Currently if a pod is being scheduled with no meta.RolesKey label
attached to it, per convention the first configured mesos (framework)
role is being used.

This is quite limiting and also lets e2e tests fail. This commit
introduces a new configuration option "--mesos-default-pod-roles" defaulting to
"*" which defines the default pod roles in case the meta.RolesKey pod
label is missing.

Switch to versioned ListOptions in client.

Move Kubmark tests to us-east1-b, to make space for enormous cluster …

Fix queryparams convert

Auto commit by PR queue bot

Auto commit by PR queue bot

Auto commit by PR queue bot

Auto commit by PR queue bot

Increase limit for listing pods in 100-node cluster

Auto commit by PR queue bot

Auto commit by PR queue bot

Auto commit by PR queue bot

renaming labels.LabelSelector to labels.internalSelector

Auto commit by PR queue bot

Auto commit by PR queue bot

Auto commit by PR queue bot

Auto commit by PR queue bot

Auto commit by PR queue bot

Revert "[hpa] Parameterize tolerance, downscale, and upscale into HPAController, and add corresponding unit test for backsolved tolerance."

Revert "[hpa] Parameterize tolerance, downscale, and upscale into HPAController, and add corresponding unit test for backsolved tolerance."

Auto commit by PR queue bot

Auto commit by PR queue bot

Auto commit by PR queue bot

Auto commit by PR queue bot

Auto commit by PR queue bot

Auto commit by PR queue bot

Auto commit by PR queue bot

Add quiet mode to gcloud calls from upgrade e2e tests

Replace podFullName with format.Pod() in logging messages

Support collection deletion in apiserver.

Validation cleanup parts 5 & 6 together

fix testcase in pkg/client/leaderelection that wasn't testing what it said it was testing

PersistentVolume Provisioner Controller