Note: This bug is displayed in read-only format because the product is no longer active in Red Hat Bugzilla.

Bug 1915802

Summary: Ironic not getting proxy configurations required for reaching redfish endpoints
Product: OpenShift Container Platform Reporter: Mario Vázquez <mavazque>
Component: Bare Metal Hardware ProvisioningAssignee: Beth White <beth.white>
Bare Metal Hardware Provisioning sub component: cluster-baremetal-operator QA Contact: Amit Ugol <augol>
Status: CLOSED DUPLICATE Docs Contact:
Severity: unspecified    
Priority: unspecified CC: aos-bugs, djuran, mavazque, pablo.iranzo, sdasu, shsaxena, stbenjam, tsedovic
Version: 4.6Keywords: Triaged
Target Milestone: ---Flags: sdasu: needinfo-
Target Release: 4.7.0   
Hardware: x86_64   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2021-01-19 17:31:29 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Bug Depends On:    
Bug Blocks: 1831748    
Attachments:
Description Flags
ironic-conductor-podman-inspect-output none

Description Mario Vázquez 2021-01-13 13:02:48 UTC 
Created attachment 1747059 [details]
ironic-conductor-podman-inspect-output

Description of problem:

In an installation requiring a proxy to reach redfish endpoints, Ironic components are not getting the proxy configuration from the installer.

Version-Release number of selected component (if applicable): 

4.6.8


How reproducible:

Always

Steps to Reproduce:

1. Configure the required proxy configurations in the install-config file
2. Export the required proxy configurations before running the OpenShift installer
3. Wait for the ironic conductor to try to power on servers via redfish
4. Conductor times out since no proxy is configured on the container environment vars or ironic config files

Actual results:

Ironic / Container doesn't have the required proxy configurations.

Expected results:

Ironic / Container has the required proxy configurations.

Additional info:

If we export the HTTPS_PROXY env var and run a curl to redfish endpoint within the ironic-conductor container it works as expected. So we're assuming it's a matter of getting these env vars configured as specified in the install-config.


Proxy settings used on the install-config:

<OMITTED_OUTPUT>
proxy:
  httpProxy: http://192.168.0.10:3128
  httpsProxy: http://192.168.0.10:3128
  noProxy: .example.com,192.168.1.0/24
<OMITTED_OUTPUT>

Environment file we source before running the installer binary:

/etc/environment:

https_proxy=192.168.0.10:3128
http_proxy=192.168.0.10:3128
no_proxy=.example.com,192.168.1.10,192.168.1.12,192.168.1.13


Ironic conductor podman inspect output attached. You can see no env vars related to proxy are set. Ironic config files inside the container don't have the proxy setting.
Comment 3 Mario Vázquez 2021-01-14 16:36:47 UTC 
We moved the BMC to a different network to avoid requiring the proxy for accessing them. Then we found a different problem after the cluster is deployed.

The global proxy setting is set correctly, but the metal3 pod as its running with the hostnetwork seems to not use the proxy setting, which means that it cannot download the RHCOS images due to missing proxy.

The init container metal3-machine-os-downloader fails.


We tried to rsh into the container and as expected, when setting the https proxy everything works as expected: 

sh-4.4# curl https://mirror.openshift.com/pub/openshift-v4/dependencies/rhcos/latest/latest/rhcos-4.6.8-x86_64-openstack.x86_64.qcow2.gz                                                                              

sh-4.4#

sh-4.4#

sh-4.4# https_proxy=http://192.168.1.10:3128 curl https://mirror.openshift.com/pub/openshift-v4/dependencies/rhcos/latest/latest/rhcos-4.6.8-x86_64-openstack.x86_64.qcow2.gz -O                                     

  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current

                                 Dload  Upload   Total   Spent    Left  Speed

  0  844M    0  864k    0     0   850k      0  0:16:57  0:00:01  0:16:56  849k^C
Comment 5 Stephen Benjamin 2021-01-15 13:12:08 UTC 
The must-gather attached to the case doesn't indicate there's any cluster-wide proxy settings set. Can you share the full install-config the customer used?
Comment 6 Stephen Benjamin 2021-01-15 13:12:59 UTC 
And if that's not the most recently must-gather, please provide an updated must-gather, or installer log bundle.

Thanks!
Comment 7 Stephen Benjamin 2021-01-15 13:21:58 UTC 
I found it, sorry - it's there in ./cluster-scoped-resources/config.openshift.io/proxies/cluster.yaml, and I see it's not on the Metal3 pod.  Still investigating why, but you don't need to ask for a new must-gather.
Comment 8 sdasu 2021-01-19 17:18:25 UTC 
This will be fixed in CBO in 4.7 and not in MAO for 4.6.
Comment 9 sdasu 2021-01-19 17:31:29 UTC 

*** This bug has been marked as a duplicate of bug 1916772 ***
Comment 11 Red Hat Bugzilla 2023-09-15 00:58:17 UTC 
The needinfo request[s] on this closed bug have been removed as they have been unresolved for 500 days