[opnfv-tech-discuss] #auto x86 POD for Auto Project
Hi, The Auto project is looking for an x86 POD for running CI jobs on and I'm reaching out to see if you have any available resources we could use. Regards, Paul Ionut Vaduva This message, including attachments, is CONFIDENTIAL. It may also be privileged or otherwise protected by law. If you received this email by mistake please let us know by reply and then delete it from your system; you should not copy it or disclose its contents to anyone. All messages sent to and from Enea may be monitored to ensure compliance with internal policies and to protect our business. Emails are not secure and cannot be guaranteed to be error free as they can be intercepted, a mended, lost or destroyed, or contain viruses. The sender therefore does not accept liability for any errors or omissions in the contents of this message, which arise as a result of email transmission. Anyone who communicates with us by email accepts these risks. -=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#22018): https://lists.opnfv.org/g/opnfv-tech-discuss/message/22018 Mute This Topic: https://lists.opnfv.org/mt/25520512/21656 Mute #auto: https://lists.opnfv.org/mk?hashtag=auto=2783016 Group Owner: opnfv-tech-discuss+ow...@lists.opnfv.org Unsubscribe: https://lists.opnfv.org/g/opnfv-tech-discuss/unsub [arch...@mail-archive.com] -=-=-=-=-=-=-=-=-=-=-=-
Re: [opnfv-tech-discuss] [Auto] Jenkins for Auto
Hi Martin, See my comments inline... -Original Message- From: Klozik Martin Sent: Tuesday, June 12, 2018 10:46 AM To: Paul Vaduva Cc: Tina Tsou ; gerard.d...@wipro.com; Cristina Pauna ; Joe Kidder ; opnfv-tech-discuss@lists.opnfv.org Subject: Re: [Auto] Jenkins for Auto Hi Paul, thanks for taking care about initial yaml file definition at jjb/auto/auto.yaml. I've also noticed your +1 to my "ci/build-auto.sh" patch. It was acked by other team members too, so we will go ahead with my proposal. Let me try to summarize my understanding of your CI job configuration. Please let me know your comments (that applies to all). Your idea is to utilize Armband project for OPNFV installation via FUEL at ARM POD. I'm not familiar with the details, but it seems, that current configuration (with PROJECT=armband) works only for armband repository. Do you know if it can be modified/configured to properly install OPNFV at our pod while still utilizing existing build framework? [Paul Vaduva] Yes, thats true I made a job for deploying fuel armband version, it there will be other installers we would probably need other ci jobs that will install (compass, apex, etc..), Yes I already proposed a patch to extend armband deploy job template to accept project parameter. https://gerrit.opnfv.org/gerrit/#/c/58415/ If not, we would probably need to do it "manually", e.g. to introduce "install_opnfv" function into ci/build-auto.sh, which will: 1) clone armband repo as part of our CI job 2) modify WORKSPACE to point to armband repo 3) execute ci/deploy.sh from armband (and other steps if required) to take care about OPNFV installation 4) restore WORKSPACE to its orignal value before ONAP deployment [Paul Vaduva] I thought about this as well (deploying Openstack as part of auto ci job but it seem to be much more effort in maintaining the deploy procedure then the previous variant (just using armband template for deploying their project). It seems to make more sense to keep the Openstack deploying separated from ONAP deployment. Of course we would have to update this logic in the future to support both arm and x64 pods. Best Regards, Martin Od: Klozik Martin Odesláno: čtvrtek 7. června 2018 9:57 Komu: Joe Kidder; Paul Vaduva Kopie: Tina Tsou; gerard.d...@wipro.com<mailto:gerard.d...@wipro.com>; Cristina Pauna Předmět: Re: Jenkins for Auto Direct link to previously mentioned gerrit draft review: https://gerrit.opnfv.org/gerrit/#/c/58307/ --Martin Od: Klozik Martin Odesláno: čtvrtek 7. června 2018 9:48 Komu: Joe Kidder; Paul Vaduva Kopie: Tina Tsou; gerard.d...@wipro.com<mailto:gerard.d...@wipro.com>; Cristina Pauna Předmět: Re: Jenkins for Auto Hi Paul, just for the sake of discussion, I've prepared a draft patch of possible ci script skeleton. In case that we will decide to go in this direction it can be used as an initial version straight away and simply called from jenkins yaml file. May be we can do a simple vote in gerrit by adding +1 or -1. Best Regards, Martin Od: Klozik Martin Odesláno: čtvrtek 7. června 2018 9:07:36 Komu: Joe Kidder; Paul Vaduva Kopie: Tina Tsou; gerard.d...@wipro.com<mailto:gerard.d...@wipro.com>; Cristina Pauna Předmět: Re: Jenkins for Auto Thanks Joe for adding me in the loop. Hi Paul, I'm sorry, I was not aware of your activities. I've sent a few thoughts about possible Auto CI (initial) configuration into the Auto mailing list. Could you please have a look at it? Based on this thread I can see that you're about to define a yaml file and prepare initial body of Auto jobs (daily/verify/merge). Let me know if I can be of any help. In a nutshell, I was proposing the same concept as is used by vswitchperf project for some time and later adopted by storperf too. It means, to define only minimalistic YAML file with definition of jobs frequency, allowed slaves, etc. and then to invoke a script stored inside Auto repo (e.g. ci/build-auto.sh) with the name of the job (e.g. ci/build-auto.sh verify). We can start with empty body, so all jobs will always end with success if slave (unh-pod1) is up and auto repository clones smoothly. Later we can add a real stuff there, based on the progress of platform installation scripts (for OPNFV and ONAP) and Auto tests automation. My idea was to start with a common functions (inside build-auto.sh) for platform installation (which can be shared among all jobs). Based on the experience we can later split it to separate jobs to get some visibility of particular "sub-task" stability directly in the Jenkins job history. Best Regards, Martin BTW, could you please CC "auto mailing list" in the future discussions? So the knowledge is spread among the team and properly archived by the mailing list. Thanks. Od: Joe Kidder mailto:joe.k
Re: [opnfv-tech-discuss] [Doctor] Integration with VNFM
Hi Ifat, We have a minor modification to the script zabbix_vitrage.py. It’s just at line 83 we split only for the first ‚=’ as sometimes the expression value may contain „=” and it confuses the split function. Best Regards, Paul From: Afek, Ifat (Nokia - IL/Kfar Sava) [mailto:ifat.a...@nokia.com] Sent: Monday, July 31, 2017 4:47 PM To: Ciprian Barbu <ciprian.ba...@enea.com> Cc: opnfv-tech-discuss@lists.opnfv.org; Paul Vaduva <paul.vad...@enea.com>; Cristina Pauna <cristina.pa...@enea.com> Subject: Re: [opnfv-tech-discuss] [Doctor] Integration with VNFM Hi, Please see my comments inline. Best regards, Ifat From: Ciprian Barbu <ciprian.ba...@enea.com<mailto:ciprian.ba...@enea.com>> Date: Monday, 31 July 2017 at 16:09 To: "Afek, Ifat (Nokia - IL/Kfar Sava)" <ifat.a...@nokia.com<mailto:ifat.a...@nokia.com>> Cc: "opnfv-tech-discuss@lists.opnfv.org<mailto:opnfv-tech-discuss@lists.opnfv.org>" <opnfv-tech-discuss@lists.opnfv.org<mailto:opnfv-tech-discuss@lists.opnfv.org>>, Paul Vaduva <paul.vad...@enea.com<mailto:paul.vad...@enea.com>>, Cristina Pauna <cristina.pa...@enea.com<mailto:cristina.pa...@enea.com>> Subject: RE: [opnfv-tech-discuss] [Doctor] Integration with VNFM Hi, Thank you for the details. We managed to replicate the scenario where nova-compute is killed, Zabbix detects the process down and Vitrage will eventually receive the „alarm” from Zabbix. There are a couple of things though: 1. Vitrage takes a long time to process the alarm from Zabbix, from what I can tell the Zabbix driver in Vitrage works by polling the Zabbix hosts at regular intervals, but we couldn’t exactly figure how to set that polling time [Ifat] The Zabbix driver can work in one of two modes: poll Zabbix periodically, or receive push notifications. Of course, push is the preferred option. If you configure Zabbix to send notification to Vitrage for every trigger change[1], Vitrage will react immediately. 2. The nova service is forced down, this can be seen in /var/log/nova/nova-api.log and in Vitrage Topology, but we couldn’t figure out yet how to verify an alarm is triggered on Nova side. I’m not too familiar with the Ceilometer and Aodh architecture, should there be an Aodh alarm generated, which can be later consumed by a VFM, like Tacker? [Ifat] You can check nova service status also by executing ‘nova service-list’. Regarding the Aodh alarm: Nova should send a notification to the OpenStack message bus. I’m not sure how this is tracked. To see the alarm in Aodh, you should do the following: · Configure Ceilometer (or Panko in newer versions?) to create an event based on this notification · Create an Aodh event-alarm that is triggered once this event is received. I’m not so familiar with the details. I had it working in an older OpenStack version, but I’m not sure how to configure it in Pike. [1] https://docs.openstack.org/vitrage/latest/contributor/zabbix_vitrage.html - see “Zabbix web UI configuration” Any help is greatly appreciated, /Ciprian From: Afek, Ifat (Nokia - IL/Kfar Sava) [mailto:ifat.a...@nokia.com] Sent: Sunday, July 30, 2017 2:43 PM To: Ciprian Barbu <ciprian.ba...@enea.com<mailto:ciprian.ba...@enea.com>> Cc: opnfv-tech-discuss@lists.opnfv.org<mailto:opnfv-tech-discuss@lists.opnfv.org> Subject: Re: [opnfv-tech-discuss] [Doctor] Integration with VNFM Hi, Two corrections: · In vitrage.conf, it should be ‘notifiers’ and not ‘Notifiers’: [DEFAULT] notifiers = nova · In order to make the Nova notifier call Nova force-down API you should add a mark_down action in one of Vitrage templates. You can see an example in the template sample: https://github.com/openstack/vitrage/blob/master/etc/vitrage/templates.sample/host_down_scenarios.yaml For additional explanations: https://docs.openstack.org/vitrage/latest/contributor/vitrage-template-format.html Best Regards, Ifat. From: <opnfv-tech-discuss-boun...@lists.opnfv.org<mailto:opnfv-tech-discuss-boun...@lists.opnfv.org>> on behalf of "Afek, Ifat (Nokia - IL/Kfar Sava)" <ifat.a...@nokia.com<mailto:ifat.a...@nokia.com>> Date: Sunday, 30 July 2017 at 10:16 To: Ciprian Barbu <ciprian.ba...@enea.com<mailto:ciprian.ba...@enea.com>> Cc: "opnfv-tech-discuss@lists.opnfv.org<mailto:opnfv-tech-discuss@lists.opnfv.org>" <opnfv-tech-discuss@lists.opnfv.org<mailto:opnfv-tech-discuss@lists.opnfv.org>> Subject: Suspected SPAM - Re: [opnfv-tech-discuss] [Doctor] Integration with VNFM Hi Ciprian, You understood correctly. The call to Nova API is done by the Vitrage Nova notifier, if it is enabled. To enable it add you /etc/vitrage/vitrage.conf: [DEFAULT] Notifiers = nova Best Regards, Ifat. From: Ciprian Barbu <ciprian.ba...@enea.com<mailto:ciprian.ba...@enea.com>> D
