In the case of multinode clusters, the http server pod and the test cluster can
spawn on different nodes, which can be problematic for poststart / prestop hooks,
as they are executed by the kubelet itself, and the cross-node lifecycle hook might
fail (according to the Kubernetes network model, it is not mandatory for kubelet to
be able to access pods on a different node).
This commit ensures that the test pod spawns on the same node as the http server pod.
If the user has not specified a pull policy we must assume a default of
v1.PullIfNotPresent.
Add some extra verbose output to help users monitor what policy is
used and what images are skipped / pulled.
Use "fallthrough" and case handle "v1.PullAlways".
Update unit test.
kubernetes#60525 introduced
Balanced attached node volumes feature gate to include volume
count for prioritizing nodes. The reason for introducing this
flag was its usefulness in Red Hat OpenShift Online environment
which is not being used any more. So, removing the flag
as it helps in maintainability of the scheduler code base
as mentioned at kubernetes#101489 (comment)
heketi that is dependency for glusterfs was updated to import a fork of jwt-go that fixes
a high severity CVE-2020-26160 by importing a forked repo (form3tech-oss/jwt-go) that fixes
the CVE as the original repo (dgrijalva/jwt-go) has a fix with breaking changes and
the repo is generally loosely maintained.
This fork is already being used in other places in kubernetes. Heketi 10.3.0 includes this fix.
NOTE: The vulnerable function is not used by heketi or kubernetes. So the vulnerable package is a
transitive dependency and this fix will reduce false postive CVE alerts on kubernetes
Bumped github.com/auth0/go-jwt-middleware to v1.0.0 which removes dgrijalva/jwt-go cleanly
Fixing test failures: pull-kubernetes-dependencies
Removing unused dependencies in replace block to fix test failures: pull-kubernetes-dependencies
Bump go-jwt-middleware to v1.0.1 to *not* pull in examples only deps
- verify memory manager data returned by `GetAllocatableResources`
- verify pod container memory manager data
Signed-off-by: Artyom Lukianov <alukiano@redhat.com>
The test was flaking because the test was creating more connections
than expected.
Disabling connection pooling removes the flakes, and no more connections
are created that the necessary.
`--command-timeout` is supported in `ETCDCTL_API` version 3 and it is NOT supported in `ETCDCTL_API` version 2. Up to etcd 3.3 (inclusive) default version of `ETCDCTL_API` was 2. This change explicitly defines used `ETCDCTL_API` and allows usage of etcd versions below 3.4.
