-
Notifications
You must be signed in to change notification settings - Fork 51
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add OpenShift deployment support #94
Conversation
5cd5cee
to
3481456
Compare
Error seen in operator logs |
9bfcec8
to
c87ffa2
Compare
Thanks @AbdYsn, I added the perms to |
Are there no more |
macvlan net-attach-def creation is handled by the operator : https://github.com/Mellanox/network-operator#macvlannetwork-crd there are examples under |
Is there a way to see the network-operator logs (or something more than the console log) as to why the CI is failing? |
The logs should be in http://13.74.249.42/nic_operator-ci//logs/ e,g. http://13.74.249.42/nic_operator-ci/68/logs/ |
@sjug, at the end of the console logs, just before the email sent, there is a |
From the logs seems like issue with rbac:
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
added some comments,
In addition we should have a way to determine if the operator is deployed in an openshift cluster. if it is not then the related objects should not be created.
manifests/stage-nv-peer-mem-driver/0050_nv-peer-mem-driver-ds.yaml
Outdated
Show resolved
Hide resolved
error is:
|
3c4d598
to
7584670
Compare
@AbdYsn Can you advise on how to get some meaningful logs out jenkins? The CI jobs failed but I have no idea why. |
@sjug So for the CI logs it is split to two parts, the console output (the link in the github status), and the job logs (a link to them is found at the end of the console output just before the email section). Most of the time if the CI is not broken, you can find the failures in the network operator pod logs found at the logs link (search for error in the logs and look for the last one). if nothing is found there then most of the time the CI itself would be broken (or the change require a change in the CI). regarding the last patchset, this error is found in the network operator logs here: |
40e3a76
to
69270b4
Compare
@adrianchiris @AbdYsn Please advise on how to fix the "code duplication" problem which seems to be the cause of the "Travis CI build" failure. |
@sjug well, CI reports these bits are very similar and perhaps some refactoring is needed to make code common. for now id try to only add the OsName to runtimeSpec of rdma shared device plugin stage. as thats the only new attribute you need. |
93d8487
to
ecdad24
Compare
@AbdYsn Both |
8549854
to
764f6bd
Compare
Any other outstanding issues? |
@e0ne and @adrianchiris PTAL |
@sjug can you rebase the PR? |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Hi Sebastian, Apologies for the late review.
I believe we are very close.
I have added some comments which proposes a way to address a comment i had in earlier review round:
In addition we should have a way to determine if the operator is deployed in an openshift cluster. if it is not then the related objects should not be created.
LMK what you think.
This PR addresses the states shared dp, mofed, nv_peer_mem deployment in openshift.
however what will happen if the user deploys nicclusterpolicy CR with other parts enabled ? e.g SecondaryNetwork
will that work or fail because of openshift scc
?
If it fails, does it fail with a reasonable error ?
(just thinking on the user experience here, im OK with not supporting all options)
if there are Openshift limitations, i think we need to have some documentation about it.
e.g a section in README about deployment in an Openshfit environment ?
I have already addressed your previous comment by adding the go template conditionals to the SCC. What is your new "proposed way" to address this? |
1880cc7
to
ef48d2c
Compare
/retest-nic_operator |
- Add new RBAC roles & clusterroles for all stages - Add OCP specific artifacts - Updated examples - Fixed some file permissions - Make SCC objects conditional on OCP - Add OSName fields back to state_shared_dp - Modify helm charts - Disable additional OCP objects with template boolean Signed-off-by: Sebastian Jug <[email protected]>
Operator successfully deploys and functions on OpenShift.