
      mode: 0644
    with_items:
      - "{{ ovirt_vdsm_trust_store }}/{{ ovirt_vdsm_ca_file }}"
      - "{{ ovirt_vdsm_trust_store }}/{{ ovirt_vdsm_spice_ca_file }}"
      - "{{ ovirt_libvirt_default_trust_store }}/{{ ovirt_libvirt_default_client_ca_file }}"
    loop_control:
      loop_var: filedest

  - name: Show cacert
    debug:
      msg: CA contents 1987 {{ ovirt_ca_cert }}

and the result was:

2020-04-01 06:02:23 EDT - TASK [ovirt-host-deploy-vdsm-certificates : Show cacert] ***********************
2020-04-01 06:02:23 EDT - ok: [lago-basic-suite-master-host-1] => {
    "msg": "CA contents 1987 -----BEGIN CERTIFICATE----- MIIDhDCCAmygAwIBAgICEAAwDQYJKoZIhvcNAQELBQAwMzELMAkGA1UEBhMCVVMxDTALBgNVBAoM BFRlc3QxFTATBgNVBAMMDGVuZ2luZS4yNDU1NTAeFw0yMDAzMzEwOTU0MjRaFw0zMDAzMzAwOTU0 MjRaMDMxCzAJBgNVBAYTAlVTMQ0wCwYDVQQKDARUZXN0MRUwEwYDVQQDDAxlbmdpbmUuMjQ1NTUw ggEiMA0GCSqGSIb3DQEBAQUAA4IBDwAwggEKAoIBAQC3c20WyiBD98u6Ty6Yjb48fx9wuYUp2MIK j7E8qlX9QvNgvuudTYugPf040xyi+pcVhbXjqc7PhJoqowzgYxuyBu7W/KZigAp2pWMl12w7J1J/ 3Hp2IXD5hM7M6aCQ1jMDLxt1YECZfw+TEFVep1z7oxGZHPRZM8MDvYdBje+oPj41kIL1XNsCOiTy J8auU5/eaFbZFjP/sCDNuN14MnmhJtlVahRouODt86N1DRf3ubkmV/Bcr/Xp4iLx4ycyFiPU31cu Gnb2x8pTMPIbgtMYJTqMnRVrzJPV+ALA/PCSOL6LKkM7Jy4ecVFcGcJfvFpmsvF+qd7NuCOfqA7u l6EnAgMBAAGjgaEwgZ4wHQYDVR0OBBYEFPK3q/RlmHfh5o0KmmTguIALVwFgMFwGA1UdIwRVMFOA FPK3q/RlmHfh5o0KmmTguIALVwFgoTekNTAzMQswCQYDVQQGEwJVUzENMAsGA1UECgwEVGVzdDEV MBMGA1UEAwwMZW5naW5lLjI0NTU1ggIQADAPBgNVHRMBAf8EBTADAQH/MA4GA1UdDwEB/wQEAwIB BjANBgkqhkiG9w0BAQsFAAOCAQEAaHQqbgeG7ReoodKwbmFOFq99YOMrYmLx2llt5s49wz+eZsMN OIja8Dilyhew+r6aM30cXHm6U8dOZpLQ9Ga0Y1hk4Edu6Vu4x51WXZdVTkxIjhD+DrHsuaM0PZsE s1tq+ngBaMFxSdXIWNf7DUEf9hymxfLDoOjjVfxxlFtaDsBmu1dup/N8shzUrZ+bTt8i7TGG/JWl F+Iyq/A1EHXywFwr/ZsEAeRjStFt0IytbYprGi98yt9LRZ4puDooio8PI57crON+Cu9vqHsYU3yc lj8vLtwcr354LlY+nLO+cnslhirZlhIuLtytDvBXA8bNJ3EdlAInCfr6SnXKC61aqA== -----END CERTIFICATE----- "

So when running the playbook it's already broken.
Artur OTOH checked the value of the variable by breaking in the engine code and it seemed ok there.
Indeed I think there's a problem in [1].

[1] https://gerrit.ovirt.org/#/c/107683/5/backend/manager/modules/bll/src/main/java/org/ovirt/engine/core/common/utils/ansible/AnsibleRunnerHTTPClient.java


Martin Necas


On Wed, Apr 1, 2020 at 1:22 PM Artur Socha <asocha@redhat.com> wrote:
Posting a public pastebin url [1]. Apologies for using the private one before.

[1] https://pastebin.com/wrw5ME7j
A.


On Wed, Apr 1, 2020 at 12:31 PM Artur Socha <asocha@redhat.com> wrote:
>
> Adding request content:
> http://pastebin.test.redhat.com/850652
>
> A.
>
> On Wed, Apr 1, 2020 at 12:28 PM Artur Socha <asocha@redhat.com> wrote:
>>
>> I have debug the flow until the moment the request is being seng via http client to ansible runner service and until that point it was correct. The json did contain correctly formatted ovirt_ca_cert.
>> Artur
>>
>> On Wed, Apr 1, 2020 at 12:26 PM Marcin Sobczyk <msobczyk@redhat.com> wrote:
>>>
>>>
>>>
>>> On 4/1/20 11:54 AM, Martin Perina wrote:
>>>
>>>
>>>
>>> On Wed, Apr 1, 2020 at 11:15 AM Marcin Sobczyk <msobczyk@redhat.com> wrote:
>>>>
>>>>
>>>>
>>>> On 4/1/20 11:06 AM, Marcin Sobczyk wrote:
>>>> >
>>>> >
>>>> > On 4/1/20 9:51 AM, Marcin Sobczyk wrote:
>>>> >> Hi,
>>>> >>
>>>> >> On 4/1/20 8:44 AM, Yedidyah Bar David wrote:
>>>> >>> On Wed, Apr 1, 2020 at 6:21 AM <jenkins@jenkins.phx.ovirt.org> wrote:
>>>> >>>> Project:
>>>> >>>> https://jenkins.ovirt.org/job/ovirt-system-tests_he-basic-suite-master/
>>>> >>>>
>>>> >>>> Build:
>>>> >>>> https://jenkins.ovirt.org/job/ovirt-system-tests_he-basic-suite-master/1548/
>>>> >>> Previous build 1547 passed!, after many months of failing, thanks to
>>>> >>> Evgeny's work
>>>> >>> in recent weeks. Above one failed.
>>>> >>> I think the root cause is that the engine tried to connect to vdsm
>>>> >>> right after
>>>> >>> successfully finishing ansible host-deploy, but failed. vdsm.log has:
>>>> >>>
>>>> >>> https://jenkins.ovirt.org/job/ovirt-system-tests_he-basic-suite-master/1548/artifact/exported-artifacts/test_logs/he-basic-suite-master/post-he_deploy/lago-he-basic-suite-master-host-0/_var_log/vdsm/vdsm.log
>>>> >>>
>>>> >>>
>>>> >>> 2020-03-31 22:58:49,773-0400 ERROR (Reactor thread) [vds.dispatcher]
>>>> >>> uncaptured python exception, closing channel
>>>> >>> <yajsonrpc.betterAsyncore.Dispatcher connected
>>>> >>> ('::ffff:192.168.222.76', 46754, 0, 0) at 0x7f416c150a90> (<class
>>>> >>> 'ssl.SSLError'>:[X509] no certificate or crl found (_ssl.c:3771)
>>>> >>> [/usr/lib64/python3.6/asyncore.py|readwrite|110]
>>>> >>> [/usr/lib64/python3.6/asyncore.py|handle_write_event|442]
>>>> >>> [/usr/lib/python3.6/site-packages/yajsonrpc/betterAsyncore.py|handle_write|74]
>>>> >>>
>>>> >>> [/usr/lib/python3.6/site-packages/yajsonrpc/betterAsyncore.py|_delegate_call|168]
>>>> >>>
>>>> >>> [/usr/lib/python3.6/site-packages/vdsm/sslutils.py|handle_write|190]
>>>> >>> [/usr/lib/python3.6/site-packages/vdsm/sslutils.py|_handle_io|194]
>>>> >>> [/usr/lib/python3.6/site-packages/vdsm/sslutils.py|_set_up_socket|154])
>>>> >>> (betterAsyncore:179)
>>>> >>>
>>>> >>> Not sure what might have caused this. Can anyone have a look? Thanks.
>>>> >> Probably caused by https://gerrit.ovirt.org/108016
>>>> >> Looking into this.
>>>> >>
>>>> > Turns out that the patch is not the cause of the error per se - it simply
>>>> > uncovered a different problem - the CA on the hosts is broken:
>>>> >
>>>> > [root@lago-basic-suite-master-host-0 certs]# openssl x509 -in
>>>> > /etc/pki/vdsm/certs/cacert.pem -text
>>>> > unable to load certificate
>>>> > 139987452258112:error:0909006C:PEM routines:get_name:no start
>>>> > line:crypto/pem/pem_lib.c:745:Expecting: TRUSTED CERTIFICATE
>>>> It looks like they have spaces instead of newlines.
>>>> When I manually replaced the spaces to newlines, openssl is able to read
>>>> them.
>>>
>>>
>>> Martin/Dana, couldn't this be caused by any recent changes in ansible-runner integrations?
>>>
>>> This looks like a suspect to me:
>>>
>>> https://gerrit.ovirt.org/#/c/107683/5/backend/manager/modules/bll/src/main/java/org/ovirt/engine/core/common/utils/ansible/AnsibleRunnerHTTPClient.java
>>>
>>>>
>>>> >
>>>> >>>
>>>> >>>> Build Number: 1548
>>>> >>>> Build Status:  Failure
>>>> >>>> Triggered By: Started by timer
>>>> >>>>
>>>> >>>> -------------------------------------
>>>> >>>> Changes Since Last Success:
>>>> >>>> -------------------------------------
>>>> >>>> Changes for Build #1548
>>>> >>>> [Galit Rosenthal] Fix the repo for suites that weren't moved to no
>>>> >>>> reposync
>>>> >>>>
>>>> >>>>
>>>> >>>>
>>>> >>>>
>>>> >>>> -----------------
>>>> >>>> Failed Tests:
>>>> >>>> -----------------
>>>> >>>> No tests ran.
>>>> >>>
>>>> >>>
>>>> >>
>>>> >
>>>>
>>>
>>>
>>> --
>>> Martin Perina
>>> Manager, Software Engineering
>>> Red Hat Czech s.r.o.
>>>
>>>
>>
>>
>> --
>>
>> Artur Socha
>>
>> Senior Software Engineer, RHV
>>
>> Red Hat
>
>
>
> --
>
> Artur Socha
>
> Senior Software Engineer, RHV
>
> Red Hat



--

Artur Socha

Senior Software Engineer, RHV

Red Hat