summaryrefslogtreecommitdiffstats
AgeCommit message (Collapse)AuthorFilesLines
2016-09-02Merge "Add print-to-a-file option when executing a command in functest_utils"jose.lausuch1-4/+11
2016-09-02Merge "Change default memory for Promise flavor"jose.lausuch1-1/+1
2016-09-02Merge "Fixes cleaning up HA Neutron objects"Jose Lausuch1-0/+11
2016-09-02Merge "Force python-keystoneclient to version 2.3.1 for tacker to work"jose.lausuch1-0/+1
2016-09-01Fixes cleaning up HA Neutron objectsTim Rozet1-0/+11
This fixes a race condition where the cleanup method is trying to remove HA networks, while Neutron is also trying to remove the network - after the HA router is removed. This only applies to installations where Neutron L3 HA is enabled (nosdn cases only). JIRA: FUNCTEST-454 Change-Id: If06b6f6297e99dcce4b0f2aff8ec80b47488ee90 Signed-off-by: Tim Rozet <trozet@redhat.com>
2016-09-01Change debug mode for commands that generate too much output in sec_scanjose.lausuch1-6/+7
The security scan test case generates sometimes too much output messages and makes Jenkins unreadable. For example: https://build.opnfv.org/ci/job/functest-apex-apex-daily-colorado-daily-colorado/65/console 80% of the output belongs to this test case. This patch will reduce it since debug information is not showed in Jenkins but it is pushed as part of functest.log to artifacts. Change-Id: I56b3adf87d302d118ee5c15eb86bc10237a7cc4e Signed-off-by: jose.lausuch <jose.lausuch@ericsson.com>
2016-09-01Change default memory for Promise flavorjose.lausuch1-1/+1
In some scenarios (bgpvpn) the timeout defined in promise to boot/delete vms is too tight. Having less memory might fix the problem as it will take less time and it doesn't impact the promise test case. Change-Id: Ib3165ea5d6518d88384f76c8edd10dddb0b529a3 Signed-off-by: jose.lausuch <jose.lausuch@ericsson.com>
2016-09-01Add router association utility functionGeorge Paraskevopoulos1-0/+5
Add function to create an association between a router and a bgpvpn Change-Id: I23449a0363f9d8f3c77ea3be1e104586d3c111d1 Signed-off-by: George Paraskevopoulos <geopar@intracom-telecom.com>
2016-09-01Add print-to-a-file option when executing a command in functest_utilsjose.lausuch1-4/+11
Change-Id: I8b91979380d0f4794d5ceaf8edb2a9d80cd7170f Signed-off-by: jose.lausuch <jose.lausuch@ericsson.com>
2016-09-01Fix some bugsboucherv2-2/+2
Change-Id: Iaf0859775933ba09cbbe170af3ce4ed1e6d3e9a1 Signed-off-by: boucherv <valentin.boucher@orange.com>
2016-09-01Force python-keystoneclient to version 2.3.1 for tacker to workjose.lausuch1-0/+1
Tacker client is installed from Tim Rozet repository, it doesn't support version 3 of the keystone client. Change-Id: Ie2005b1590dfccf93442995e96e0c02b4629edea Signed-off-by: jose.lausuch <jose.lausuch@ericsson.com>
2016-09-01stop hardcoded FUNCTEST_REPO path everywhereSerenaFeng17-72/+54
functest repo is used by almost every testcase, each place we get it like: REPOS_DIR = os.getenv('repos_dir') FUNCTEST_REPO = ("%s/functest" % REPOS_DIR) provide a common interface in functest_util.py JIRA: FUNCTEST-453 Change-Id: Ie0635dacc761ed2d05b7e606530368844f32ebaf Signed-off-by: SerenaFeng <feng.xiaowei@zte.com.cn>
2016-09-01Merge "unify functest_yaml obtain process"Jose Lausuch14-71/+38
2016-09-01Merge "Add a specific logger for orchestrator class"Jose Lausuch2-60/+45
2016-08-31Remove copper from lxd scenarioMorgan Richomme1-1/+1
JIRA: FUNCTEST-452 Change-Id: I35c6417e6650ce190e499384f9cf9cca1fa35eb1 Signed-off-by: Morgan Richomme <morgan.richomme@orange.com>
2016-08-31Refactor SFC testcaseGeorge Paraskevopoulos1-49/+54
- Change the way the test checks if the vxlan_tool has started. Now the vxlan_tool is detected even if another python process is up. - Log the exceptions in the vxlan_tool check block. The exceptions that occur are not destructive for the testcase but it's better to be visible. - Add a 60 seconds sleep before the testcase starts to give plenty of time to ODL to add the correct flows. - Move the testcase success/fail check blocks inside the try blocks. The checks are performed on the stdout variable which is in scope only inside the try blocks. Change-Id: I49dec176fb83117e9896ac1d19e8a5bd95e3d7cf Signed-off-by: George Paraskevopoulos <geopar@intracom-telecom.com>
2016-08-31Add a specific logger for orchestrator classboucherv2-60/+45
JIRA: FUNCTEST-428 Change-Id: I1aab08f8f92df3b5527e6f8121cdf45aec2f5c17 Signed-off-by: boucherv <valentin.boucher@orange.com>
2016-08-31unify functest_yaml obtain processSerenaFeng14-71/+38
functest_yaml is needed in almost all of the testcases, the obtain process is the same: with open(os.environ["CONFIG_FUNCTEST_YAML"]) as f: functest_yaml = yaml.safe_load(f) f.close() abstract a method to unify the process, and provide the interface JIRA: FUNCTEST-447 Change-Id: I96a9a9962d7b466f10bbd3b5ab2495957524e22a Signed-off-by: SerenaFeng <feng.xiaowei@zte.com.cn>
2016-08-31Remove all logers as utils method args.Cédric Ollivier16-69/+50
As it modifies method params, it could break the testcases. JIRA: FUNCTEST-428 Change-Id: I4263d9bd812f1803e513464e3c321cc4a7281958 Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
2016-08-31Use a global logger instead of method argsCédric Ollivier1-74/+44
It's the first step to safely remove all logers as method args. Outputs printed by bash scripts called by execute_command are no longer written in functest.log. Else logs printed by python scripts will be duplicated in it. JIRA: FUNCTEST-428 Change-Id: I60702f3b62121f731cd6845c335e227f97df37c9 Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
2016-08-31Merge "unify testcases.yaml obtain process"Morgan Richomme4-8/+8
2016-08-31Merge "delete useless FUNCTEST_REPO variable"Morgan Richomme5-5/+0
2016-08-31parser fails on virtual environementSerenaFeng1-1/+1
Parser tests needs lots of resources, such as vcpu/memory/disk, virtuals pod may not meet the requirements, so exclude Parser running on virtual pods by exclude 'noha' scenario JIRA: FUNCTEST-444 Change-Id: I6f2832fa0f727f762987c41d7a3485dede041151 Signed-off-by: SerenaFeng <feng.xiaowei@zte.com.cn>
2016-08-31Fix problem with SFC testBrady Johnson1-0/+5
- If its not possible to get the IPs, then exit - Patch Set 2: addressed code review comments Change-Id: I16e22513b32e2b0da311ebcc280f3611570634a3 Signed-off-by: Brady Johnson <brady.allen.johnson@ericsson.com>
2016-08-31Committer promotion: Cedric OllivierMorgan Richomme1-0/+1
Cedric joind Functest for Colorado He fully refactored the ODL section including work upstream within ODL community In addition of ODL skills, Cedric is a great benevolent reviewer. Even if he is famous in the project for his gerrit -1, his reviews are always usefull and clearly consolidate the code baseline for functest. He is very relevant on all the technical aspects and we always need a Debianist among commiters Change-Id: I4bd0f778edfaa0d8969bdda3387d5a5ebc8b0f93 Signed-off-by: Morgan Richomme <morgan.richomme@orange.com>
2016-08-31Merge "Committer promotion: Serena Feng"Morgan Richomme1-0/+1
2016-08-31unify testcases.yaml obtain processSerenaFeng4-8/+8
hardcoded "''FUNCTEST_REPO + "/ci/testcases.yaml"''' appears wherever testcases.yaml is needed, abstract a method to provide unified interface JIRA: FUNCTEST-449 Change-Id: Ida8e3d90926d3e2245883a29f577ecb121bd5d6c Signed-off-by: SerenaFeng <feng.xiaowei@zte.com.cn>
2016-08-31delete useless FUNCTEST_REPO variableSerenaFeng5-5/+0
FUNCTEST_REPO appears in lots of feature testcases, but never be used, delete it clean the code JIRA: FUNCTEST-448 Change-Id: I4a9a10eac29e09b481563ed2c33f4122dd9ee3de Signed-off-by: SerenaFeng <feng.xiaowei@zte.com.cn>
2016-08-30Release note: add restriction for joid/dominoMorgan Richomme2-12/+20
JIRA: FUNCTEST-436 Change-Id: I44fb50a2ef58095aff54f03b837bf456850f3057 Signed-off-by: Morgan Richomme <morgan.richomme@orange.com>
2016-08-30disable test case does not make sense for LXD.Narinder Gupta1-0/+11
This patch will disable the test cases which does not make sense to run on LXD deployments. Change-Id: I683b3714aa5af3c750ca57fab9fceff907b7d646 Signed-off-by: Narinder Gupta <narinder.gupta@canonical.com>
2016-08-30Merge "update fuel ref in tacker setup script"Jose Lausuch1-1/+1
2016-08-30Exclude healthcheck from lxd scenariosMorgan Richomme1-1/+1
JIRA: FUNCTEST-442 Change-Id: I1f6888dbddebf28af4b565ca87d9791a21cd4d7e Signed-off-by: Morgan Richomme <morgan.richomme@orange.com>
2016-08-30update fuel ref in tacker setup scriptFerenc Cserepkei1-1/+1
Since the poc script is fixed it's fuel commit hash in the setup script have to be updated to select the fixed poc script. Change-Id: I1fd22e4dc216fd2889adea0c9889df1d4a1bf476 Signed-off-by: Ferenc Cserepkei <ferenc.cserepkei@ericsson.com>
2016-08-30Allow realtime Jenkins' console outputCédric Ollivier1-1/+1
Change-Id: I6005000762eb6e7bb7e40ccaea35004a07b47aa1 Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
2016-08-30Force stdin, stdout and stderr to be totally unbufferedCédric Ollivier1-0/+1
It protects against mixed logs as lots of python scripts directly call print which doesn't flush stdout as opposed to log.info(). This requirement is induced by subprocess.Popen() initiated by run_tests.py. The following programs could switch to logging instead of print [1] [1] $ find . -name "*.py" |xargs grep -v ^[[:space:]]*# \ |grep \ print\ |cut -d: -f 1 |sort -u ./testcases/Controllers/ODL/OpenDaylightTesting.py ./testcases/Controllers/ONOS/Sfc/Sfc_fun.py ./testcases/Controllers/ONOS/Sfc/Sfc.py ./testcases/Controllers/ONOS/Teston/adapters/client.py ./testcases/Controllers/ONOS/Teston/adapters/connection.py ./testcases/Controllers/ONOS/Teston/adapters/environment.py ./testcases/Controllers/ONOS/Teston/adapters/foundation.py ./testcases/OpenStack/vPing/vping_util.py ./testcases/security_scan/connect.py ./utils/functest_utils.py ./utils/functest_vacation.py ./utils/openstack_clean.py JIRA: FUNCTEST-441 Change-Id: I7e9a888507f10352be7e85abd7f27dba4b1a7b53 Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
2016-08-30Merge "Stop hard-coding db url"Morgan Richomme1-3/+3
2016-08-30Merge "Flush stdout as it can be redirected to log"Morgan Richomme1-0/+1
2016-08-30Exclude floating-ip related tempest tests to blacklist for bgpvpn scenariojose.lausuch1-0/+2
Change-Id: I7115018e64a2413209313b9a432d3d1b59e841d1 Signed-off-by: jose.lausuch <jose.lausuch@ericsson.com>
2016-08-30Stop hard-coding db urlCédric Ollivier1-3/+3
Change-Id: I587ac6d210135e7124ba006a977a675a4d38b518 Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
2016-08-30Flush stdout as it can be redirected to logCédric Ollivier1-0/+1
It avoids mixed logs in Jenkins' console by flushing every message printed by the forked programs as logger.info() does. It must be complete by refactoring logger management in util functions [1] and by preventing output buffering in forked programs. [1] https://jira.opnfv.org/browse/FUNCTEST-428 JIRA: FUNCTEST-441 Change-Id: Id8da814d79ea6b1e6b91b49e23a85c3802e11515 Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
2016-08-30Merge "Set env vars as mandatory to push to DB"Morgan Richomme2-11/+25
2016-08-30Merge "Update Release note"Morgan Richomme1-21/+40
2016-08-29Update Release noteMorgan Richomme1-21/+40
JIRA: FUNCTEST-436 Change-Id: I3cbf6b276494c77afb10802af6735647ea6e62d5 Signed-off-by: Morgan Richomme <morgan.richomme@orange.com>
2016-08-29Merge "Review dev guide"Morgan Richomme1-98/+147
2016-08-29Review dev guideMorgan Richomme1-98/+147
JIRA: FUNCTEST-433 Change-Id: Ic141a5b74d9b3afc165e1f1f65f3dc1b05d77f4e Signed-off-by: Morgan Richomme <morgan.richomme@orange.com>
2016-08-26Set env vars as mandatory to push to DBCédric Ollivier2-11/+25
Now env vars are mandatory in push_results_to_db. It avoids testing programs to check them even if they are useless from their points of view. The related getters aren't removed to help debugging via logger_test_results which could be run without report. Change-Id: Iadb94319529f5ec4230fb62042511948183b93ed Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
2016-08-26Merge "configuration guide review"Morgan Richomme2-42/+48
2016-08-26configuration guide reviewMorgan Richomme2-42/+48
JIRA: FUNCTEST-435 Change-Id: I5b18f1223a5892d084c761f174ac5f713ceed476 Signed-off-by: Morgan Richomme <morgan.richomme@orange.com>
2016-08-26refactor create shared network process to eliminate reduplicateSerenaFeng3-32/+33
Some places shared network is created through create_network_full first, then update it to shared, abstract a uniformed method to eliminate reduplicate JIRA: FUNCTEST-437 Change-Id: I13c20ab061323188c5eb3b8c861674df38083c62 Signed-off-by: SerenaFeng <feng.xiaowei@zte.com.cn>
2016-08-26Merge "refactor create or get image process to eliminate reduplicate"Morgan Richomme4-56/+35