post-rebase TODOs

[deads2k]

• UPSTREAM: <carry>: make namespace lifecycle access review aware should be fixed by discovery. If you can't get it, then it's not really stored. - Remove ns lifecycle carry #15255
• find all the timeout increase carries. Either make them configurable or bump them in kube. - move timeout increases upstream #15256
• UPSTREAM: <carry>: Disable file locking in clientcmd why would we do this. Seems like it gets removed upstream or we deal with it. - Remove UPSTREAM: <carry>: Disable file locking in clientcmd #15257
• UPSTREAM: <carry>: match kube rbac setup in e2es with openshift kinds drop when we move to RBAC authorizer - remove UPSTREAM: <carry>: match kube rbac setup in e2es with openshift kinds #15258
• UPSTREAM: 00000: disambiguate operation names for legacy discovery move upstream - Move upstream UPSTREAM: 00000: disambiguate operation names for legacy discovery  #15259
• UPSTREAM: <drop>: drop post 3.7 rebase. allows disabled aggregator force aggregator always on and remove this - Remove UPSTREAM: <drop>: drop post 3.7 rebase. allows disabled aggregator #15260
• UPSTREAM: <carry>: Add our types to kubectl get error should be handled generically. upstream needs to discover too - remove UPSTREAM: <carry>: Add our types to kubectl get error #15261
• UPSTREAM: <carry>: add kubelet timeouts move upstream - move timeout increases upstream #15256
• UPSTREAM: <carry>: Allow overriding default generators for run - just fix the upstream already - move UPSTREAM: <carry>: Allow overriding default generators for run #15262
• UPSTREAM: <carry>: fix fifo resync, remove after FIFO is dead fix this whereever it needs fixing - remove UPSTREAM: <carry>: fix fifo resync, remove after FIFO is dead #15263
• UPSTREAM: <carry>: Fix to avoid REST API calls at log level 2. - move upstream or drop it @aveshagarwal - remove UPSTREAM: <carry>: Fix to avoid REST API calls at log level 2. #15264
• UPSTREAM: <drop>: Set the log level for iptables rule dump to 5 - move upstream or drop it (ben bennet) - remove UPSTREAM: <drop>: Set the log level for iptables rule dump to 5 #15265
• remove reflector.Run carry. - remove reflector.Run carry #15266
• bump kube github.com/Azure/go-ansiterm - I think this is fixed when we move the registry out
• bump kube github.com/Sirupsen/logrus - make upstream logrus match our #15295
• bump kube github.com/cloudflare/cfssl/auth - bump "github.com/cloudflare/cfssl/auth" #15297
• bump kube github.com/coreos/go-semver/semver - bump "github.com/coreos/go-semver/semver" #15299
• bump kube github.com/docker/distribution/digest - I think this is fixed moving the registry out
• bump kube github.com/google/certificate-transparency/go - bump "github.com/google/certificate-transparency/go" #15300
• bump kube github.com/jonboulle/clockwork - bump "github.com/jonboulle/clockwork" #15301
• bump kube github.com/mailru/easyjson/buffer - bump "github.com/mailru/easyjson/buffer" #15302
• bump kube github.com/spf13/cobra - bump "github.com/spf13/cobra" #15303
• bump kube golang.org/x/sys/unix - I think this is a transitive from registry again
• bump kube golang.org/x/text/cases - bump "golang.org/x/text/cases" #15304
• bump kube gopkg.in/yaml.v2 - bump "gopkg.in/yaml.v2" #15305
• type assertions on all storage - type assertions on all reststorage #15268
• standardize GetAttrs - refactor GetAttrs for all RESTStorage #15267
• must remove UPSTREAM: : allow to use * as a capability in Security Context - Remove CapabilitiesAll from main api types #15273
• remove UPSTREAM: : keep old pod available - Remove UPSTREAM: <drop>: keep old pod available #15274
• use upstream namespace lifecycle and remove project lifecycle one - Use upstream namespace lifecycle controller, remove project lifecycle #15275
• move non build admission plugins from admission package - Move non-admission plugins out of pkg/build/admission #15276
• kill github.com/openshift/origin/pkg/client - Remove use of legacy client #15272
• fix remaining controllers - Sweep openshift controllers for "standard" construction #15277
• fix router metrics called out in pull. haproxy.go and router.go MustRegister - Restore router metrics using gauge #15279
• remove our shortcut expander. - Remove the origin shortcutexpander restmapper #15282
• use factory for building resourcebuidlers - Use NewBuilder function in the factory from upstream #15278
• remove NewProxyCommand - Remove NewProxyCommand #15286
• remove sdn's GetNodeCondition - remove sdn's GetNodeCondition #15283
• make GetStandardPrinter respect the factory - Fix the kubectl printing stack #15280
• make upstream tests enforce ownerref restrictions in admission - Upstream e2e tests need to enforce ownerref restrictions for admission #15281
• re-enable portworx - re-enable portworx #15269
• add test to check configapi defaulting - add test to check configapi defaulting #15287
• add integration tests for checking the ability to enable and disable versions - Add integration tests to ensure we can enable and disable versions #15285
• gofmt version update - fix disable go version checking #15270
• wire-up the new WithAudit filter. Now we do only WithLegacyAudit. - Wire up the new WithAudit filter #15271
• add new oc commands: "apiversions", "clusterinfo", "plugin", "resize", "rollingupdate", "run-container", "update", - Figure out why deprecated kube commands are showing up in our test. #15288
  • looks like these are all deprecated (@sttts)
• refactor SA token verifier for REST - Refactor the SA token verifier to use a REST client #15284
• make master config clientgo client: privilegedLoopbackKubeClientsetClientGo - cleanup master config clients and other #15293
• simplify createAPIExtensionsConfig to streamline it. no second copy for instance - streamline apiextensions wiring #15289
• decide what to do about new admission plugins. list in pkg/cmd/server/origin/admission/register.go - decide what to do about new admission plugins #15290
• reenable statefulset e2e tests: Re-enable stateful set tests #15252
• reenable node port service e2e tests: fix e2e test: Services should be able to create a functioning NodePort service #15253
• reenable schedulerpriorities e2e tests: fix e2e tests: SchedulerPriorities #15254
• reenable if _, err := url.Parse(o.RegistryUrlOverride); err != nil { in pkg/cmd/admin/prune/images.go - add back validation (maybe?) for disable registry url validation in prune #15291
• cleanup InfraPersistentVolumeRecyclerControllerServiceAccountName creation in pkg/cmd/server/start/start_master.go - cleanup creation of InfraPersistentVolumeRecyclerControllerServiceAccountName #15292
• re-enable unidler: Re-enable idler #15101

Mismatch on github.com/coreos/etcd/alarm:
    Origin:  61fc123e7a8b14a0a258aa3f5c4159861b1ec2e7 (2017-06-22 09:47:21 -0700, discontinuous)
    K8s:    *20490caaf0dcd96bb4a95e40625559def8ef5b04 (2017-03-27 10:20:28 -0700, discontinuous)
    To check out newest:
    cd /home/deads/workspaces/openshift3/src/github.com/coreos/etcd/alarm && git checkout 61fc123e7a8b14a0a258aa3f5c4159861b1ec2e7

[smarterclayton]

• cAdvisor metrics are no longer shown After upgrading to 1.7.0, Kubelet no longer reports cAdvisor stats kubernetes/kubernetes#48483

[liggitt]

bump coreos/etcd upstream to v3.2.1 - #15322

[smarterclayton]

Seeing this in the logs on a real data set: ``` E0718 22:37:20.981215 74010 cacher.go:274] unexpected ListAndWatch error: github.com/openshift/origin/vendor/k8s.io/apiserver/pkg/storage/cacher.go:215: Failed to list *autoscaling.HorizontalPodAutoscaler: no kind "HorizontalPodAutoscaler" is registered for version "extensions/v1beta1" I0718 22:37:21.982476 74010 reflector.go:236] Listing and watching *autoscaling.HorizontalPodAutoscaler from github.com/openshift/origin/vendor/k8s.io/apiserver/pkg/storage/cacher.go:215 ```

…

On Tue, Jul 18, 2017 at 10:07 PM, Jordan Liggitt ***@***.***> wrote: bump coreos/etcd upstream to v3.2.1 - #15322 <#15322> — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#14647 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ABG_p-ae_dc5Ns4Pmav3ATBZ4PLpR9N0ks5sPWTugaJpZM4N6Ndo> .

[liggitt]

did you run oadm migrate? HPAs are written to etcd as autoscaling.v1.

https://github.com/openshift/origin/blob/release-3.6/test/integration/etcd_storage_path_test.go#L506-L510

[smarterclayton]

Did we drop support for that in 1.6? Because this is restore from a backup onto 1.7

…

On Tue, Jul 18, 2017 at 10:41 PM, Jordan Liggitt ***@***.***> wrote: did you run oadm migrate? should be written to autoscaling v1. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#14647 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ABG_py-QSi6_IBlu-0Hy8aG1l3RICkiSks5sPWzRgaJpZM4N6Ndo> .

[liggitt]

dropped support in 1.7. upstream dropped it in 1.5, we carried it in 1.6 to allow the migrate

[smarterclayton]

Ok, something I can test then

…

On Tue, Jul 18, 2017 at 10:48 PM, Jordan Liggitt ***@***.***> wrote: dropped support in 1.7 — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#14647 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ABG_p5biSJnjTsVGtrbjGW3GdIzgYTK3ks5sPW6NgaJpZM4N6Ndo> .

[smarterclayton]

Attach looks broken on Mac @fabianofranz

$ oc run -it --attach --image="gcr.io/google_containers/hostexec:1.2" --restart=Never test
If you don't see a command prompt, try pressing enter.

I hit enter and nothing ever happens. Did we fix something like this in upstream?

If I hit CTRL+C I get:

If you don't see a command prompt, try pressing enter.

^C/bin/sh: can't open /fifo: Interrupted system call

[smarterclayton]

Which is novel.

[0xmichalis]

@deads2k all TODOs seem ticked