Bug 1649886

Note: I could see it on 5.10.0.23 as well.

Hi Boris,

Is this same bug as the one we found - ovn-provider stopped supporting "limit" parameter?

I do not think so, the limit issue did not cause an error - it just did not get the right data.

Hi Kedar,

I couldn't reproduce the error.

Boris fixed some bugs in the network provider refresh.
Can you please check if the bug is still reproduced with the latest nightly build?

If you manage to reproduce the error, please attach the evm.log and the ovirt-provider-ovn.log.

Hi Kedar, any progress with reproduction on latest version?

Created attachment 1514227 [details]
Screenshot

I was able to reproduce it on 5.10.0.29.

(In reply to Kedar Kulkarni from comment #8)
> Created attachment 1514227 [details]
> Screenshot
> 
> I was able to reproduce it on 5.10.0.29.

Alona, could you please take a look?

Hi Kedar,

Please attach the evm.log and the ovirt-provider-ovn.log

(In reply to Alona Kaplan from comment #10)
> Hi Kedar,
> 
> Please attach the evm.log and the ovirt-provider-ovn.log

The ovirt-provider-ovn.log should be in DEBUG level.

(In reply to Alona Kaplan from comment #11)
> (In reply to Alona Kaplan from comment #10)
> > Hi Kedar,
> > 
> > Please attach the evm.log and the ovirt-provider-ovn.log
> 
> The ovirt-provider-ovn.log should be in DEBUG level.

Hi Kedar,

could you please attach requested logs?

Thanks

Martin

This BZ seems to be duplicate of BZ #1661874

*** Bug 1661874 has been marked as a duplicate of this bug. ***

Is this a duplicate of https://github.com/ManageIQ/manageiq/issues/18457 ?

(replying to Dominik Holler, comment#19)
> Is this a duplicate of https://github.com/ManageIQ/manageiq/issues/18457 ?

I think so, but cannot say for sure, since there aren't attached logs, and thus I can't check the ovirt-provider-ovn version nor ovirt engine version.

It looks exactly like the same error...
#<Excon::Error::InternalServerError: Expected(200) <=> Actual(500 InternalServerError)
excon.error.response
  :body          => "{\n  \"error\": {\n    \"message\": \"Port_Group\",\n    \"code\": 500,\n    \"title\": \"Internal Server Error\"\n  }\n}\n"
  :cookies       => [
  ]
  :headers       => {
    "Connection"   => "close"
    "Content-Type" => "application/json"
    "Date"         => "Thu, 02 May 2019 12:23:49 GMT"
    "Server"       => "BaseHTTP/0.3 Python/2.7.5"
  }
  :host          => "env-rhv42-01-mgr.cfme2.lab.eng.rdu2.redhat.com"
  :local_address => "10.35.1.93"
  :local_port    => 41238
  :path          => "/v2.0/security-groups"
  :port          => 9696
  :reason_phrase => "Internal Server Error"
  :remote_ip     => "10.8.197.37"
  :status        => 500
  :status_line   => "HTTP/1.0 500 Internal Server Error\r\n"
>

Would it be possible to unblock the CloudForms NetworkManager for ovirt-provider-ovn of RHV 4.2, even if /v2.0/security-groups returns 500?

Marku, as mentioned in the email, would it be possible to fix that issue in CFME OpenStack Provider?

Hi, from OpenStack side, we do require security groups query in refresh. I don't think it makes sense ignore network security groups in OpenStack provider.

If RHV cannot respond to security groups query with at least empty response, I'd suggest make a temporary skip in ovirt provider (or we could add a switch to settings or somewhere which would keep current behaviour in OpenStack provider and allow ovirt provider skip these failures, but OpenStack provider doesn't seem to me as a right place).

(In reply to Marek Aufart from comment #24)
> Hi, from OpenStack side, we do require security groups query in refresh. I
> don't think it makes sense ignore network security groups in OpenStack
> provider.
> 
> If RHV cannot respond to security groups query with at least empty response,

Please note that only the most recent releases of RHV 4.2.x are affected.
The behavior of old RHV 4.2.x and RHV 4.3 is like expected.

> I'd suggest make a temporary skip in ovirt provider (or we could add a
> switch to settings or somewhere which would keep current behaviour in
> OpenStack provider and allow ovirt provider skip these failures, but
> OpenStack provider doesn't seem to me as a right place).

https://github.com/ManageIQ/manageiq-providers-ovirt/pull/365

New commit detected on ManageIQ/manageiq-providers-ovirt/master:

https://github.com/ManageIQ/manageiq-providers-ovirt/commit/9aecec67ac09448c8a4edb683cad0c60316d66f6
commit 9aecec67ac09448c8a4edb683cad0c60316d66f6
Author:     Adam Grare <agrare>
AuthorDate: Wed May 15 09:01:09 2019 -0400
Commit:     Adam Grare <agrare>
CommitDate: Wed May 15 09:01:09 2019 -0400

    Don't collect security groups in rhev refresh

    Skip collection of security groups in the network manager refresh.

    Fixes https://bugzilla.redhat.com/show_bug.cgi?id=1649886

 app/models/manageiq/providers/redhat/inventory/collector/network_manager.rb | 4 +
 1 file changed, 4 insertions(+)

Retargeting to 5.11 as we already have 5.10.5 clone BZ1669023

Verified in CFME 5.11.0.5.20190521144849_d0a06e6 and RHV 4.2.8.5-0.1.el7ev.
Network manager is refreshed without errors.