Under very heavy loads (e.g. 100 pods created at the same time), VM
based runtimes can take more than 10 seconds to create a pod.
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
When cri-o assumes the container creation failed, we need to let the
runtime know that we're bailing out so that it cancels all ongoing
operation.
In container creation timeout situations for example, failing to
explictly request the runtime for container deletion can lead to large
resource leaks as kubelet re-creates a failing container, while the
runtime finishes creating the previous one(s).
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
Set the exitsdir for kpod back to /var/run/crio... so kpod can benefit
from the container exit file.
Because 0 is the int32 blank value, kpod needs its own container state
struct with the omitempty removed so it can actually display 0 in
its default json output.
Signed-off-by: baude <bbaude@redhat.com>
When a user enters a CLI with a StringFlags or StringSliceFlags and does not add
a value the CLI mistakently takes the next option and uses it as a value.
This usually ends up with an error like not enough options or others. Some times
it could also succeed, with weird results. This patch looks for any values that
begin with a "-" and return an error.
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
* Remove duplicate definitions of storage-related flags for kpod, since
we set them in helpers.bash now, and the other locations that were
also setting it were doing so after loading the definitions in
helpers.
* Set kpod storage flags after checking if we need to force use of the
"vfs" storage driver for cri-o, to make sure kpod also ends up with
the same override if we're using one.
Signed-off-by: Nalin Dahyabhai <nalin@redhat.com>
Move kpod tests from kpod.bats to kpod_[commandname].bats
Also make sure all status checks have a echo $output before them.
Signed-off-by: Ryan Cole <rcyoalne@gmail.com>
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
Signed-off-by: umohnani8 <umohnani@redhat.com>
Packages are no longer available to build on RHEL and CentOS and
btrfs is not longer supported, so we should not build with it.
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
When running cri-tests with cri-o, I found out that cri-o panicked
immediately with the following message. Fix it by accessing to the
labels map only if it's non-nil.
```
panic: assignment to entry in nil map
goroutine 57 [running]:
.../cri-o/server.(*Server).RunPodSandbox(0xc42048e000, 0x7efcad4cd400,
0xc42066ec90, 0xc4201703d0, 0x0, 0x0, 0x0)
.../cri-o/server/sandbox_run.go:225 +0xda5
.../cri-o/vendor/k8s.io/kubernetes/pkg/kubelet/apis/cri/v1alpha1/runtime
._RuntimeService_RunPodSandbox_Handler(0x21793e0, 0xc42048e000,
0x7efcad4cd400, 0xc42066ec90, 0xc4204fe780, 0x0, 0x0, 0x0, 0x0, 0x0)
.../cri-o/vendor/k8s.io/kubernetes/pkg/kubelet/apis/cri/v1alpha1/runtime/api.pb.go:3645 +0x279
.../cri-o/vendor/google.golang.org/grpc.(*Server).processUnaryRPC(0xc420
09e3c0, 0x33e79c0, 0xc4203d1950, 0xc42080a000, 0xc4202bb980, 0x33b1d58,
0xc42066ec60, 0x0, 0x0)
.../cri-o/vendor/google.golang.org/grpc/server.go:638 +0x99c
```
Signed-off-by: Dongsu Park <dongsu@kinvolk.io>
Previously the wrapper changed to $WORKSPACE prior to executing
the ansible-playbook command. This has the unintended consequence of
preventing use of relative paths. Fix this by using absolute paths in
wrapper script instead of changing directories.
Signed-off-by: Chris Evich <cevich@redhat.com>
It's unsightly and hard to maintain collections of references and long
lists across multiple playbooks/include files. Centralize them all
in ``vars.yml``, then include that in all plays.
Minor: Update all files with a newline at the start and end.
Signed-off-by: Chris Evich <cevich@redhat.com>
Add a playbook to pull down the integration and e2e testing
logs/xml. By default they will appear in a 'artifacts' subdirectory
of wherever the ``results.yml`` playbook lives. If the ``$WORKSPACE``
env. var is set and non-empty, the subdirectory will be created
there instead.
Inside the ``artifacts`` directory, further sub-directories are created,
one for each subject's Ansible inventory name. Within those
sub-directories are all the collected logs from that host. In this way,
automation may simply archive the entire 'artifacts' directory to
capture the important log files.
(Depends on PR #935)
Signed-off-by: Chris Evich <cevich@redhat.com>
Processing node-e2e.log into jUnit format is insane, it's chock-full of
terminal escape codes. They would either need to be scraped/removed or
disabled somehow. Instead, take advantage of ``e2e.go --report-dir=``
option. This will cause it to store native jUnit results in the
specified directory for later collection. The jUnit results are also
needed for the google test grid.
Signed-off-by: Chris Evich <cevich@redhat.com>
When run by hand, it's much easier to spot things going wrong when
they're colored in red. Add an ansible.cfg to make that happen. This
also sets a default output log file (``$ARTIFACTS/main.log``) - that
doesn't contain color-codes.
When executing against multple hosts, the output can sometimes become
difficult to read, esp. with lots of async. tasks. The callback_plugin
script reorganizes how the console and log is organized, making it
clearer which host did what and when.
Signed-off-by: Chris Evich <cevich@redhat.com>
