Fix SetInterfaceMTU not working on Windows bug

[lzhecheng]

For Windows, antrea-gw0 interface must be enabled before
MTU is set, so MTU is not configured with current code.
Currently the result of SetInterfaceMTU is not checked, so
the error is not found.

Signed-off-by: Zhecheng Li lzhecheng@vmware.com

[lzhecheng]

/test-all /test-windows-all

[codecov-commenter]

Codecov Report

Merging #3043 (b938cfb) into main (db0c92d) will increase coverage by 0.41%.
The diff coverage is 0.00%.

@@            Coverage Diff             @@
##             main    #3043      +/-   ##
==========================================
+ Coverage   60.61%   61.03%   +0.41%     
==========================================
  Files         292      292              
  Lines       24708    24722      +14     
==========================================
+ Hits        14976    15088     +112     
+ Misses       8092     7994      -98     
  Partials     1640     1640              

Flag	Coverage Δ
kind-e2e-tests	48.24% <0.00%> (+0.49%)	⬆️
unit-tests	40.18% <0.00%> (+0.04%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
pkg/agent/agent.go	51.85% <0.00%> (-0.11%)	⬇️
pkg/apiserver/handlers/endpoint/handler.go	58.82% <0.00%> (-11.77%)	⬇️
...kg/agent/flowexporter/connections/conntrack_ovs.go	77.57% <0.00%> (-1.82%)	⬇️
pkg/controller/types/group.go	0.00% <0.00%> (ø)
pkg/agent/route/route_linux.go	46.82% <0.00%> (+0.14%)	⬆️
pkg/agent/cniserver/server.go	68.88% <0.00%> (+0.31%)	⬆️
pkg/agent/openflow/client.go	57.76% <0.00%> (+0.65%)	⬆️
pkg/agent/controller/networkpolicy/reconciler.go	78.03% <0.00%> (+1.04%)	⬆️
pkg/controller/networkpolicy/clustergroup.go	76.99% <0.00%> (+1.37%)	⬆️
... and 9 more

[tnqn]

LGTM, was there any issue caused by it?

[lzhecheng]

LGTM, was there any issue caused by it?

I don't think currently there's any antrea issue on it.

[tnqn]

LGTM, was there any issue caused by it?

I don't think currently there's any antrea issue on it.

In encap mode, when we access a Pod running on Node B from Node A, wouldn't the packet in maximum size be dropped because the encapsulated packet exceeds MTU?

[tnqn]

/test-integration
/test-e2e
/test-conformance

[lzhecheng]

LGTM, was there any issue caused by it?

I don't think currently there's any antrea issue on it.

In encap mode, when we access a Pod running on Node B from Node A, wouldn't the packet in maximum size be dropped because the encapsulated packet exceeds MTU?

To be clear, before I meant that there's no opening Github issue in Antrea repo about this problem. So I didn't link this PR to anything.

Yes, use UDP protocol and there will be drop.

[lzhecheng]

/test-e2e

[tnqn]

@lzhecheng You might have noticed that the commit messages and PR descriptions appears in weird format when you cherry-pick PRs. I guess you committed the changes on Windows platform and line endings are handled differently.

You can find that the commit message of your PRs don't have newlines:
https://patch-diff.githubusercontent.com/raw/antrea-io/antrea/pull/3043.patch
https://patch-diff.githubusercontent.com/raw/antrea-io/antrea/pull/2976.patch

Please try to set git config --global core.autocrlf true to convert line endings to Unix style.