ovirt e2e: fix nodeport issue on CI bug 1794714#7614
ovirt e2e: fix nodeport issue on CI bug 1794714#7614Gal-Zaidman wants to merge 1 commit intoopenshift:masterfrom
Conversation
openshift-ci-robot
added
the
size/S
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: Gal-Zaidman The full list of commands accepted by this bot can be found here. DetailsNeeds approval from an approver in each of these files:Approvers can indicate their approval by writing |
Gal-Zaidman
force-pushed
the
fix-node-port
branch
from
7edcdca to
fe6c0d2
Compare
openshift-ci-robot
added
size/XS
Gal-Zaidman
force-pushed
the
fix-node-port
branch
2 times, most recently
from
1bb6881 to
5d1df3f
Compare
This patch adds the workaround suggested on [1] to make nodeport work and avoid the conformance tests failures. [1] https://bugzilla.redhat.com/show_bug.cgi?id=1794714 Signed-off-by: Gal Zaidman <gzaidman@redhat.com>
Gal-Zaidman
force-pushed
the
fix-node-port
branch
from
5d1df3f to
260124d
Compare
|
/test pj-rehearse |
|
/assign @wking |
| # Runs an install | ||
| - name: setup | ||
| # A midstep till we have the installer work merged, then we | ||
| # can use the CI artifact |
There was a problem hiding this comment.
Looks like this is your comment from #4340. Can you explain (ideally in the commit message), why you're removing it here?
There was a problem hiding this comment.
I will edit the commit message
| oc --insecure-skip-tls-verify get configs.imageregistry.operator.openshift.io/cluster > /dev/null && break | ||
| done | ||
| oc patch configs.imageregistry.operator.openshift.io cluster --type merge --patch '{"spec":{"managementState":"Managed","storage":{"emptyDir":{}}}}' | ||
| oc --insecure-skip-tls-verify patch configs.imageregistry.operator.openshift.io cluster --type merge --patch '{"spec":{"managementState":"Managed","storage":{"emptyDir":{}}}}' |
There was a problem hiding this comment.
Why are you adding these? Don't we have a kubeconfig with the CA for the Kube API? If that CA doesn't match what the cluster serves us, I'd rather error out here instead of ignoring the mismatch.
There was a problem hiding this comment.
+1
I will remove it, it worked fine before.
I have added it because for some reason the "get nodes" fail for me locally without it.
| sleep 10m | ||
| oc get co/image-registry | ||
| # This is a workaround for https://bugzilla.redhat.com/show_bug.cgi?id=1794714 | ||
| for n in $(oc --insecure-skip-tls-verify get nodes|awk '{print $1}'|grep worker);do |
There was a problem hiding this comment.
I don't like operating on the table output of get nodes. Can you use -o jsonpath=... to have it spit out the field you want?
There was a problem hiding this comment.
sure didn't know the option, I will try it
| oc get co/image-registry | ||
| # This is a workaround for https://bugzilla.redhat.com/show_bug.cgi?id=1794714 | ||
| for n in $(oc --insecure-skip-tls-verify get nodes|awk '{print $1}'|grep worker);do | ||
| oc -n default --insecure-skip-tls-verify debug node/$n --image=centos/tools -- ethtool --offload vxlan_sys_4789 tx off |
There was a problem hiding this comment.
I feel like this is the sort of change that should be handled by a MachineConfig, although I don't know what you'd put in the config to apply this specific ethtool tweak. Dropping into a debug session on each node and bumping things yourself seems very brittle.
There was a problem hiding this comment.
When we talked about it we said that we don't want to apply it by default on a cluster because it is a tweak that disables checksum and it seems to us like a decision that we don't want by default, plus it is a high priority bug which affects most platforms, happens only when using OpenshiftSDN , and has network people working on it.
We wanted to apply the fix in the CI because it made network test flaky during conformance.
|
Not relevant any more, decided to implement a MCO fix |
3 participants
This patch adds the workaround suggested on [1]
to make nodeport work and avoid the conformance
tests failures.
[1] https://bugzilla.redhat.com/show_bug.cgi?id=1794714
Signed-off-by: Gal Zaidman gzaidman@redhat.com