2015-03-31 04:53:30 +00:00
|
|
|
#!/bin/bash
|
|
|
|
|
Rework functional test runner
This simplifies and enhances the functional-test runner script for
much better interactive behaviour and to give us the ability to better
choose what is running in CI.
Firstly, I have split the image-output testing into a separate script.
This is not actually part of the functional testing of elements and is
both logically and functionally different. It currently does not run
in upstream CI because we don't have docker in the images. I have
nothing against it, but it can be it's own thing.
run_functests.sh is overhauled to have a useful interactive interface,
e.g.
---
$ ./run_functests.sh -h
run_functests.sh [-h] [-l] <test> <test> ...
-h : show this help
-l : list available tests
<test> : functional test to run
Special test 'all' will run all tests
$ ./run_functests.sh -l
The available functional tests are:
apt-sources/test-sources
debian/build-succeeds
fedora/build-succeeds
fedora/build-succeeds-f21
ironic-agent/build-succeeds-fedora
---
As described there, you can run a single test, a number of tests, the
default tests (as CI will do) or all tests. Running all tests is too
much for regular CI, but currently the only way to stop a low priority
test running, or temporarily pause is to remove it completely --
clearly sub-optimal (see I93c2990472e88ab3e5ff14db56b4ff1b4dd965ef).
There is nothing complicated about this, and to further simplify I
have merged the runner functions back into run_functests.sh which
remains a very modest ~150 lines, with most of that being argument
sanity. With that and the image-format cleanup, we can remove the
indirection of the 3 small library files.
For consistency, I have renamed the "dib_functions_test" (that tests
things from the dib functions library) with a run_* prefix.
Because the default list is the same as the current functional tests
run, this does not modify the status-quo. I plan to modify this,
however, to run fedora-minimal & centos-minimal tests in a future
change, as these are required to be stable for openstack ci.
Documentation is updated, and a README.rst is added in the tests
directory for discoverability.
Change-Id: I86d208bd34ff09a29fdb916a4e7ef740c7f65af8
2016-02-05 03:54:50 +00:00
|
|
|
set -eu
|
2015-03-31 04:53:30 +00:00
|
|
|
set -o pipefail
|
|
|
|
|
Rework functional test runner
This simplifies and enhances the functional-test runner script for
much better interactive behaviour and to give us the ability to better
choose what is running in CI.
Firstly, I have split the image-output testing into a separate script.
This is not actually part of the functional testing of elements and is
both logically and functionally different. It currently does not run
in upstream CI because we don't have docker in the images. I have
nothing against it, but it can be it's own thing.
run_functests.sh is overhauled to have a useful interactive interface,
e.g.
---
$ ./run_functests.sh -h
run_functests.sh [-h] [-l] <test> <test> ...
-h : show this help
-l : list available tests
<test> : functional test to run
Special test 'all' will run all tests
$ ./run_functests.sh -l
The available functional tests are:
apt-sources/test-sources
debian/build-succeeds
fedora/build-succeeds
fedora/build-succeeds-f21
ironic-agent/build-succeeds-fedora
---
As described there, you can run a single test, a number of tests, the
default tests (as CI will do) or all tests. Running all tests is too
much for regular CI, but currently the only way to stop a low priority
test running, or temporarily pause is to remove it completely --
clearly sub-optimal (see I93c2990472e88ab3e5ff14db56b4ff1b4dd965ef).
There is nothing complicated about this, and to further simplify I
have merged the runner functions back into run_functests.sh which
remains a very modest ~150 lines, with most of that being argument
sanity. With that and the image-format cleanup, we can remove the
indirection of the 3 small library files.
For consistency, I have renamed the "dib_functions_test" (that tests
things from the dib functions library) with a run_* prefix.
Because the default list is the same as the current functional tests
run, this does not modify the status-quo. I plan to modify this,
however, to run fedora-minimal & centos-minimal tests in a future
change, as these are required to be stable for openstack ci.
Documentation is updated, and a README.rst is added in the tests
directory for discoverability.
Change-Id: I86d208bd34ff09a29fdb916a4e7ef740c7f65af8
2016-02-05 03:54:50 +00:00
|
|
|
BASE_DIR=$(cd $(dirname "$0")/.. && pwd)
|
|
|
|
export DIB_ELEMENTS=$BASE_DIR/elements
|
|
|
|
export DIB_CMD=$BASE_DIR/bin/disk-image-create
|
2015-09-30 09:55:40 +00:00
|
|
|
|
Rework functional test runner
This simplifies and enhances the functional-test runner script for
much better interactive behaviour and to give us the ability to better
choose what is running in CI.
Firstly, I have split the image-output testing into a separate script.
This is not actually part of the functional testing of elements and is
both logically and functionally different. It currently does not run
in upstream CI because we don't have docker in the images. I have
nothing against it, but it can be it's own thing.
run_functests.sh is overhauled to have a useful interactive interface,
e.g.
---
$ ./run_functests.sh -h
run_functests.sh [-h] [-l] <test> <test> ...
-h : show this help
-l : list available tests
<test> : functional test to run
Special test 'all' will run all tests
$ ./run_functests.sh -l
The available functional tests are:
apt-sources/test-sources
debian/build-succeeds
fedora/build-succeeds
fedora/build-succeeds-f21
ironic-agent/build-succeeds-fedora
---
As described there, you can run a single test, a number of tests, the
default tests (as CI will do) or all tests. Running all tests is too
much for regular CI, but currently the only way to stop a low priority
test running, or temporarily pause is to remove it completely --
clearly sub-optimal (see I93c2990472e88ab3e5ff14db56b4ff1b4dd965ef).
There is nothing complicated about this, and to further simplify I
have merged the runner functions back into run_functests.sh which
remains a very modest ~150 lines, with most of that being argument
sanity. With that and the image-format cleanup, we can remove the
indirection of the 3 small library files.
For consistency, I have renamed the "dib_functions_test" (that tests
things from the dib functions library) with a run_* prefix.
Because the default list is the same as the current functional tests
run, this does not modify the status-quo. I plan to modify this,
however, to run fedora-minimal & centos-minimal tests in a future
change, as these are required to be stable for openstack ci.
Documentation is updated, and a README.rst is added in the tests
directory for discoverability.
Change-Id: I86d208bd34ff09a29fdb916a4e7ef740c7f65af8
2016-02-05 03:54:50 +00:00
|
|
|
#
|
|
|
|
# Default skip tests
|
|
|
|
#
|
|
|
|
# For time reasons, we do not run these tests by default; i.e. these
|
|
|
|
# tests are not run by "tox -e func" in the gate.
|
|
|
|
#
|
|
|
|
DEFAULT_SKIP_TESTS=(
|
|
|
|
# we run version pinned test in gate (this just runs latest)
|
|
|
|
fedora/build-succeeds
|
2016-04-07 05:12:51 +00:00
|
|
|
# in non-voting
|
|
|
|
gentoo/build-succeeds
|
2016-12-09 10:33:56 +00:00
|
|
|
opensuse/build-succeeds
|
2016-07-17 17:03:55 +00:00
|
|
|
# good to have the test case around - but because of changes
|
|
|
|
# in testing does not work always.
|
|
|
|
debian-minimal/testing-build-succeeds
|
2015-05-07 21:49:09 +00:00
|
|
|
# Currently failing due to bug in locale generation
|
|
|
|
centos-minimal/build-succeeds
|
2016-12-09 03:40:08 +00:00
|
|
|
# No longer reasonable to test upstream (lacks a mirror in infra)
|
|
|
|
centos/build-succeeds
|
Rework functional test runner
This simplifies and enhances the functional-test runner script for
much better interactive behaviour and to give us the ability to better
choose what is running in CI.
Firstly, I have split the image-output testing into a separate script.
This is not actually part of the functional testing of elements and is
both logically and functionally different. It currently does not run
in upstream CI because we don't have docker in the images. I have
nothing against it, but it can be it's own thing.
run_functests.sh is overhauled to have a useful interactive interface,
e.g.
---
$ ./run_functests.sh -h
run_functests.sh [-h] [-l] <test> <test> ...
-h : show this help
-l : list available tests
<test> : functional test to run
Special test 'all' will run all tests
$ ./run_functests.sh -l
The available functional tests are:
apt-sources/test-sources
debian/build-succeeds
fedora/build-succeeds
fedora/build-succeeds-f21
ironic-agent/build-succeeds-fedora
---
As described there, you can run a single test, a number of tests, the
default tests (as CI will do) or all tests. Running all tests is too
much for regular CI, but currently the only way to stop a low priority
test running, or temporarily pause is to remove it completely --
clearly sub-optimal (see I93c2990472e88ab3e5ff14db56b4ff1b4dd965ef).
There is nothing complicated about this, and to further simplify I
have merged the runner functions back into run_functests.sh which
remains a very modest ~150 lines, with most of that being argument
sanity. With that and the image-format cleanup, we can remove the
indirection of the 3 small library files.
For consistency, I have renamed the "dib_functions_test" (that tests
things from the dib functions library) with a run_* prefix.
Because the default list is the same as the current functional tests
run, this does not modify the status-quo. I plan to modify this,
however, to run fedora-minimal & centos-minimal tests in a future
change, as these are required to be stable for openstack ci.
Documentation is updated, and a README.rst is added in the tests
directory for discoverability.
Change-Id: I86d208bd34ff09a29fdb916a4e7ef740c7f65af8
2016-02-05 03:54:50 +00:00
|
|
|
)
|
|
|
|
|
2016-07-20 21:34:50 +00:00
|
|
|
function log_with_prefix {
|
|
|
|
local pr=$1
|
2016-10-24 00:17:28 +00:00
|
|
|
local log
|
2016-07-20 21:34:50 +00:00
|
|
|
|
|
|
|
while read a; do
|
2016-10-24 00:17:28 +00:00
|
|
|
log="[$pr] $a"
|
|
|
|
if [[ ${LOG_DATESTAMP} -ne 0 ]]; then
|
|
|
|
log="$(date +"%Y%m%d-%H%M%S.%N") ${log}"
|
|
|
|
fi
|
|
|
|
echo "${log}"
|
2016-07-20 21:34:50 +00:00
|
|
|
done
|
|
|
|
}
|
|
|
|
|
|
|
|
# Log job control messages
|
|
|
|
function log_jc {
|
|
|
|
local msg="$1"
|
2016-10-24 00:17:28 +00:00
|
|
|
local log="[JOB-CONTROL] ${msg}"
|
|
|
|
|
|
|
|
if [[ ${LOG_DATESTAMP} -ne 0 ]]; then
|
|
|
|
log="$(date +"%Y%m%d-%H%M%S.%N") ${log}"
|
|
|
|
fi
|
|
|
|
echo "${log}"
|
2016-07-20 21:34:50 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
function job_cnt {
|
|
|
|
running_jobs=$(jobs -p)
|
|
|
|
echo ${running_jobs} | wc -w
|
|
|
|
}
|
|
|
|
|
|
|
|
# This is needed, because the better 'wait -n' is
|
|
|
|
# available since bash 4.3 only.
|
|
|
|
function wait_minus_n {
|
|
|
|
if [ "${BASH_VERSINFO[0]}" -gt 4 \
|
|
|
|
-o "${BASH_VERSINFO[0]}" = 4 \
|
|
|
|
-a "${BASH_VERSINFO[1]}" -ge 3 ]; then
|
|
|
|
# Good way: wait on any job
|
|
|
|
wait -n
|
|
|
|
return $?
|
|
|
|
else
|
|
|
|
# Not that good way: wait on one specific job
|
|
|
|
# (others may be finished in the mean time)
|
|
|
|
local wait_for_pid=$(jobs -p | head -1)
|
|
|
|
wait ${wait_for_pid}
|
|
|
|
return $?
|
|
|
|
fi
|
|
|
|
}
|
|
|
|
|
Rework functional test runner
This simplifies and enhances the functional-test runner script for
much better interactive behaviour and to give us the ability to better
choose what is running in CI.
Firstly, I have split the image-output testing into a separate script.
This is not actually part of the functional testing of elements and is
both logically and functionally different. It currently does not run
in upstream CI because we don't have docker in the images. I have
nothing against it, but it can be it's own thing.
run_functests.sh is overhauled to have a useful interactive interface,
e.g.
---
$ ./run_functests.sh -h
run_functests.sh [-h] [-l] <test> <test> ...
-h : show this help
-l : list available tests
<test> : functional test to run
Special test 'all' will run all tests
$ ./run_functests.sh -l
The available functional tests are:
apt-sources/test-sources
debian/build-succeeds
fedora/build-succeeds
fedora/build-succeeds-f21
ironic-agent/build-succeeds-fedora
---
As described there, you can run a single test, a number of tests, the
default tests (as CI will do) or all tests. Running all tests is too
much for regular CI, but currently the only way to stop a low priority
test running, or temporarily pause is to remove it completely --
clearly sub-optimal (see I93c2990472e88ab3e5ff14db56b4ff1b4dd965ef).
There is nothing complicated about this, and to further simplify I
have merged the runner functions back into run_functests.sh which
remains a very modest ~150 lines, with most of that being argument
sanity. With that and the image-format cleanup, we can remove the
indirection of the 3 small library files.
For consistency, I have renamed the "dib_functions_test" (that tests
things from the dib functions library) with a run_* prefix.
Because the default list is the same as the current functional tests
run, this does not modify the status-quo. I plan to modify this,
however, to run fedora-minimal & centos-minimal tests in a future
change, as these are required to be stable for openstack ci.
Documentation is updated, and a README.rst is added in the tests
directory for discoverability.
Change-Id: I86d208bd34ff09a29fdb916a4e7ef740c7f65af8
2016-02-05 03:54:50 +00:00
|
|
|
# run_disk_element_test <test_element> <element>
|
|
|
|
# Run a disk-image-build .tar build of ELEMENT including any elements
|
|
|
|
# specified by TEST_ELEMENT
|
|
|
|
function run_disk_element_test() {
|
|
|
|
local test_element=$1
|
|
|
|
local element=$2
|
2016-07-20 21:34:50 +00:00
|
|
|
local dont_use_tmp=$3
|
|
|
|
local use_tmp_flag=""
|
Rework functional test runner
This simplifies and enhances the functional-test runner script for
much better interactive behaviour and to give us the ability to better
choose what is running in CI.
Firstly, I have split the image-output testing into a separate script.
This is not actually part of the functional testing of elements and is
both logically and functionally different. It currently does not run
in upstream CI because we don't have docker in the images. I have
nothing against it, but it can be it's own thing.
run_functests.sh is overhauled to have a useful interactive interface,
e.g.
---
$ ./run_functests.sh -h
run_functests.sh [-h] [-l] <test> <test> ...
-h : show this help
-l : list available tests
<test> : functional test to run
Special test 'all' will run all tests
$ ./run_functests.sh -l
The available functional tests are:
apt-sources/test-sources
debian/build-succeeds
fedora/build-succeeds
fedora/build-succeeds-f21
ironic-agent/build-succeeds-fedora
---
As described there, you can run a single test, a number of tests, the
default tests (as CI will do) or all tests. Running all tests is too
much for regular CI, but currently the only way to stop a low priority
test running, or temporarily pause is to remove it completely --
clearly sub-optimal (see I93c2990472e88ab3e5ff14db56b4ff1b4dd965ef).
There is nothing complicated about this, and to further simplify I
have merged the runner functions back into run_functests.sh which
remains a very modest ~150 lines, with most of that being argument
sanity. With that and the image-format cleanup, we can remove the
indirection of the 3 small library files.
For consistency, I have renamed the "dib_functions_test" (that tests
things from the dib functions library) with a run_* prefix.
Because the default list is the same as the current functional tests
run, this does not modify the status-quo. I plan to modify this,
however, to run fedora-minimal & centos-minimal tests in a future
change, as these are required to be stable for openstack ci.
Documentation is updated, and a README.rst is added in the tests
directory for discoverability.
Change-Id: I86d208bd34ff09a29fdb916a4e7ef740c7f65af8
2016-02-05 03:54:50 +00:00
|
|
|
local dest_dir=$(mktemp -d)
|
|
|
|
|
2016-07-20 21:34:50 +00:00
|
|
|
trap "rm -rf $dest_dir" EXIT
|
|
|
|
|
|
|
|
if [ "${dont_use_tmp}" = "yes" ]; then
|
|
|
|
use_tmp_flag="--no-tmpfs"
|
|
|
|
fi
|
Rework functional test runner
This simplifies and enhances the functional-test runner script for
much better interactive behaviour and to give us the ability to better
choose what is running in CI.
Firstly, I have split the image-output testing into a separate script.
This is not actually part of the functional testing of elements and is
both logically and functionally different. It currently does not run
in upstream CI because we don't have docker in the images. I have
nothing against it, but it can be it's own thing.
run_functests.sh is overhauled to have a useful interactive interface,
e.g.
---
$ ./run_functests.sh -h
run_functests.sh [-h] [-l] <test> <test> ...
-h : show this help
-l : list available tests
<test> : functional test to run
Special test 'all' will run all tests
$ ./run_functests.sh -l
The available functional tests are:
apt-sources/test-sources
debian/build-succeeds
fedora/build-succeeds
fedora/build-succeeds-f21
ironic-agent/build-succeeds-fedora
---
As described there, you can run a single test, a number of tests, the
default tests (as CI will do) or all tests. Running all tests is too
much for regular CI, but currently the only way to stop a low priority
test running, or temporarily pause is to remove it completely --
clearly sub-optimal (see I93c2990472e88ab3e5ff14db56b4ff1b4dd965ef).
There is nothing complicated about this, and to further simplify I
have merged the runner functions back into run_functests.sh which
remains a very modest ~150 lines, with most of that being argument
sanity. With that and the image-format cleanup, we can remove the
indirection of the 3 small library files.
For consistency, I have renamed the "dib_functions_test" (that tests
things from the dib functions library) with a run_* prefix.
Because the default list is the same as the current functional tests
run, this does not modify the status-quo. I plan to modify this,
however, to run fedora-minimal & centos-minimal tests in a future
change, as these are required to be stable for openstack ci.
Documentation is updated, and a README.rst is added in the tests
directory for discoverability.
Change-Id: I86d208bd34ff09a29fdb916a4e7ef740c7f65af8
2016-02-05 03:54:50 +00:00
|
|
|
|
|
|
|
if break="after-error" break_outside_target=1 \
|
2016-07-20 21:34:50 +00:00
|
|
|
break_cmd="cp -v \$TMP_MOUNT_PATH/tmp/dib-test-should-fail ${dest_dir} || true" \
|
2016-04-07 04:45:43 +00:00
|
|
|
DIB_SHOW_IMAGE_USAGE=1 \
|
Rework functional test runner
This simplifies and enhances the functional-test runner script for
much better interactive behaviour and to give us the ability to better
choose what is running in CI.
Firstly, I have split the image-output testing into a separate script.
This is not actually part of the functional testing of elements and is
both logically and functionally different. It currently does not run
in upstream CI because we don't have docker in the images. I have
nothing against it, but it can be it's own thing.
run_functests.sh is overhauled to have a useful interactive interface,
e.g.
---
$ ./run_functests.sh -h
run_functests.sh [-h] [-l] <test> <test> ...
-h : show this help
-l : list available tests
<test> : functional test to run
Special test 'all' will run all tests
$ ./run_functests.sh -l
The available functional tests are:
apt-sources/test-sources
debian/build-succeeds
fedora/build-succeeds
fedora/build-succeeds-f21
ironic-agent/build-succeeds-fedora
---
As described there, you can run a single test, a number of tests, the
default tests (as CI will do) or all tests. Running all tests is too
much for regular CI, but currently the only way to stop a low priority
test running, or temporarily pause is to remove it completely --
clearly sub-optimal (see I93c2990472e88ab3e5ff14db56b4ff1b4dd965ef).
There is nothing complicated about this, and to further simplify I
have merged the runner functions back into run_functests.sh which
remains a very modest ~150 lines, with most of that being argument
sanity. With that and the image-format cleanup, we can remove the
indirection of the 3 small library files.
For consistency, I have renamed the "dib_functions_test" (that tests
things from the dib functions library) with a run_* prefix.
Because the default list is the same as the current functional tests
run, this does not modify the status-quo. I plan to modify this,
however, to run fedora-minimal & centos-minimal tests in a future
change, as these are required to be stable for openstack ci.
Documentation is updated, and a README.rst is added in the tests
directory for discoverability.
Change-Id: I86d208bd34ff09a29fdb916a4e7ef740c7f65af8
2016-02-05 03:54:50 +00:00
|
|
|
ELEMENTS_PATH=$DIB_ELEMENTS:$DIB_ELEMENTS/$element/test-elements \
|
2016-07-20 21:34:50 +00:00
|
|
|
$DIB_CMD -x -t tar,qcow2 ${use_tmp_flag} -o $dest_dir/image -n $element $test_element 2>&1 \
|
|
|
|
| log_with_prefix "${element}/${test_element}"; then
|
2016-04-07 05:21:45 +00:00
|
|
|
|
|
|
|
if ! [ -f "$dest_dir/image.qcow2" ]; then
|
|
|
|
echo "Error: qcow2 build failed for element: $element, test-element: $test_element."
|
|
|
|
echo "No image $dest_dir/image.qcow2 found!"
|
|
|
|
fi
|
|
|
|
|
|
|
|
# check inside the tar for sentinel files
|
Rework functional test runner
This simplifies and enhances the functional-test runner script for
much better interactive behaviour and to give us the ability to better
choose what is running in CI.
Firstly, I have split the image-output testing into a separate script.
This is not actually part of the functional testing of elements and is
both logically and functionally different. It currently does not run
in upstream CI because we don't have docker in the images. I have
nothing against it, but it can be it's own thing.
run_functests.sh is overhauled to have a useful interactive interface,
e.g.
---
$ ./run_functests.sh -h
run_functests.sh [-h] [-l] <test> <test> ...
-h : show this help
-l : list available tests
<test> : functional test to run
Special test 'all' will run all tests
$ ./run_functests.sh -l
The available functional tests are:
apt-sources/test-sources
debian/build-succeeds
fedora/build-succeeds
fedora/build-succeeds-f21
ironic-agent/build-succeeds-fedora
---
As described there, you can run a single test, a number of tests, the
default tests (as CI will do) or all tests. Running all tests is too
much for regular CI, but currently the only way to stop a low priority
test running, or temporarily pause is to remove it completely --
clearly sub-optimal (see I93c2990472e88ab3e5ff14db56b4ff1b4dd965ef).
There is nothing complicated about this, and to further simplify I
have merged the runner functions back into run_functests.sh which
remains a very modest ~150 lines, with most of that being argument
sanity. With that and the image-format cleanup, we can remove the
indirection of the 3 small library files.
For consistency, I have renamed the "dib_functions_test" (that tests
things from the dib functions library) with a run_* prefix.
Because the default list is the same as the current functional tests
run, this does not modify the status-quo. I plan to modify this,
however, to run fedora-minimal & centos-minimal tests in a future
change, as these are required to be stable for openstack ci.
Documentation is updated, and a README.rst is added in the tests
directory for discoverability.
Change-Id: I86d208bd34ff09a29fdb916a4e7ef740c7f65af8
2016-02-05 03:54:50 +00:00
|
|
|
if ! [ -f "$dest_dir/image.tar" ]; then
|
|
|
|
echo "Error: Build failed for element: $element, test-element: $test_element."
|
|
|
|
echo "No image $dest_dir/image.tar found!"
|
|
|
|
exit 1
|
|
|
|
else
|
|
|
|
if tar -tf $dest_dir/image.tar | grep -q /tmp/dib-test-should-fail; then
|
|
|
|
echo "Error: Element: $element, test-element $test_element should have failed, but passed."
|
|
|
|
exit 1
|
|
|
|
else
|
|
|
|
echo "PASS: Element $element, test-element: $test_element"
|
|
|
|
fi
|
|
|
|
fi
|
|
|
|
else
|
2016-07-20 21:34:50 +00:00
|
|
|
if [ -f "${dest_dir}/dib-test-should-fail" ]; then
|
Rework functional test runner
This simplifies and enhances the functional-test runner script for
much better interactive behaviour and to give us the ability to better
choose what is running in CI.
Firstly, I have split the image-output testing into a separate script.
This is not actually part of the functional testing of elements and is
both logically and functionally different. It currently does not run
in upstream CI because we don't have docker in the images. I have
nothing against it, but it can be it's own thing.
run_functests.sh is overhauled to have a useful interactive interface,
e.g.
---
$ ./run_functests.sh -h
run_functests.sh [-h] [-l] <test> <test> ...
-h : show this help
-l : list available tests
<test> : functional test to run
Special test 'all' will run all tests
$ ./run_functests.sh -l
The available functional tests are:
apt-sources/test-sources
debian/build-succeeds
fedora/build-succeeds
fedora/build-succeeds-f21
ironic-agent/build-succeeds-fedora
---
As described there, you can run a single test, a number of tests, the
default tests (as CI will do) or all tests. Running all tests is too
much for regular CI, but currently the only way to stop a low priority
test running, or temporarily pause is to remove it completely --
clearly sub-optimal (see I93c2990472e88ab3e5ff14db56b4ff1b4dd965ef).
There is nothing complicated about this, and to further simplify I
have merged the runner functions back into run_functests.sh which
remains a very modest ~150 lines, with most of that being argument
sanity. With that and the image-format cleanup, we can remove the
indirection of the 3 small library files.
For consistency, I have renamed the "dib_functions_test" (that tests
things from the dib functions library) with a run_* prefix.
Because the default list is the same as the current functional tests
run, this does not modify the status-quo. I plan to modify this,
however, to run fedora-minimal & centos-minimal tests in a future
change, as these are required to be stable for openstack ci.
Documentation is updated, and a README.rst is added in the tests
directory for discoverability.
Change-Id: I86d208bd34ff09a29fdb916a4e7ef740c7f65af8
2016-02-05 03:54:50 +00:00
|
|
|
echo "PASS: Element $element, test-element: $test_element"
|
|
|
|
else
|
|
|
|
echo "Error: Build failed for element: $element, test-element: $test_element."
|
|
|
|
exit 1
|
|
|
|
fi
|
|
|
|
fi
|
|
|
|
|
|
|
|
trap EXIT
|
|
|
|
rm -rf $dest_dir /tmp/dib-test-should-fail
|
|
|
|
}
|
|
|
|
|
|
|
|
# run_ramdisk_element_test <test_element> <element>
|
|
|
|
# Run a disk-image-builder default build of ELEMENT including any
|
|
|
|
# elements specified by TEST_ELEMENT
|
|
|
|
function run_ramdisk_element_test() {
|
|
|
|
local test_element=$1
|
|
|
|
local element=$2
|
|
|
|
local dest_dir=$(mktemp -d)
|
|
|
|
|
|
|
|
if ELEMENTS_PATH=$DIB_ELEMENTS/$element/test-elements \
|
2016-07-20 21:34:50 +00:00
|
|
|
$DIB_CMD -x -o $dest_dir/image $element $test_element \
|
|
|
|
| log_with_prefix "${element}/${test_element}"; then
|
Rework functional test runner
This simplifies and enhances the functional-test runner script for
much better interactive behaviour and to give us the ability to better
choose what is running in CI.
Firstly, I have split the image-output testing into a separate script.
This is not actually part of the functional testing of elements and is
both logically and functionally different. It currently does not run
in upstream CI because we don't have docker in the images. I have
nothing against it, but it can be it's own thing.
run_functests.sh is overhauled to have a useful interactive interface,
e.g.
---
$ ./run_functests.sh -h
run_functests.sh [-h] [-l] <test> <test> ...
-h : show this help
-l : list available tests
<test> : functional test to run
Special test 'all' will run all tests
$ ./run_functests.sh -l
The available functional tests are:
apt-sources/test-sources
debian/build-succeeds
fedora/build-succeeds
fedora/build-succeeds-f21
ironic-agent/build-succeeds-fedora
---
As described there, you can run a single test, a number of tests, the
default tests (as CI will do) or all tests. Running all tests is too
much for regular CI, but currently the only way to stop a low priority
test running, or temporarily pause is to remove it completely --
clearly sub-optimal (see I93c2990472e88ab3e5ff14db56b4ff1b4dd965ef).
There is nothing complicated about this, and to further simplify I
have merged the runner functions back into run_functests.sh which
remains a very modest ~150 lines, with most of that being argument
sanity. With that and the image-format cleanup, we can remove the
indirection of the 3 small library files.
For consistency, I have renamed the "dib_functions_test" (that tests
things from the dib functions library) with a run_* prefix.
Because the default list is the same as the current functional tests
run, this does not modify the status-quo. I plan to modify this,
however, to run fedora-minimal & centos-minimal tests in a future
change, as these are required to be stable for openstack ci.
Documentation is updated, and a README.rst is added in the tests
directory for discoverability.
Change-Id: I86d208bd34ff09a29fdb916a4e7ef740c7f65af8
2016-02-05 03:54:50 +00:00
|
|
|
# TODO(dtantsur): test also kernel presence once we sort out its naming
|
|
|
|
# problem (vmlinuz vs kernel)
|
|
|
|
if ! [ -f "$dest_dir/image.initramfs" ]; then
|
|
|
|
echo "Error: Build failed for element: $element, test-element: $test_element."
|
|
|
|
echo "No image $dest_dir/image.initramfs found!"
|
|
|
|
exit 1
|
|
|
|
else
|
|
|
|
echo "PASS: Element $element, test-element: $test_element"
|
|
|
|
fi
|
|
|
|
else
|
|
|
|
echo "Error: Build failed for element: $element, test-element: $test_element."
|
|
|
|
exit 1
|
|
|
|
fi
|
|
|
|
}
|
|
|
|
|
|
|
|
#
|
|
|
|
# run_functests.sh
|
|
|
|
# run the functional tests for dib elements
|
|
|
|
#
|
|
|
|
|
|
|
|
# find elements that have functional test elements. TESTS will be an
|
|
|
|
# array with each value being "element/test-element"
|
|
|
|
TESTS=()
|
|
|
|
for e in $DIB_ELEMENTS/*/test-elements/*; do
|
|
|
|
test_element=$(echo $e | awk 'BEGIN {FS="/"}{print $NF}')
|
|
|
|
element=$(echo $e | awk 'BEGIN {FS="/"}{print $(NF-2)}')
|
|
|
|
TESTS+=("$element/$test_element")
|
|
|
|
done
|
|
|
|
|
2016-10-24 00:17:28 +00:00
|
|
|
#
|
|
|
|
# Default values
|
|
|
|
#
|
2016-07-20 21:34:50 +00:00
|
|
|
JOB_MAX_CNT=1
|
2016-10-24 00:17:28 +00:00
|
|
|
LOG_DATESTAMP=0
|
2016-07-20 21:34:50 +00:00
|
|
|
|
2016-10-24 00:17:28 +00:00
|
|
|
#
|
|
|
|
# Parse args
|
|
|
|
#
|
|
|
|
while getopts ":hlj:t" opt; do
|
Rework functional test runner
This simplifies and enhances the functional-test runner script for
much better interactive behaviour and to give us the ability to better
choose what is running in CI.
Firstly, I have split the image-output testing into a separate script.
This is not actually part of the functional testing of elements and is
both logically and functionally different. It currently does not run
in upstream CI because we don't have docker in the images. I have
nothing against it, but it can be it's own thing.
run_functests.sh is overhauled to have a useful interactive interface,
e.g.
---
$ ./run_functests.sh -h
run_functests.sh [-h] [-l] <test> <test> ...
-h : show this help
-l : list available tests
<test> : functional test to run
Special test 'all' will run all tests
$ ./run_functests.sh -l
The available functional tests are:
apt-sources/test-sources
debian/build-succeeds
fedora/build-succeeds
fedora/build-succeeds-f21
ironic-agent/build-succeeds-fedora
---
As described there, you can run a single test, a number of tests, the
default tests (as CI will do) or all tests. Running all tests is too
much for regular CI, but currently the only way to stop a low priority
test running, or temporarily pause is to remove it completely --
clearly sub-optimal (see I93c2990472e88ab3e5ff14db56b4ff1b4dd965ef).
There is nothing complicated about this, and to further simplify I
have merged the runner functions back into run_functests.sh which
remains a very modest ~150 lines, with most of that being argument
sanity. With that and the image-format cleanup, we can remove the
indirection of the 3 small library files.
For consistency, I have renamed the "dib_functions_test" (that tests
things from the dib functions library) with a run_* prefix.
Because the default list is the same as the current functional tests
run, this does not modify the status-quo. I plan to modify this,
however, to run fedora-minimal & centos-minimal tests in a future
change, as these are required to be stable for openstack ci.
Documentation is updated, and a README.rst is added in the tests
directory for discoverability.
Change-Id: I86d208bd34ff09a29fdb916a4e7ef740c7f65af8
2016-02-05 03:54:50 +00:00
|
|
|
case $opt in
|
|
|
|
h)
|
|
|
|
echo "run_functests.sh [-h] [-l] <test> <test> ..."
|
|
|
|
echo " -h : show this help"
|
|
|
|
echo " -l : list available tests"
|
2016-10-24 00:17:28 +00:00
|
|
|
echo " -j : parallel job count (default to 1)"
|
|
|
|
echo " -t : prefix log messages with timestamp"
|
Rework functional test runner
This simplifies and enhances the functional-test runner script for
much better interactive behaviour and to give us the ability to better
choose what is running in CI.
Firstly, I have split the image-output testing into a separate script.
This is not actually part of the functional testing of elements and is
both logically and functionally different. It currently does not run
in upstream CI because we don't have docker in the images. I have
nothing against it, but it can be it's own thing.
run_functests.sh is overhauled to have a useful interactive interface,
e.g.
---
$ ./run_functests.sh -h
run_functests.sh [-h] [-l] <test> <test> ...
-h : show this help
-l : list available tests
<test> : functional test to run
Special test 'all' will run all tests
$ ./run_functests.sh -l
The available functional tests are:
apt-sources/test-sources
debian/build-succeeds
fedora/build-succeeds
fedora/build-succeeds-f21
ironic-agent/build-succeeds-fedora
---
As described there, you can run a single test, a number of tests, the
default tests (as CI will do) or all tests. Running all tests is too
much for regular CI, but currently the only way to stop a low priority
test running, or temporarily pause is to remove it completely --
clearly sub-optimal (see I93c2990472e88ab3e5ff14db56b4ff1b4dd965ef).
There is nothing complicated about this, and to further simplify I
have merged the runner functions back into run_functests.sh which
remains a very modest ~150 lines, with most of that being argument
sanity. With that and the image-format cleanup, we can remove the
indirection of the 3 small library files.
For consistency, I have renamed the "dib_functions_test" (that tests
things from the dib functions library) with a run_* prefix.
Because the default list is the same as the current functional tests
run, this does not modify the status-quo. I plan to modify this,
however, to run fedora-minimal & centos-minimal tests in a future
change, as these are required to be stable for openstack ci.
Documentation is updated, and a README.rst is added in the tests
directory for discoverability.
Change-Id: I86d208bd34ff09a29fdb916a4e7ef740c7f65af8
2016-02-05 03:54:50 +00:00
|
|
|
echo " <test> : functional test to run"
|
|
|
|
echo " Special test 'all' will run all tests"
|
|
|
|
exit 0
|
|
|
|
;;
|
|
|
|
l)
|
|
|
|
echo "The available functional tests are:"
|
|
|
|
echo
|
|
|
|
for t in ${TESTS[@]}; do
|
|
|
|
echo " $t"
|
|
|
|
done
|
|
|
|
echo
|
|
|
|
exit 0
|
|
|
|
;;
|
2016-07-20 21:34:50 +00:00
|
|
|
j)
|
|
|
|
JOB_MAX_CNT=${OPTARG}
|
|
|
|
echo "Running parallel - using [${JOB_MAX_CNT}] jobs"
|
|
|
|
;;
|
2016-10-24 00:17:28 +00:00
|
|
|
t)
|
|
|
|
LOG_DATESTAMP=1
|
|
|
|
;;
|
Rework functional test runner
This simplifies and enhances the functional-test runner script for
much better interactive behaviour and to give us the ability to better
choose what is running in CI.
Firstly, I have split the image-output testing into a separate script.
This is not actually part of the functional testing of elements and is
both logically and functionally different. It currently does not run
in upstream CI because we don't have docker in the images. I have
nothing against it, but it can be it's own thing.
run_functests.sh is overhauled to have a useful interactive interface,
e.g.
---
$ ./run_functests.sh -h
run_functests.sh [-h] [-l] <test> <test> ...
-h : show this help
-l : list available tests
<test> : functional test to run
Special test 'all' will run all tests
$ ./run_functests.sh -l
The available functional tests are:
apt-sources/test-sources
debian/build-succeeds
fedora/build-succeeds
fedora/build-succeeds-f21
ironic-agent/build-succeeds-fedora
---
As described there, you can run a single test, a number of tests, the
default tests (as CI will do) or all tests. Running all tests is too
much for regular CI, but currently the only way to stop a low priority
test running, or temporarily pause is to remove it completely --
clearly sub-optimal (see I93c2990472e88ab3e5ff14db56b4ff1b4dd965ef).
There is nothing complicated about this, and to further simplify I
have merged the runner functions back into run_functests.sh which
remains a very modest ~150 lines, with most of that being argument
sanity. With that and the image-format cleanup, we can remove the
indirection of the 3 small library files.
For consistency, I have renamed the "dib_functions_test" (that tests
things from the dib functions library) with a run_* prefix.
Because the default list is the same as the current functional tests
run, this does not modify the status-quo. I plan to modify this,
however, to run fedora-minimal & centos-minimal tests in a future
change, as these are required to be stable for openstack ci.
Documentation is updated, and a README.rst is added in the tests
directory for discoverability.
Change-Id: I86d208bd34ff09a29fdb916a4e7ef740c7f65af8
2016-02-05 03:54:50 +00:00
|
|
|
\?)
|
|
|
|
echo "Invalid option: -$OPTARG"
|
|
|
|
exit 1
|
|
|
|
;;
|
|
|
|
esac
|
|
|
|
done
|
|
|
|
shift $((OPTIND-1))
|
|
|
|
|
2016-07-20 21:34:50 +00:00
|
|
|
DONT_USE_TMP="no"
|
|
|
|
if [ "${JOB_MAX_CNT}" -gt 1 ]; then
|
|
|
|
# switch off using tmp dir for image building
|
|
|
|
# (The mem check using the tmp dir is currently done
|
|
|
|
# based on the available memory - and not on the free.
|
|
|
|
# See #1618124 for more details)
|
|
|
|
DONT_USE_TMP="yes"
|
|
|
|
fi
|
|
|
|
|
Rework functional test runner
This simplifies and enhances the functional-test runner script for
much better interactive behaviour and to give us the ability to better
choose what is running in CI.
Firstly, I have split the image-output testing into a separate script.
This is not actually part of the functional testing of elements and is
both logically and functionally different. It currently does not run
in upstream CI because we don't have docker in the images. I have
nothing against it, but it can be it's own thing.
run_functests.sh is overhauled to have a useful interactive interface,
e.g.
---
$ ./run_functests.sh -h
run_functests.sh [-h] [-l] <test> <test> ...
-h : show this help
-l : list available tests
<test> : functional test to run
Special test 'all' will run all tests
$ ./run_functests.sh -l
The available functional tests are:
apt-sources/test-sources
debian/build-succeeds
fedora/build-succeeds
fedora/build-succeeds-f21
ironic-agent/build-succeeds-fedora
---
As described there, you can run a single test, a number of tests, the
default tests (as CI will do) or all tests. Running all tests is too
much for regular CI, but currently the only way to stop a low priority
test running, or temporarily pause is to remove it completely --
clearly sub-optimal (see I93c2990472e88ab3e5ff14db56b4ff1b4dd965ef).
There is nothing complicated about this, and to further simplify I
have merged the runner functions back into run_functests.sh which
remains a very modest ~150 lines, with most of that being argument
sanity. With that and the image-format cleanup, we can remove the
indirection of the 3 small library files.
For consistency, I have renamed the "dib_functions_test" (that tests
things from the dib functions library) with a run_* prefix.
Because the default list is the same as the current functional tests
run, this does not modify the status-quo. I plan to modify this,
however, to run fedora-minimal & centos-minimal tests in a future
change, as these are required to be stable for openstack ci.
Documentation is updated, and a README.rst is added in the tests
directory for discoverability.
Change-Id: I86d208bd34ff09a29fdb916a4e7ef740c7f65af8
2016-02-05 03:54:50 +00:00
|
|
|
# cull the list of tests to run into TESTS_TO_RUN
|
|
|
|
TESTS_TO_RUN=()
|
|
|
|
title=""
|
|
|
|
if [[ -z "$@" ]]; then
|
|
|
|
# remove the skipped tests
|
|
|
|
title="Running default tests:"
|
|
|
|
for test in "${TESTS[@]}"; do
|
|
|
|
if [[ " ${DEFAULT_SKIP_TESTS[@]} " =~ " ${test} " ]]; then
|
|
|
|
continue
|
|
|
|
else
|
|
|
|
TESTS_TO_RUN+=("${test}")
|
|
|
|
fi
|
|
|
|
done
|
|
|
|
elif [[ $1 == "all" ]]; then
|
|
|
|
title="Running all tests:"
|
|
|
|
TESTS_TO_RUN=("${TESTS[@]}")
|
|
|
|
else
|
|
|
|
title="Running specified tests:"
|
|
|
|
for test in $@; do
|
|
|
|
if [[ ! " ${TESTS[@]} " =~ " ${test} " ]]; then
|
|
|
|
echo "${test} : not a known test (see -l)"
|
|
|
|
exit 1
|
|
|
|
fi
|
|
|
|
TESTS_TO_RUN+=("${test}")
|
|
|
|
done
|
2015-09-30 09:55:40 +00:00
|
|
|
fi
|
Rework functional test runner
This simplifies and enhances the functional-test runner script for
much better interactive behaviour and to give us the ability to better
choose what is running in CI.
Firstly, I have split the image-output testing into a separate script.
This is not actually part of the functional testing of elements and is
both logically and functionally different. It currently does not run
in upstream CI because we don't have docker in the images. I have
nothing against it, but it can be it's own thing.
run_functests.sh is overhauled to have a useful interactive interface,
e.g.
---
$ ./run_functests.sh -h
run_functests.sh [-h] [-l] <test> <test> ...
-h : show this help
-l : list available tests
<test> : functional test to run
Special test 'all' will run all tests
$ ./run_functests.sh -l
The available functional tests are:
apt-sources/test-sources
debian/build-succeeds
fedora/build-succeeds
fedora/build-succeeds-f21
ironic-agent/build-succeeds-fedora
---
As described there, you can run a single test, a number of tests, the
default tests (as CI will do) or all tests. Running all tests is too
much for regular CI, but currently the only way to stop a low priority
test running, or temporarily pause is to remove it completely --
clearly sub-optimal (see I93c2990472e88ab3e5ff14db56b4ff1b4dd965ef).
There is nothing complicated about this, and to further simplify I
have merged the runner functions back into run_functests.sh which
remains a very modest ~150 lines, with most of that being argument
sanity. With that and the image-format cleanup, we can remove the
indirection of the 3 small library files.
For consistency, I have renamed the "dib_functions_test" (that tests
things from the dib functions library) with a run_* prefix.
Because the default list is the same as the current functional tests
run, this does not modify the status-quo. I plan to modify this,
however, to run fedora-minimal & centos-minimal tests in a future
change, as these are required to be stable for openstack ci.
Documentation is updated, and a README.rst is added in the tests
directory for discoverability.
Change-Id: I86d208bd34ff09a29fdb916a4e7ef740c7f65af8
2016-02-05 03:54:50 +00:00
|
|
|
|
|
|
|
# print a little status info
|
|
|
|
echo "------"
|
|
|
|
echo ${title}
|
|
|
|
for test in "${TESTS_TO_RUN[@]}"; do
|
|
|
|
echo " ${test}"
|
|
|
|
done
|
|
|
|
echo "------"
|
|
|
|
|
2016-07-20 21:34:50 +00:00
|
|
|
function wait_and_exit_on_failure {
|
|
|
|
local pid=$1
|
|
|
|
|
|
|
|
wait ${pid}
|
|
|
|
result=$?
|
|
|
|
|
|
|
|
if [ "${result}" -ne 0 ]; then
|
|
|
|
exit ${result}
|
|
|
|
fi
|
|
|
|
return 0
|
|
|
|
}
|
|
|
|
|
|
|
|
EXIT_CODE=0
|
Rework functional test runner
This simplifies and enhances the functional-test runner script for
much better interactive behaviour and to give us the ability to better
choose what is running in CI.
Firstly, I have split the image-output testing into a separate script.
This is not actually part of the functional testing of elements and is
both logically and functionally different. It currently does not run
in upstream CI because we don't have docker in the images. I have
nothing against it, but it can be it's own thing.
run_functests.sh is overhauled to have a useful interactive interface,
e.g.
---
$ ./run_functests.sh -h
run_functests.sh [-h] [-l] <test> <test> ...
-h : show this help
-l : list available tests
<test> : functional test to run
Special test 'all' will run all tests
$ ./run_functests.sh -l
The available functional tests are:
apt-sources/test-sources
debian/build-succeeds
fedora/build-succeeds
fedora/build-succeeds-f21
ironic-agent/build-succeeds-fedora
---
As described there, you can run a single test, a number of tests, the
default tests (as CI will do) or all tests. Running all tests is too
much for regular CI, but currently the only way to stop a low priority
test running, or temporarily pause is to remove it completely --
clearly sub-optimal (see I93c2990472e88ab3e5ff14db56b4ff1b4dd965ef).
There is nothing complicated about this, and to further simplify I
have merged the runner functions back into run_functests.sh which
remains a very modest ~150 lines, with most of that being argument
sanity. With that and the image-format cleanup, we can remove the
indirection of the 3 small library files.
For consistency, I have renamed the "dib_functions_test" (that tests
things from the dib functions library) with a run_* prefix.
Because the default list is the same as the current functional tests
run, this does not modify the status-quo. I plan to modify this,
however, to run fedora-minimal & centos-minimal tests in a future
change, as these are required to be stable for openstack ci.
Documentation is updated, and a README.rst is added in the tests
directory for discoverability.
Change-Id: I86d208bd34ff09a29fdb916a4e7ef740c7f65af8
2016-02-05 03:54:50 +00:00
|
|
|
for test in "${TESTS_TO_RUN[@]}"; do
|
2016-07-20 21:34:50 +00:00
|
|
|
running_jobs_cnt=$(job_cnt)
|
|
|
|
log_jc "Number of running jobs [${running_jobs_cnt}] max jobs [${JOB_MAX_CNT}]"
|
|
|
|
if [ "${running_jobs_cnt}" -ge "${JOB_MAX_CNT}" ]; then
|
|
|
|
log_jc "Waiting for job to finish"
|
|
|
|
wait_minus_n
|
|
|
|
result=$?
|
|
|
|
|
|
|
|
if [ "${result}" -ne 0 ]; then
|
|
|
|
EXIT_CODE=1
|
|
|
|
# If a job fails, do not start any new ones.
|
|
|
|
break
|
|
|
|
fi
|
|
|
|
fi
|
|
|
|
|
|
|
|
log_jc "Starting new job"
|
|
|
|
|
Rework functional test runner
This simplifies and enhances the functional-test runner script for
much better interactive behaviour and to give us the ability to better
choose what is running in CI.
Firstly, I have split the image-output testing into a separate script.
This is not actually part of the functional testing of elements and is
both logically and functionally different. It currently does not run
in upstream CI because we don't have docker in the images. I have
nothing against it, but it can be it's own thing.
run_functests.sh is overhauled to have a useful interactive interface,
e.g.
---
$ ./run_functests.sh -h
run_functests.sh [-h] [-l] <test> <test> ...
-h : show this help
-l : list available tests
<test> : functional test to run
Special test 'all' will run all tests
$ ./run_functests.sh -l
The available functional tests are:
apt-sources/test-sources
debian/build-succeeds
fedora/build-succeeds
fedora/build-succeeds-f21
ironic-agent/build-succeeds-fedora
---
As described there, you can run a single test, a number of tests, the
default tests (as CI will do) or all tests. Running all tests is too
much for regular CI, but currently the only way to stop a low priority
test running, or temporarily pause is to remove it completely --
clearly sub-optimal (see I93c2990472e88ab3e5ff14db56b4ff1b4dd965ef).
There is nothing complicated about this, and to further simplify I
have merged the runner functions back into run_functests.sh which
remains a very modest ~150 lines, with most of that being argument
sanity. With that and the image-format cleanup, we can remove the
indirection of the 3 small library files.
For consistency, I have renamed the "dib_functions_test" (that tests
things from the dib functions library) with a run_* prefix.
Because the default list is the same as the current functional tests
run, this does not modify the status-quo. I plan to modify this,
however, to run fedora-minimal & centos-minimal tests in a future
change, as these are required to be stable for openstack ci.
Documentation is updated, and a README.rst is added in the tests
directory for discoverability.
Change-Id: I86d208bd34ff09a29fdb916a4e7ef740c7f65af8
2016-02-05 03:54:50 +00:00
|
|
|
# from above; each array value is element/test_element. split it
|
|
|
|
# back up
|
|
|
|
element=${test%/*}
|
|
|
|
test_element=${test#*/}
|
|
|
|
|
|
|
|
# tests default to disk-based, but "element-type" can optionally
|
|
|
|
# override that
|
|
|
|
element_type=disk
|
|
|
|
element_type_override=$DIB_ELEMENTS/${element}/test-elements/${test_element}/element-type
|
|
|
|
if [ -f ${element_type_override} ]; then
|
|
|
|
element_type=$(cat ${element_type_override})
|
|
|
|
fi
|
|
|
|
|
|
|
|
echo "Running $test ($element_type)"
|
2016-07-20 21:34:50 +00:00
|
|
|
run_${element_type}_element_test $test_element $element ${DONT_USE_TMP} &
|
Rework functional test runner
This simplifies and enhances the functional-test runner script for
much better interactive behaviour and to give us the ability to better
choose what is running in CI.
Firstly, I have split the image-output testing into a separate script.
This is not actually part of the functional testing of elements and is
both logically and functionally different. It currently does not run
in upstream CI because we don't have docker in the images. I have
nothing against it, but it can be it's own thing.
run_functests.sh is overhauled to have a useful interactive interface,
e.g.
---
$ ./run_functests.sh -h
run_functests.sh [-h] [-l] <test> <test> ...
-h : show this help
-l : list available tests
<test> : functional test to run
Special test 'all' will run all tests
$ ./run_functests.sh -l
The available functional tests are:
apt-sources/test-sources
debian/build-succeeds
fedora/build-succeeds
fedora/build-succeeds-f21
ironic-agent/build-succeeds-fedora
---
As described there, you can run a single test, a number of tests, the
default tests (as CI will do) or all tests. Running all tests is too
much for regular CI, but currently the only way to stop a low priority
test running, or temporarily pause is to remove it completely --
clearly sub-optimal (see I93c2990472e88ab3e5ff14db56b4ff1b4dd965ef).
There is nothing complicated about this, and to further simplify I
have merged the runner functions back into run_functests.sh which
remains a very modest ~150 lines, with most of that being argument
sanity. With that and the image-format cleanup, we can remove the
indirection of the 3 small library files.
For consistency, I have renamed the "dib_functions_test" (that tests
things from the dib functions library) with a run_* prefix.
Because the default list is the same as the current functional tests
run, this does not modify the status-quo. I plan to modify this,
however, to run fedora-minimal & centos-minimal tests in a future
change, as these are required to be stable for openstack ci.
Documentation is updated, and a README.rst is added in the tests
directory for discoverability.
Change-Id: I86d208bd34ff09a29fdb916a4e7ef740c7f65af8
2016-02-05 03:54:50 +00:00
|
|
|
done
|
2015-03-31 04:53:30 +00:00
|
|
|
|
2016-07-20 21:34:50 +00:00
|
|
|
# Wait for the rest of the jobs
|
|
|
|
while true; do
|
|
|
|
running_jobs_cnt=$(job_cnt)
|
|
|
|
log_jc "Number of running jobs left [${running_jobs_cnt}]"
|
|
|
|
|
|
|
|
if [ "${running_jobs_cnt}" -eq 0 ]; then
|
|
|
|
break;
|
|
|
|
fi
|
|
|
|
|
|
|
|
wait_minus_n
|
|
|
|
result=$?
|
|
|
|
|
|
|
|
if [ "${result}" -ne 0 ]; then
|
|
|
|
EXIT_CODE=1
|
|
|
|
fi
|
|
|
|
done
|
|
|
|
|
|
|
|
if [ "${EXIT_CODE}" -eq 0 ]; then
|
|
|
|
echo "Tests passed!"
|
|
|
|
exit 0
|
|
|
|
else
|
|
|
|
echo "At least one test failed"
|
|
|
|
exit 1
|
|
|
|
fi
|