[ovirt-users] External Provider https (unknown error)

[Stefan Wendler]

On 04/06/2018 19:43, Dominik Holler wrote:
> On Wed, 4 Apr 2018 13:29:56 +0200
> Stefan Wendler <stefan.wendler at tngtech.com> wrote:
> 
>> Hi,
>>
>> I am currently trying to attach Glance (OpenStack Image) and Cinder
>> (OpenStack Volume) as external provider and am facing a problem when
>> trying to use https in the Provider-URL on an ovirt 3.6 and 4.1
>> cluster.
>>
>> The Provider-URL I am using is in the form:
>> https://<host>:9292 (or port 8776 for Cinder - <host> is either a fqdn
>> or an IP-Address)
>>
>> Whenever I press the "Test" button in the "Add Provider" dialog I get
>> the message "Test Failed (unknown error)." There is no entry in any
>> logfile whatsoever (at least not in any logs that are associated with
>> ovirt). I would expect an ssl cert dialog here. I can telnet to the
>> destination ports from the engine and the nodes so Clance and Cinder
>> are reachable
>>
>> I have also read that this might happen, if there is a corrupted
>> /var/lib/ovirt-engine/external_truststore
>> But this file is not even existing and when i create it by hand, it is
>> not touched.
>>
>> How can I get this to work or even get an error message that gives me
>> a hint where to look?
>>
> 
> If there is something logged, it would be in engine.log.
> Can you please re-check if there is something related logged in
> engine.log?
> 
> Are you using authentication?
> 
> Do you use HTTPS for Glance/Cinder and authentication?

Thanks for the reply. The log file states nothing whatsoever.
Theresponse is so insanely fast that it feels as if something fails an
early validation. It does not matter, if I use authentication or not. If
I use http instead of https it logs the following error:

---- engine.log ----
2018-04-09 15:39:39,298+02 INFO
[org.ovirt.engine.core.bll.provider.TestProviderConnectivityCommand]
(default task-77) [b91f14f3-d8f6-43a1-a377-2d1523ee58dd] Running
command: TestProviderConnectivityCommand internal: false. Entities
affected :  ID: aaa00000-0000-0000-0000-123456789aaa Type: SystemAction
group CREATE_STORAGE_POOL with role type ADMIN
2018-04-09 15:39:39,403+02 ERROR
[org.ovirt.engine.core.bll.provider.storage.AbstractOpenStackStorageProviderProxy]
(default task-77) [b91f14f3-d8f6-43a1-a377-2d1523ee58dd] Not Found
(OpenStack response error code: 404)
2018-04-09 15:39:39,404+02 ERROR
[org.ovirt.engine.core.bll.provider.TestProviderConnectivityCommand]
(default task-77) [b91f14f3-d8f6-43a1-a377-2d1523ee58dd] Command
'org.ovirt.engine.core.bll.provider.TestProviderConnectivityCommand'
failed: EngineException: (Failed with error PROVIDER_FAILURE and code 5050)
---- EOM ----

So http goes thru and responds with the expected error (connection via
http is reset immediately). But for https it does not even get this far.
I can access the Provider-URL and Auth-URL via browser and from the
engine's command line and get a valid json response.

It also does not matter, if I use authentication or not. Error end
response rate stays the same. I checked with our hosting provider that
the urls are correct.

I have attached a screenshot of the "Edit Provider"-Window with the
message. It's the only place something is logged. Setting log level to
debug shows nothing either.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ovirt-glance.jpg
Type: image/jpeg
Size: 43467 bytes
Desc: not available
URL: <http://lists.ovirt.org/pipermail/users/attachments/20180410/0bd39852/attachment.jpg>