Hi Cedric, I don’t really have any results which failed because of the timeout. I am just curious about the timeout Functest set. Because as I know, the default timeout of Tempest is 300 and 180 for shade. So a little worried about the 60 Functest set.
Yes. The default wait is False for shade, but Functest has already set it to be True when calling create_server(). So the wait_for_server() seems to be duplicate. For more information about the vping test case on VLAN environment, I have attached them in the JIRA ticket [1]. And as I mentioned in the comments, I have a simple discussion with the shade team about this question [2]. Hope this can help you. [1] https://jira.opnfv.org/browse/FUNCTEST-996 [2] http://lists.openstack.org/pipermail/openstack-infra/2018-June/005983.html From: opnfv-tech-discuss@lists.opnfv.org [mailto:opnfv-tech-discuss@lists.opnfv.org] On Behalf Of Cedric OLLIVIER Sent: Monday, July 30, 2018 8:22 PM To: xudan (N) <xuda...@huawei.com> Cc: opnfv-tech-discuss@lists.opnfv.org; Georg Kunz <georg.k...@ericsson.com>; louie long <yl...@biigroup.cn> Subject: Re: [opnfv-tech-discuss] [Functest] vping failed Hello Xudan, 60s is just in case of Cirros. That duration is very big regarding the one required: vping testcases ends before that minute. In case of Cloudify, the image is much more bigger which increases the duration when caching images. Why not setting 180 but all installers are passing all Cirros-based testcases. But be free to send us any log showing that the timeout has to be increased. It has to be guided by experiment. Our target is simply to select the right timeout according to the image we have selected (shade selects an arbitrary value). We should detect if the duration is abnormally long instead of setting the biggest timeouts. At the end of the day, the end users will face with that issues. As far as I know, wait arg is set to False by default in create_server(). Then we do wait via wait_for_server() or setting wait=True (both are equal). Yes we would like to support vlan (even if few ids are supported) but badly it's out of OPNFV gates. To check your results, we need all logs and config files (please send me them by email). I don't know if you correctly modify · network_type · physical_network · segmentation_id We can implement a fallback mechanism based on the port created if Shade team doesn't update its code. But following our leitmotiv "upstream first", we should work on shade instead of Functest here. Cédric 2018-07-30 10:33 GMT+02:00 xudan <xuda...@huawei.com<mailto:xuda...@huawei.com>>: Hi Functest team, I found there is a patch [1] aims at increasing timeout when creating servers. It sets the timeout of creating servers to be 60 while the default value of shade is 180s. Does 60 a little short? I suggest to keep it as the shade default value. Another question is what the function ‘wait_for_server’ [2] for? It already waits during the function ‘create_server’. The Functest vping test cases failed on my local environment. I guess it’s because of that my environment is VLAN and shade failed to get the server’s private ip in this case. I have created a JIRA ticket. Can someone help to check it? @louie I remembered that you have met a similar problem several days before. Could you help to confirm if the error is the same as yours and can it work well on your environment now? [1] https://gerrit.opnfv.org/gerrit/#/c/60243/ [2] https://github.com/opnfv/functest/blob/stable/fraser/functest/core/singlevm.py#L184 [3] https://jira.opnfv.org/browse/FUNCTEST-996 Thanks, Xudan -=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#21654): https://lists.opnfv.org/g/opnfv-tech-discuss/message/21654 Mute This Topic: https://lists.opnfv.org/mt/23858006/21656 Group Owner: opnfv-tech-discuss+ow...@lists.opnfv.org Unsubscribe: https://lists.opnfv.org/g/opnfv-tech-discuss/unsub [arch...@mail-archive.com] -=-=-=-=-=-=-=-=-=-=-=-
