update deploy file: use install_multus bin to update cni file

[41tair]

install cni initcontiainer can't execute sometimes

a sample error logs looks like:

cp: cannot create regular file '/host/opt/cni/bin/multus-shim': Text file busy

Fixes #1221

[coveralls]

coverage: 63.051% (-0.4%) from 63.432%
when pulling efa51cf on 41tair:fix/use-multus-install
into 3477c9c on k8snetworkplumbingwg:master.

[s1061123]

Hi @41tair, thank you for the PR!

Once we use install_multus for thick pluigin, then we could change container base image to distroless from

multus-cni/images/Dockerfile.thick

Line 10 in ba18cf5

FROM debian:stable-slim

to such like that as thin Dockerfile

multus-cni/images/Dockerfile

Line 11 in ba18cf5

FROM gcr.io/distroless/base-debian11:latest

Could you pelase take care of that?

[kfox1111]

Anything left to do on this? Think mode seems broken/risky without it.

[dougbtv]

LGTM! Thanks for the update

[kfox1111]

Reluctant to deploy a new instance of multus until this bug is fixed.

[dustinrouillard]

Reluctant to deploy a new instance of multus until this bug is fixed.

Would love to see this fixed.

[hansbogert]

as an end-user I'm wondering what's blocking this? I'm kinda stuck, 4.0.2 is non working for my cluster, but so is master, b/c of this: #1221

#1221 mentions that that issue should be fixed by this change

[iiiceoo]

/cc @s1061123 May I ask why this PR is blocked? It actually fixes a critical bug.

[sandj1438]

@dougbtv can we merge this and release a new image ? it is a critical bug fix right?

[dustinrouillard]

Hey maintainers, can we please get some action on this?!?!?!?!?!?!?!?! People are tired of manually patching and fixing their installs.. Merge. The. Fix.

[kub3let]

@s1061123 sorry to bother you but please give a statement about this, it's breaking a lot of clusters.

What has to be done to get this merged ?

[dougbtv]

Please join the maintainer's call (see k8snetworkplumbingwg/community github page for details) and express this stuff if it's causing you problems! we'd also love tests to prove that it works.

in the meanwhile -- we're re-running the tests in #1364 (and... Tomo also figured out how to kick the tests again, so, I appreciate it)

Note that the install yaml is meant as a basis to use for your own environments and comes with no guarantees (other than it passes the e2e tests, so we want to see it tested that way too) which is what slowed down this process and made it difficult to get maintainer feedback on. My apologies to those impacted by the usage of this yaml.

[kfox1111]

@dougbtv kubevirt/cluster-network-addons-operator#1810 is the same issue too. They seem to just be following this repo's lead and running into the same issue. :/