[opnfv-tech-discuss] [barometer] Update

2017-09-21 Thread Tahhan, Maryam
Hi folks

[barometer] Update 



Actions:

* Maryam or team to create a wiki on the collectd VES application

* Maryam to review/help out with Al's definitions.

* Al to look at LLC definition --> Done but clarification needed on 
whether it is Kilobytes/Kibibytes.  ACTION Maryam to follow up



Dev updates: 
https://wiki.opnfv.org/display/fastpath/Barometer+Development+Updates



Functest update:

* issues since move to new model of split containers in functest.

* some JIRA bugs have been filed with functest.

* Emma has posted some patches to fix the long term testing issues encountered 
with Gnocchi



IFA027:

* Network Functions Virtualization (NFV) Release 2; Management and 
Orchestration; Performance Measurements Specification Perf. Measurements Spec

* Specify the performance measurements (i.e. performance metrics, performance 
values) and use cases for descriptors and interfaces, including Or-Vnfm 
reference point, Ve-Vnfm reference point, Vi-Vnfm reference point, Or-Vi 
reference point, and Os-Ma-nfvo reference point, based on the performance 
metrics collected from NFVI.

* References TST008

* 
https://docbox.etsi.org/isg/nfv/open/drafts/IFA027_Perf_Measurements_Spec/NFV-IFA027v050.zip

* Not looking at adding new metrics that go beyond the scope of past standard.

* Need to add a clarification note to TST008 about memory used as it's directly 
referenced by this doc.





Runtime analysis: 
https://wiki.opnfv.org/display/fastpath/Runtime+analysis+of+the+monitoring+agents+methodology





Please let me know if you have any questions or comments.

BR
Maryam
___
opnfv-tech-discuss mailing list
opnfv-tech-discuss@lists.opnfv.org
https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss


[opnfv-tech-discuss] [barometer] Update

2017-09-15 Thread Tahhan, Maryam
Hi folks

[barometer] Update 



Actions:

* Maryam to extend the meeting next week and arrange a readout on 
Prometheus and review the runtime analysis. --> DONE

* Maryam to provide feedback to Sridhar on the runtime performance 
analysis. --> DONE

* Maryam to review TST008 updates --> DONE

* Al to look at the local and remote memory BW explanations for a 
definition. --> DONE Wiki additions in place 
https://wiki.opnfv.org/display/fastpath/Runtime+analysis+of+the+monitoring+agents+methodology

* Sridhar to publish test steps, HW and SW configurations and initial 
test results to runtime analysis wiki --> DONE 
https://wiki.opnfv.org/display/fastpath/Runtime+analysis+of+the+monitoring+agents+methodology

* Maryam to make sure the kafka user guide is pushed to gerrit for 
review --> DONE https://gerrit.opnfv.org/gerrit/#/c/40753/

* Maryam or team to create a wiki on the collectd VES application

* Maryam to review/help out with Al's defintions.

* Maryam + Al to look at LLC definition



Dev updates: 
https://wiki.opnfv.org/display/fastpath/Barometer+Development+Updates



Functest update:

  * Successfully deploying Apex now with a work around in the env.

  * Still some instability issues with gnocchi beyond testing on a fresh env. 
Will need to log a bug and document it in the barometer documentation.

  * Need to look at stress/longer term tests in F.

  * Issue spotted in virtual and baremetal env.



Prometheus readout:  
https://wiki.opnfv.org/display/fastpath/Collectd+integration+with+prometheus

* Uses a pull model only

* collectd has a Prometheus plugin that sets up a server that can be queried 
for stats. It natively exports metrics into the Prometheus format.

* Storage in Prometheus is on the local host.

* runs as a pod in your Kubernetes cluster.

* It has its own dashboard.





Runtime analysis: 
https://wiki.opnfv.org/display/fastpath/Runtime+analysis+of+the+monitoring+agents+methodology





Please let me know if you have any questions or comments.

BR
Maryam
___
opnfv-tech-discuss mailing list
opnfv-tech-discuss@lists.opnfv.org
https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss


Re: [opnfv-tech-discuss] [barometer] Update

2017-09-06 Thread Tahhan, Maryam
Hi Aaron

The overview will be along the lines of how we can get the platform telemetry 
integrated into a container environment. One of the Guys from the Intel SA team 
will give the overview. He also looked at C-advisor and Heapster and how they 
are used… so they can provide a comparison there, as part of the overview if 
that’s of interest.

BR
Maryam

From: Aaron Smith [mailto:aasm...@redhat.com]
Sent: Wednesday, September 6, 2017 2:26 PM
To: Tahhan, Maryam <maryam.tah...@intel.com>
Cc: opnfv-tech-discuss@lists.opnfv.org; Power, Damien <damien.po...@intel.com>; 
Mcmahon, Tony B <tony.b.mcma...@intel.com>; Gherghe, Calin 
<calin.gher...@intel.com>
Subject: Re: [opnfv-tech-discuss] [barometer] Update 

Hi Maryam,
  Apologies for missing the first part of this weeks status update.  Who will 
be giving the Prometheus review
and is there any preliminary documentation?  We have done a Prometheus review 
but only from the standpoint of integration,
 not runtime performance.  Looking forward to hearing the analysis.

Aaron

On Wed, Sep 6, 2017 at 8:13 AM, Tahhan, Maryam 
<maryam.tah...@intel.com<mailto:maryam.tah...@intel.com>> wrote:
Hi folks
[barometer] Update 



Actions:

-  Maryam to extend the meeting next week and arrange a readout on 
Prometheus and review the runtime analysis.

-  Maryam to provide feedback to Sridhar on the runtime performance 
analysis.

-  Maryam to review TST008 updates

-  Al to look at the local and remote memory BW explanations for a 
definition.

-  Maryam to make sure the kafka user guide is pushed to gerrit for 
review --> DONE https://gerrit.opnfv.org/gerrit/#/c/40753/

-  Maryam or team to create a wiki on the collectd VES application



Dev updates: 
https://wiki.opnfv.org/display/fastpath/Barometer+Development+Updates



Functest update:

-  MS 5 + 6 met

-  Still some failures due to authentication problems in the nightly 
build env. Working with the Apex team to resolve.

-  Also encountering some proxy issues.

o   Learning for next release to have a POD backup available for project 
functesting in the OPNFV lab.





Please let me know if you have any questions or comments.

BR
Maryam

___
opnfv-tech-discuss mailing list
opnfv-tech-discuss@lists.opnfv.org<mailto:opnfv-tech-discuss@lists.opnfv.org>
https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss



--

AARON SMITH

SENIOR PRINCIPAL SOFTWARE ENGINEER, NFVPE

Red Hat

<https://www.redhat.com/>

314 Littleton Rd, Westford, MA 01886

aasm...@redhat.com<mailto:aasm...@redhat.com>M: 
617.877.4814
[https://www.redhat.com/files/brand/email/sig-redhat.png]<https://red.ht/sig>


___
opnfv-tech-discuss mailing list
opnfv-tech-discuss@lists.opnfv.org
https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss


Re: [opnfv-tech-discuss] [barometer] Update

2017-09-06 Thread Aaron Smith
Hi Maryam,
  Apologies for missing the first part of this weeks status update.  Who
will be giving the Prometheus review
and is there any preliminary documentation?  We have done a Prometheus
review but only from the standpoint of integration,
 not runtime performance.  Looking forward to hearing the analysis.

Aaron

On Wed, Sep 6, 2017 at 8:13 AM, Tahhan, Maryam 
wrote:

> Hi folks
>
> [barometer] Update 
>
>
>
> Actions:
>
> -  *Maryam *to extend the meeting next week and arrange a readout
> on Prometheus and review the runtime analysis.
>
> -  *Maryam *to provide feedback to Sridhar on the runtime
> performance analysis.
>
> -  *Maryam *to review TST008 updates
>
> -  *Al *to look at the local and remote memory BW explanations
> for a definition.
>
> -  *Maryam *to make sure the kafka user guide is pushed to gerrit
> for review à *DONE https://gerrit.opnfv.org/gerrit/#/c/40753/
>  *
>
> -  *Maryam *or team to create a wiki on the collectd VES
> application
>
>
>
> Dev updates: https://wiki.opnfv.org/display/fastpath/Barometer+
> Development+Updates
>
>
>
> Functest update:
>
> -  MS 5 + 6 met
>
> -  Still some failures due to authentication problems in the
> nightly build env. Working with the Apex team to resolve.
>
> -  Also encountering some proxy issues.
>
> o   *Learning for next release *to have a POD backup available for
> project functesting in the OPNFV lab.
>
>
>
>
>
> Please let me know if you have any questions or comments.
>
> BR
> Maryam
>
> ___
> opnfv-tech-discuss mailing list
> opnfv-tech-discuss@lists.opnfv.org
> https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss
>
>


-- 

AARON SMITH

SENIOR PRINCIPAL SOFTWARE ENGINEER, NFVPE

Red Hat



314 Littleton Rd, Westford, MA 01886

aasm...@redhat.comM: 617.877.4814

___
opnfv-tech-discuss mailing list
opnfv-tech-discuss@lists.opnfv.org
https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss


[opnfv-tech-discuss] [barometer] Update

2017-09-06 Thread Tahhan, Maryam
Hi folks
[barometer] Update 



Actions:

-  Maryam to extend the meeting next week and arrange a readout on 
Prometheus and review the runtime analysis.

-  Maryam to provide feedback to Sridhar on the runtime performance 
analysis.

-  Maryam to review TST008 updates

-  Al to look at the local and remote memory BW explanations for a 
definition.

-  Maryam to make sure the kafka user guide is pushed to gerrit for 
review --> DONE https://gerrit.opnfv.org/gerrit/#/c/40753/

-  Maryam or team to create a wiki on the collectd VES application



Dev updates: 
https://wiki.opnfv.org/display/fastpath/Barometer+Development+Updates



Functest update:

-  MS 5 + 6 met

-  Still some failures due to authentication problems in the nightly 
build env. Working with the Apex team to resolve.

-  Also encountering some proxy issues.

o   Learning for next release to have a POD backup available for project 
functesting in the OPNFV lab.





Please let me know if you have any questions or comments.

BR
Maryam
___
opnfv-tech-discuss mailing list
opnfv-tech-discuss@lists.opnfv.org
https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss


Re: [opnfv-tech-discuss] [barometer] Update

2017-08-08 Thread Tahhan, Maryam
Hi folks
[barometer] Update 



Congratulations to our new committers: Al Morton, Aaron Smith, Emma Foley, and 
Calin Gherghe.



Maryam OOO till the 29th of August. Please complete any code reviews you are 
added to for MS5 and MS6.


* E Release
  * MS 5 pushed out to Aug 11, MS6 pushed out to August 25

  * We are very much in the middle of Apex integration. 2 scenarios are 
created:  os-nosdn-bar-ha and os-nosdn-bar-noha @ 
https://build.opnfv.org/ci/view/apex/job/apex-daily-master/
  * Apex integration for MS5:


Plugin


integration with Apex for OPNFV E


Documentation Review


Comments


Dpdkstats





To be done


No DPDK specific scenario


Dpdkevents





To be done


No DPDK specific scenario


Hugepages


X


To be done





Intel RDT


Requires an RPM fix


To be done





IPMI





To be done


NOT MERGED TO MASTER YET will not make E


Mcelog (memory RAS)


X


To be done


Needs to be updated when pull request gets merged


OvS stats


X


To be done


No multi instance


OvS events


X


To be done





SNMP-Agent


Requires an RPM fix


To be done





Virt


X


To be done





PMU stats


Will see if we can get this in


To be done


MERGED TO MASTER


OvS DPDK PMD





To be done


No DPDK specific scenario


Gnocchi


X


To be done





Aodh


X


To be done





Ceilometer plugin








Deprecated for E - based on Gnocchi availability


PCIe Errors











RAS Other













Please let me know if you have any questions or comments.

BR
Maryam
___
opnfv-tech-discuss mailing list
opnfv-tech-discuss@lists.opnfv.org
https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss


[opnfv-tech-discuss] [barometer] Update

2017-08-01 Thread Tahhan, Maryam
Hi folks
[barometer] Update 


* E Release
  * MS 5 pushed out to Aug 11, MS6 pushed out to August 25

  * We are very much in the middle of Apex integration. It looks like we are 
nearly in with our 2 scenarios: 

  os-nosdn-bar-ha and os-nosdn-bar-noha
  * Apex integration for MS5:


Plugin


integration with Apex for OPNFV E


Documentation Review


Comments


Dpdkstats





To be done


No DPDK specific scenario


Dpdkevents





To be done


No DPDK specific scenario


Hugepages


X


To be done





Intel RDT


Requires an RPM fix


To be done





IPMI





To be done


NOT MERGED TO MASTER YET will not make E


Mcelog (memory RAS)


X


To be done


Needs to be updated when pull request gets merged


OvS stats


X


To be done


No multi instance


OvS events


X


To be done





SNMP-Agent


Requires an RPM fix


To be done





Virt


X


To be done





PMU stats


Will see if we can get this in


To be done


MERGED TO MASTER


OvS DPDK PMD





To be done


No DPDK specific scenario


Gnocchi


X


To be done





Aodh


X


To be done





Ceilometer plugin








Deprecated for E - based on Gnocchi availability


PCIe Errors











RAS Other













Please let me know if you have any questions or comments.

BR
Maryam
___
opnfv-tech-discuss mailing list
opnfv-tech-discuss@lists.opnfv.org
https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss


Re: [opnfv-tech-discuss] [barometer] Update

2017-07-27 Thread Tahhan, Maryam
Hi folks
[barometer] Update 

* Gurpreet - Action to send on link for the demo and the github agent code base.

* E Release
  * Next Milestone is MS 5 - 7/27/2017 - scenario integration
  * Src scripts

Plugin


documentation in OPNFV


integration with the src script in OPNFV


integration with CI RPM - (becomes eligible for E through Apex)


Comments


Dpdkstats


X


X





No DPDK specific scenario


Dpdkevents


X


X





No DPDK specific scenario


Hugepages


X


X


X





Intel RDT


X


X


X





IPMI


X








NOT MERGED TO MASTER YET will not make E


Mcelog (memory RAS)


X


X


X


Needs to be updated when pull request gets merged


OvS stats


X


X


X


No multi instance


OvS events


X


X


X





SNMP-Agent


X


X


X





Virt


X


X


X





PMU stats











MERGED TO MASTER but timeline was too close to code freeze- will not make E


OvS DPDK PMD


MUST DO for E


X





No DPDK specific scenario


Gnocchi


MUST DO for E


MUST DO for E


through tarballs rather than RPMs)





Aodh


MUST DO for E


MUST DO for E


through tarballs rather than RPMs)





Ceilometer plugin











Deprecated for E - based on Gnocchi availability


PCIe Errors














RAS Other














* Dev update: 
https://wiki.opnfv.org/display/fastpath/Barometer+Development+Updates

* Test Update:

* E release

  * still blocked by Apex
* Barometer Apex integration - reviews reviewed by Tim Rozet, thanks Tim 
and John - 
https://gerrit.opnfv.org/gerrit/#/q/owner:john.hinman%2540intel.com+status:open

* ETSI document TST008:
   * Most of the updates were discussed and provisionally accepted
   * Definitions for memory parameters need to be looked at again - for 
physical and virtualized views


* Comparative Agent:
  * Sridhar shared some interesting results
  * 
https://wiki.opnfv.org/display/fastpath/Runtime+analysis+of+the+monitoring+agents+methodology
* Maryam has feedback on the Monitoring Agent inference questions deck -  
[ACTION] Maryam to send out in a separate email with comments inline in the 
Deck.

VES
  * kafka mechanism which will allow us to relay stats from collectd to the VES 
mapping plugin - ready
  * kafka support for notifications started.
  * Schema update through configuration design + implementation underway.

* discussion on BAROMETER-63
"Proposed Redfish Telemetry Model"
(WIP, April 2017): 
http://www.dmtf.org/sites/default/files/standards/documents/DSP-IS0002_0.9a.zip

The Redfishtool is a simple python application of CLI for Redfish; available in:
https://github.com/DMTF/Redfishtool

catfish mockup from:
https://github.com/openstack/python-redfish/blob/master/dmtf/DSP2043_1.1.0.zip


Future discussion items:
* Inventory management - in scope for barometer or some other OPNFV project - 
discussion.

* CI - RPM builds
   * [ACTION] Maryam to follow up with Aric/Fatih on storage - note sent to 
Aric, need to follow up. - Maryam to reach out on IRC.


BR
Maryam
___
opnfv-tech-discuss mailing list
opnfv-tech-discuss@lists.opnfv.org
https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss


[opnfv-tech-discuss] [barometer] Update

2017-07-21 Thread Tahhan, Maryam
Hi folks
[barometer] Update 

* Gurpreet - Action to send on link for the demo and the github agent code base.

* D3:
   * repo tagged with a new Tag D3.1

* E Release
  * Next Milestone is MS 5 - 7/27/2017 - scenario integration
  * Src scripts

Plugin


documentation in OPNFV


integration with the src script in OPNFV


integration with CI RPM - (becomes eligible for E through Apex)


Comments


Dpdkstats


X


X





No DPDK specific scenario


Dpdkevents


X


X





No DPDK specific scenario


Hugepages


X


X


X





Intel RDT


X


X


X





IPMI


X








NOT MERGED TO MASTER YET will not make E


Mcelog (memory RAS)


X


X


X


Needs to be updated when pull request gets merged


OvS stats


X


Will be pushed for review


X


No multi instance


OvS events


X


X


X





SNMP-Agent


X


Will be pushed for review


X





Virt


X


Will be pushed for review


X





PMU stats











NOT MERGED TO MASTER YET - will not make E


OvS DPDK PMD


MUST DO for E


Will be pushed for review





No DPDK specific scenario


Gnocchi


MUST DO for E


MUST DO for E


MUST DO for E (but through tarballs rather than RPMs)





Aodh


MUST DO for E


MUST DO for E


MUST DO for E (but through tarballs rather than RPMs)





Ceilometer plugin











Deprecated for E - based on Gnocchi availability


PCIe Errors














RAS Other














* Dev update: 
https://wiki.opnfv.org/display/fastpath/Barometer+Development+Updates

* CI - RPM builds
   * [ACTION] Maryam to follow up with Aric/Fatih on storage - note sent to 
Aric, need to follow up. - Maryam to reach out on IRC.

* Test Update:

* E release
* Barometer Apex integration - reviews reviewed by Tim Rozet, thanks Tim 
and John - 
https://gerrit.opnfv.org/gerrit/#/q/owner:john.hinman%2540intel.com+status:open

* ETSI document TST008:
   * Al made the updates based on the discussion and pushed the doc for review 
- thanks Al


* Next week - discussion on BAROMETER-63


* Comparative Agent:
  * Sridhar shared some interesting results
  * 
https://wiki.opnfv.org/display/fastpath/Runtime+analysis+of+the+monitoring+agents+methodology
* Maryam has feedback on the Monitoring Agent inference questions deck -  
[ACTION] Maryam to send out in a separate email with comments inline in the 
Deck.
VES
   * Schema has been published for a while - need to look at this from an 
information model perspective, engaging an SDO.
   * Modelling info based on type vs significance? Next big step/decision.
   * kafka mechanism which will allow us to relay stats from collectd to the 
VES mapping plugin - ready
  * [Action] - Maryam to follow up on topics subscription for aggregation for 
clusters.
   * Schema update through configuration is under investigation.


Future discussion items:
* Inventory management - in scope for barometer or some other OPNFV project - 
discussion.

BR
Maryam
___
opnfv-tech-discuss mailing list
opnfv-tech-discuss@lists.opnfv.org
https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss


[opnfv-tech-discuss] [barometer] Update

2017-06-20 Thread Tahhan, Maryam
Hi folks
[barometer] Update 

Shout outs
* Bryan, Aimee and Volodymyr for their work on the VES Demo.

* D3:
  * CI jobs were green till 2 days ago - the fail seemed to be related to 
yardstick
  * Looks like there might be more delays due to environment instability

* E Release
  * Next Milestone is MS 5 - 7/21/2017 - scenario integration

* Dev update: 
https://wiki.opnfv.org/display/fastpath/Barometer+Development+Updates

* Test Update:

* E release

* porting all the scripts to work on Apex

* working on integrating the RPMs for the plugins in the environment
* [ACTION] Maryam to bottom out on timing for APEX scenarios and 
installations - on hold till next week.
* [ACTION] Maryam to arrange readout from John on Apex and integration work 
done to date.

* CI - RPM builds
   * https://gerrit.opnfv.org/gerrit/#/c/35831/4
   * [ACTION] Maryam to follow up with Aric/Fatih on storage - note sent to 
Aric, need to follow up.

VES
* VES 5.1 integration merged to barometer repo - for Beijing Demo.
   * Some bugs still open as regards some of the values that were being pushed.
   * Schema has been published for a while - need to look at this from an 
information model perspective, engaging an SDO.
   * Modelling info based on type vs significance? Next big step/decision.
   * kafka mechanism which will allow us to relay stats from collectd to the 
VES mapping plugin - under internal review.
  * [Action] - Maryam to follow up on topics subscription for aggregation for 
clusters.
   * Schema update through configuration is under investigation.

* ETSI document TST008:
   * Officially part of release 2 - officially opened for maintenance
   * [ACTION] - Maryam to do gap analysis with the provided telemetry and 
present back to barometer.

* OpenStack
* Monasca integration
   * Trying to bring up a Monasca setup


* Comparative Agent:
  * Looking at how much resources agents in run time will use when gathering 
metrics
  * 
https://wiki.opnfv.org/display/fastpath/Runtime+analysis+of+the+monitoring+agents+methodology
  * Need to make sure we have a reasonable default baseline
  * Need to also make sure we have small representative workloads - so we 
can determine the impact...
  * Maryam has feedback on the Monitoring Agent inference questions deck -  
[ACTION] Maryam to send out in a separate email with comments inline in the 
Deck.
 * Process to be followed:
1.  Isolate the CPUs on the monitoring node. [ Added isolcpus option in the 
grub]
2.  Run collectd on the isolated CPU. [ Used taskset command to run 
collectd with appropriate CPU-mask]
3.  Plugins: Make collectd to monitor following metrics [CPU, Memory, Disk, 
Interface, IPMI, processes, libvirt, Caches, OVS, hugepages]
4.  Output: Make collectd to send metrics to influxdb running on separate 
node.
5.  Workload: stress-ng + iperf.
6.  Monitoring duration: 5 minutes.
7.  Frequency: 1sec, 10 seconds, 60 seconds.
8.  Collected Metrics to analyze collectd's runtime performance  [ Used 
Snap to collect 'collectd-process' metrics and CPU and memory data]
9.  Note the iperf performance ( to study any effect on it due to collectd]
10.   Currently seeing if I can get more information from LTT-NG.

*** Repeat the above process for other monitoring agents ***
 * [ACTION] Sridhar to share the the LTTng output
 * [ACTION] Maryam to add a comment on baseline benchmarking and add 5.a. 
and with workload 5.b -

Next Week
* Gurpreet - demo of NFVI telemetry tool
* Inventory management - in scope for barometer or some other OPNFV project - 
discussion.

BR
Maryam
___
opnfv-tech-discuss mailing list
opnfv-tech-discuss@lists.opnfv.org
https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss


[opnfv-tech-discuss] [barometer] Update

2017-05-31 Thread Tahhan, Maryam
Hi folks
[barometer] Update 

* Dev update: 
https://wiki.opnfv.org/display/fastpath/Barometer+Development+Updates

* Test Update:

* E release
   * Base Apex environment brought up in the Lab  - Excellent base for future 
functests.
* [ACTION] Maryam to bottom out on timing for APEX scenarios and 
installations - Meeting setup for Thurs

* CI - RPM builds
   * CI job moved to run on a CentOS platform and pushes RPMs to 
artifacts.opnfv.org nightly
   * TODO Maryam to add gerrit review.
   * [ACTION] Maryam to follow up with Aric/Fatih on storage.

* E Release
  * Next Milestone is MS 5 - 7/21/2017 - scenario integration
  * Test cases complete MS 6 - 8/11/2017

VES 5.0
  * pushed code for review https://gerrit.opnfv.org/gerrit/#/c/35489/
  * Ongoing discussion on 4 fields on the mailing list

* ETSI document TST008:
  * Published 
http://www.etsi.org/deliver/etsi_gs/NFV-TST/001_099/008/02.01.01_60/gs_NFV-TST008v020101p.pdf
  * [ACTION] Al to find out process for issue tracker


* Comparative Agent:
  * Looking at how much resources agents in run time will use when gathering 
metrics
  * 
https://wiki.opnfv.org/display/fastpath/Runtime+analysis+of+the+monitoring+agents+methodology
  * Need to make sure we have a reasonable default baseline
  * Need to also make sure we have small representative workloads - so we 
can determine the impact...
  * Maryam has feedback on the Monitoring Agent inference questions deck -  
[ACTION] Maryam to send out in a separate email with comments inline in the 
Deck.
 * Process to be followed:
1.  Isolate the CPUs on the monitoring node. [ Added isolcpus option in the 
grub]
2.  Run collectd on the isolated CPU. [ Used taskset command to run 
collectd with appropriate CPU-mask]
3.  Plugins: Make collectd to monitor following metrics [CPU, Memory, Disk, 
Interface, IPMI, processes, libvirt, Caches, OVS, hugepages]
4.  Output: Make collectd to send metrics to influxdb running on separate 
node.
5.  Workload: stress-ng + iperf.
6.  Monitoring duration: 5 minutes.
7.  Frequency: 1sec, 10 seconds, 60 seconds.
8.  Collected Metrics to analyze collectd's runtime performance  [ Used 
Snap to collect 'collectd-process' metrics and CPU and memory data]
9.  Note the iperf performance ( to study any effect on it due to collectd]
10.   Currently seeing if I can get more information from LTT-NG.

*** Repeat the above process for other monitoring agents ***
 * [ACTION] Sridhar to share the the LTTng output
 * [ACTION] Maryam to add a comment on baseline benchmarking and add 5.a. 
and with workload 5.b -

* D3:
  * CI jobs were green till yesterday - the fail seemed to be related to the 
scenario name, waiting on today's job to complete before submitting a ticket 
https://build.opnfv.org/ci/view/fuel/job/fuel-os-nosdn-kvm_ovs_dpdk_bar-ha-baremetal-daily-danube/
  * Final documentation is done for Danube: please review: 
https://gerrit.opnfv.org/gerrit/#/c/35185/

* Wiki updates [ACTION] ALL to review and provide feedback/requirements

* Collectd 101

  * Collectd how to implement a simple 
plugin

  * Installing and configuring InfluxDB and Grafana to display metrics with 
collectd

  * Monitoring, Metrics and Events Requirements High Level 
List


* Kafka benchmarking - Installer support gap - had ran a bunch of tests - but 
need to go through these again.

* OpenStack Boston
* Monasca integration is targeted
   *  Reached out to PTL - NFVI monitoring is in scope, we should be able to 
integrate collectd as a data source for stats
   * API today only accepts stats - we need to work with the PTL on extending 
to accept notifications also.

* [ACTIONs] Maryam to :
  1. extend next week's session to 1 hour to cover the process plugin design.
 2. arrange the last 15 mins of the meeting to show  Demo from OpenStack Boston 
- https://www.youtube.com/watch?v=DmSiXS_tUEY
  3. arrange for email updates outside of the meeting

BR
Maryam
___
opnfv-tech-discuss mailing list
opnfv-tech-discuss@lists.opnfv.org
https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss


[opnfv-tech-discuss] [barometer] Update

2017-05-23 Thread Tahhan, Maryam
Hi folks
[barometer] Update 

* Dev update: 
https://wiki.opnfv.org/display/fastpath/Barometer+Development+Updates

* Test Update:

* E release
* Preparing tests for PMU plugin
* Still on Fuel setups - no APEX environments yet
* [ACTION] Maryam to bottom out on timing for APEX scenarios and 
installations.


* CI - RPM builds
* updated the ci job to run on a CentOS platform and push RPMs to 
artifacts.opnfv.org nightly
* Apex team to consume these RPMs for the APEX deployment and scenarios.

E Release
  * MS 2 passed - declaring the test cases for the E release 
https://wiki.opnfv.org/display/fastpath/Testing+with+Functest

* Systems/installation scripts:
  * updated the installation scripts for Ubuntu, centOS and RHEL to install 
collectd  + all the dependencies.
  * cleaned up any additional installation scripts

VES 5.1
  * collectd VES  plugin is currently being updated to  support the latest 
schema - 1 bug uncovered - now fixed...
  * More testing to be done.

* ETSI document TST008:
  * Spec approved last Friday with 2 change requests
  * CR1: Processor usage metrics also cover accelerated processor metrics
  * CR2: Timestamping update
   * Timestamp to be reported is that at the end of the measurement interval 
rather than the start.
   * If collection systems add a timestamp - it should be provided in addition 
to the measured timestamp.



* Comparative Agent:
  * Looking at how much resources agents in run time will use when gathering 
metrics
* Working with 3 agents at the moment.
* Open discussion on the metrics being used:
  * Created a wiki to track this: 
https://wiki.opnfv.org/display/fastpath/Runtime+analysis+of+the+monitoring+agents+methodology
 open for feedback and tidy up
  * Need to make sure we have a reasonable default baseline
  * Need to also make sure we have small representative workloads - so we 
can determine the impact... [ACTION] Sridhar to share the stress-ng  loads to 
be used and the LTTng output
  * Maryam has feedback on the Monitoring Agent inference questions deck -  
[ACTION] Maryam to send out in a separate email with comments inline in the 
Deck.

D3:
  * CI jobs were green till a functest update broke them - should be green 
again when https://gerrit.opnfv.org/gerrit/#/c/35205/ gets merged
  * final documentation is done for Danube: please review: 
https://gerrit.opnfv.org/gerrit/#/c/35185/

NEXT Week Agenda

[ACTION] Maryam to extend the call to 1 hour.

Wiki updates

 * Collectd 101

* Collectd how to implement a simple 
plugin

* Installing and configuring InfluxDB and Grafana to display metrics with 
collectd

* Monitoring, Metrics and Events Requirements High Level 
List


OpenStack Boston
* Monasca integration is targeted
   *  Reached out to PTL - NFVI monitoring is in scope, we should be able to 
integrate collectd as a data source for stats
   * API today only accepts stats - we need to work with the PTL on extending 
to accept notifications also.


Demo from OpenStack Boston - Next https://www.youtube.com/watch?v=DmSiXS_tUEY

Kafka...

BR
Maryam
___
opnfv-tech-discuss mailing list
opnfv-tech-discuss@lists.opnfv.org
https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss


Re: [opnfv-tech-discuss] [barometer] Update

2017-04-21 Thread Beierl, Mark
Hello,

I'm not part of the Barometer project, however I found this thread interesting, 
so here's my $0.02 (CAD).  On an internal project, we have started evaluating 
the use of NATS [1] vs. Kafka for high throughput collection of telemetry.  I 
don't have any stats collected at this time.

[1] http://nats.io/about/

Regards,
Mark

Mark Beierl
SW System Sr Principal Engineer
Dell EMC | Office of the CTO
mobile +1 613 314 8106
mark.bei...@dell.com

On Apr 21, 2017, at 11:45, Aaron Smith 
> wrote:

Hi Maryam,
 We have just finished a preliminary evaluation of Kakfa + collectd.
On the Kafka side, we have simulated up to 700 servers sending telemetry data
and have experimented with different broker setups with regards to number of
partitions, replication factor, etc...  What we see is average latency values 
with
traffic around 8ms.  However, the CDF has a long tail (out to 200ms) which have
not tried to address yet.

We have prototyped an event-driven, ethernet interface link status plugin for 
collectd.
The plugin produces no measurable load on the system and has a time from 
interrupt
delivery to the plugin to a consumer on the Kafka bus of around 8-10ms.  We 
would
be interested in seeing your results as well.  We are finishing up a document 
with the results.

Aaron

On Fri, Apr 21, 2017 at 8:04 AM, Tahhan, Maryam 
> wrote:
Hi folks

I will be reducing the call next week to 30mins.

[barometer] Update 

* Dev update: 
https://wiki.opnfv.org/display/fastpath/Barometer+Development+Updates

* Test Update:

  * E release
* Internally prepared scripts for IPMI functest – nearly complete
* Mastered generating RPM under fuel.
* Started preparing the Apex Setup.

* No Apex Based RPM/packaging of barometer code: RPM package build 
initiated --> in progress: https://gerrit.opnfv.org/gerrit/#/c/32797/2



* E release planning, please provide your feedback 24th April.

  * Updated the release page with a proposal: 
https://wiki.opnfv.org/display/fastpath/Barometer+Release+Plan but still 
looking for feedback/input

 * Action Maryam to add clock speed and power state (CPU + platform) to the 
list of metrics for E.

* Kafka to be added based on test results [ACTION] Maryam to follow up with the 
team – might be beyond E release


* Monitoring Agents Comparative study  
https://wiki.opnfv.org/display/fastpath/Monitoring+Agents+Comparative+Study
  * Completed answers to the question list – slides to be uploaded to the wiki.

* ETSI document TST008:
  * Barometer requirements for the metrics context reporting are now in sync 
with TST008.
  * Approval request being put forward in May.


* VES update

  * General VES code in the process of being refactored to python or Java – to 
be determined by ONAP project.
* Wiki updated for changes anticipated integration with DCAE and PNDA.

* [TODO] run some tests on the host and the guest side to determine accuracy of 
CPU metrics – In progress 
https://wiki.opnfv.org/display/fastpath/CPU+Utilisation+Case+Study
  * [ACTION] Maryam to update the wiki with results --> Done
* Results for a Test with a system (VM) under load --> Done
   * Use 10 second averages --> Done
   * create a wiki with the steps to reproduce --> Done
   * use stress as the app in the VM --> Done
* [ACTION] Maryam to create a wiki with the setup steps --> Done, 
https://wiki.opnfv.org/display/fastpath/CPU+Utilisation+case+study+Configuration
* [ACTION] Maryam to test with user time and system time
* [ACTION] Maryam to test with synced time for the 1 and 10 second interval --> 
I think this can be done with cron
* [ACTION] Maryam to check time sync between the host and the guest


Kafka:
Still investigating 
this
 kafka support in OpenStack to see what openstack projects and installers 
support it.
The next step is to install it and do some benchmarking with collectd and kafka 
then collectd and rabbitmq… note some benchmarks are included in the link 
above. --> In progress

Discussion topics wiki:
https://wiki.opnfv.org/display/fastpath/Discussion+Topics

All the best
Maryam

___
opnfv-tech-discuss mailing list
opnfv-tech-discuss@lists.opnfv.org
https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss




--
AARON SMITH

SENIOR PRINCIPAL SOFTWARE ENGINEER, NFVPE

Red Hat


314 Littleton Rd, Westford, MA 01886

aasm...@redhat.comM: 
617.877.4814

[https://www.redhat.com/files/brand/email/sig-redhat.png]

___
opnfv-tech-discuss mailing list
opnfv-tech-discuss@lists.opnfv.org

Re: [opnfv-tech-discuss] [barometer] Update

2017-04-21 Thread Aaron Smith
Hi Maryam,
 We have just finished a preliminary evaluation of Kakfa + collectd.
On the Kafka side, we have simulated up to 700 servers sending telemetry
data
and have experimented with different broker setups with regards to number
of
partitions, replication factor, etc...  What we see is average latency
values with
traffic around 8ms.  However, the CDF has a long tail (out to 200ms) which
have
not tried to address yet.

We have prototyped an event-driven, ethernet interface link status plugin
for collectd.
The plugin produces no measurable load on the system and has a time from
interrupt
delivery to the plugin to a consumer on the Kafka bus of around 8-10ms.  We
would
be interested in seeing your results as well.  We are finishing up a
document with the results.

Aaron

On Fri, Apr 21, 2017 at 8:04 AM, Tahhan, Maryam 
wrote:

> Hi folks
>
>
>
> I will be reducing the call next week to 30mins.
>
>
>
> [barometer] Update 
>
>
>
> * Dev update: https://wiki.opnfv.org/display/fastpath/Barometer+
> Development+Updates
>
>
>
> * Test Update:
>
>   * E release
>
> * Internally prepared scripts for IPMI functest – nearly complete
>
> * Mastered generating RPM under fuel.
>
> * Started preparing the Apex Setup.
>
> * No Apex Based RPM/packaging of barometer code: RPM package build
> initiated à in progress: https://gerrit.opnfv.org/gerrit/#/c/32797/2
>
>
>
> * E release planning, please provide your feedback 24th April.
>
>   * Updated the release page with a proposal: https://wiki.opnfv.org/
> display/fastpath/Barometer+Release+Plan but still looking for
> feedback/input
>
>  * Action Maryam to add clock speed and power state (CPU + platform) to
> the list of metrics for E.
>
> * Kafka to be added based on test results [ACTION] Maryam to follow up
> with the team – might be beyond E release
>
>
>
> * Monitoring Agents Comparative study  https://wiki.opnfv.org/
> display/fastpath/Monitoring+Agents+Comparative+Study
>
>   * Completed answers to the question list – slides to be uploaded to the
> wiki.
>
>
>
> * ETSI document TST008:
>
>   * Barometer requirements for the metrics context reporting are now in
> sync with TST008.
>
>   * Approval request being put forward in May.
>
>
>
> * VES update
>
>   * General VES code in the process of being refactored to python or Java
> – to be determined by ONAP project.
>
> * Wiki updated for changes anticipated integration with DCAE and PNDA.
>
>
>
> * [TODO] run some tests on the host and the guest side to determine
> accuracy of CPU metrics – In progress https://wiki.opnfv.org/
> display/fastpath/CPU+Utilisation+Case+Study
>
>   * [*ACTION*] Maryam to update the wiki with results à *Done*
>
> * Results for a Test with a system (VM) under load à *Done*
>
>* Use 10 second averages à *Done*
>
>* create a wiki with the steps to reproduce à *Done*
>
>* use stress as the app in the VM à *Done*
>
> * [*ACTION*] Maryam to create a wiki with the setup steps à *Done*,
> https://wiki.opnfv.org/display/fastpath/CPU+Utilisation+case+study+
> Configuration
>
> * [*ACTION*] Maryam to test with user time and system time
>
> * [*ACTION*] Maryam to test with synced time for the 1 and 10 second
> interval à I think this can be done with cron
>
> * [*ACTION*] Maryam to check time sync between the host and the guest
>
>
>
> Kafka:
>
> Still investigating this
> 
>  kafka
> support in OpenStack to see what openstack projects and installers support
> it.
>
> The next step is to install it and do some benchmarking with collectd and
> kafka then collectd and rabbitmq… note some benchmarks are included in the
> link above. à In progress
>
>
>
> Discussion topics wiki:
>
> https://wiki.opnfv.org/display/fastpath/Discussion+Topics
>
>
>
> All the best
>
> Maryam
>
> ___
> opnfv-tech-discuss mailing list
> opnfv-tech-discuss@lists.opnfv.org
> https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss
>
>


-- 

AARON SMITH

SENIOR PRINCIPAL SOFTWARE ENGINEER, NFVPE

Red Hat



314 Littleton Rd, Westford, MA 01886

aasm...@redhat.comM: 617.877.4814

___
opnfv-tech-discuss mailing list
opnfv-tech-discuss@lists.opnfv.org
https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss


[opnfv-tech-discuss] [barometer] Update

2017-04-21 Thread Tahhan, Maryam
Hi folks

I will be reducing the call next week to 30mins.

[barometer] Update 

* Dev update: 
https://wiki.opnfv.org/display/fastpath/Barometer+Development+Updates

* Test Update:

  * E release
* Internally prepared scripts for IPMI functest - nearly complete
* Mastered generating RPM under fuel.
* Started preparing the Apex Setup.

* No Apex Based RPM/packaging of barometer code: RPM package build 
initiated --> in progress: https://gerrit.opnfv.org/gerrit/#/c/32797/2



* E release planning, please provide your feedback 24th April.

  * Updated the release page with a proposal: 
https://wiki.opnfv.org/display/fastpath/Barometer+Release+Plan but still 
looking for feedback/input

 * Action Maryam to add clock speed and power state (CPU + platform) to the 
list of metrics for E.

* Kafka to be added based on test results [ACTION] Maryam to follow up with the 
team - might be beyond E release


* Monitoring Agents Comparative study  
https://wiki.opnfv.org/display/fastpath/Monitoring+Agents+Comparative+Study
  * Completed answers to the question list - slides to be uploaded to the wiki.

* ETSI document TST008:
  * Barometer requirements for the metrics context reporting are now in sync 
with TST008.
  * Approval request being put forward in May.


* VES update

  * General VES code in the process of being refactored to python or Java - to 
be determined by ONAP project.
* Wiki updated for changes anticipated integration with DCAE and PNDA.

* [TODO] run some tests on the host and the guest side to determine accuracy of 
CPU metrics - In progress 
https://wiki.opnfv.org/display/fastpath/CPU+Utilisation+Case+Study
  * [ACTION] Maryam to update the wiki with results --> Done
* Results for a Test with a system (VM) under load --> Done
   * Use 10 second averages --> Done
   * create a wiki with the steps to reproduce --> Done
   * use stress as the app in the VM --> Done
* [ACTION] Maryam to create a wiki with the setup steps --> Done, 
https://wiki.opnfv.org/display/fastpath/CPU+Utilisation+case+study+Configuration
* [ACTION] Maryam to test with user time and system time
* [ACTION] Maryam to test with synced time for the 1 and 10 second interval --> 
I think this can be done with cron
* [ACTION] Maryam to check time sync between the host and the guest


Kafka:
Still investigating 
this
 kafka support in OpenStack to see what openstack projects and installers 
support it.
The next step is to install it and do some benchmarking with collectd and kafka 
then collectd and rabbitmq... note some benchmarks are included in the link 
above. --> In progress

Discussion topics wiki:
https://wiki.opnfv.org/display/fastpath/Discussion+Topics

All the best
Maryam
___
opnfv-tech-discuss mailing list
opnfv-tech-discuss@lists.opnfv.org
https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss


Re: [opnfv-tech-discuss] [barometer] Update

2017-03-21 Thread Tahhan, Maryam
Hi folks

Heads up - Maryam OOO March 30th 2017 - April 10th 2017 with no access to email.

[barometer] Update 

* Dev update: 
https://wiki.opnfv.org/display/fastpath/Barometer+Danube+Development+Updates
  * Documentation update on dpdkstat initialization merged: 
https://github.com/collectd/collectd/commit/8be4f037005245971f3ffee6623e69226790ca72
  * OVS plugins fixed to only be configured on supported platforms if the 
dependency is met: 
https://github.com/collectd/collectd/commit/0400307a674fcf134fba40aa4129b8ea6b09b098
  * Change OVS events plugin default behavior - from poll driven to events 
driven 
https://github.com/collectd/collectd/commit/1d49aafd2022aa3b03e8fc79edfd6f8da00b9405

* Test Update:
   * Starting to look at testing the new features for the OPNFV E release.
   * Looking at building the Fuel plugin and packaging.
   * Still have confidence that the barometer only functest tests are passing. 
Full scenario having issues deploying the scenario at the moment.
 * There was a successful pass on March 18th - but there are still issues 
with deploying on the Virtual environment.

* ETSI document TST008:

  * Al has updated 
https://wiki.opnfv.org/display/fastpath/Metrics+and+Events+Requirements+High+Level+List
 for closer alignment with TST008 - Thanks Al.

  * The wiki now includes the list of fields that need to be sent with metrics.

* E release planning, please provide your feedback.

  * Updated the release page with a proposal: 
https://wiki.opnfv.org/display/fastpath/Barometer+Release+Plan but still 
looking for feedback/input

  * Alignment with ETSI TST008 - needs to be added [ACTION] Maryam - DONE

  * Plugin descriptions to be added to the wiki [ACTION] Maryam - DONE

* Monitoring Agents Comparative study  
https://wiki.opnfv.org/display/fastpath/Monitoring+Agents+Comparative+Study


* [TODO] run some tests on the host and the guest side to determine accuracy of 
CPU metrics - In progress 
https://wiki.opnfv.org/display/fastpath/CPU+Utilisation+Case+Study
  * [ACTION] Maryam to update the wiki with  --> still to be done.
* Results for a Test with a system (VM) under load
   * Use 10 second averages
   * separate out user time and system time
   * create a wiki with the steps to reproduce
   * use stress as the app in the VM




Kafka:
Still investigating 
this
 kafka support in OpenStack to see what openstack projects and installers 
support it.
The next step is to install it and do some benchmarking with collectd and kafka 
then collectd and rabbitmq... note some benchmarks are included in the link 
above. --> In progress

Discussion topics wiki:
https://wiki.opnfv.org/display/fastpath/Discussion+Topics

All the best
Maryam
___
opnfv-tech-discuss mailing list
opnfv-tech-discuss@lists.opnfv.org
https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss


Re: [opnfv-tech-discuss] [barometer] Update

2017-03-21 Thread Morales, Victor
Hi there,

AFAIR, there were some issues in Monasca to keep support to this driver[1] and 
Mehdi was working in solving some of the roadblocks[2], more likely he can 
provide some status about the current state and future plans of that.

Regards/Saludos
Victor Morales
irc: electrocucaracha

[1] http://lists.openstack.org/pipermail/openstack-dev/2016-December/108245.html
[2] https://review.openstack.org/#/q/topic:sileht/kafka-fix

From: <opnfv-tech-discuss-boun...@lists.opnfv.org> on behalf of "Tahhan, 
Maryam" <maryam.tah...@intel.com>
Date: Tuesday, March 21, 2017 at 10:10 AM
To: "opnfv-tech-discuss@lists.opnfv.org" <opnfv-tech-discuss@lists.opnfv.org>, 
"Power, Damien" <damien.po...@intel.com>, "Mcmahon, Tony B" 
<tony.b.mcma...@intel.com>
Subject: Re: [opnfv-tech-discuss] [barometer] Update 


Kafka:
Investigating 
this<https://specs.openstack.org/openstack/oslo-specs/specs/liberty/adding-kafka-support.html>
 kafka support in OpenStack to see what openstack projects and installers 
support it.
The next step is to install it and do some benchmarking with collectd and kafka 
then collectd and rabbitmq… note some benchmarks are included in the link above.
___
opnfv-tech-discuss mailing list
opnfv-tech-discuss@lists.opnfv.org
https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss


Re: [opnfv-tech-discuss] [barometer] Update

2017-03-21 Thread Tahhan, Maryam
Hi folks
Sorry for the delay in sending these out

[barometer] Update 

* Dev update: 
https://wiki.opnfv.org/display/fastpath/Barometer+Danube+Development+Updates

* Test Update:
   * Two patches submitted to correct - issue blocking the execution to obtain 
ssh keys for fuel master using the functest APIs is now resolved.
   * Enhance mcelog tests - adding the tool to inject events to be able to 
monitor and test - tested offline, not In functest yet
   * Tests for release D are complete now and all tests are passing.
   * Barometer combined (KVM4NFV and OVS4NFV) scenario testing - can reproduce 
the failure manually, look at reproducing with CI, in the process of being root 
caused.

* ETSI document:
  * Standardize key measurements of the infrastructure - definitions are 
formally written.
  * To what extent are the metrics defined in the spec supported in existing 
agents...? Need to explore the alignment and where we are unaligned - need to 
address as part of the E release.
  * 
https://wiki.opnfv.org/display/fastpath/Metrics+and+Events+Requirements+High+Level+List

* Monitoring Agents Comparative study  
https://wiki.opnfv.org/display/fastpath/Monitoring+Agents+Comparative+Study
  * Snap info added in. remaining 6/7 plugins only.

* [TODO] run some tests on the host and the guest side to determine accuracy of 
CPU metrics - In progress 
https://wiki.opnfv.org/display/fastpath/CPU+Utilisation+Case+Study
  * [ACTION] Maryam to update the wiki with
* Results for a Test with a system (VM) under load
   * Use 10 second averages
   * separate out user time and system time
   * create a wiki with the steps to reproduce
   * use stress as the app in the VM



* E release planning, please provide your feedback.

  * updated the release page with a proposal: 
https://wiki.opnfv.org/display/fastpath/Barometer+Release+Plan but still 
looking for feedback/input

  * Alignment with ETSI TST008 - needs to be added [ACTION] Maryam

  * Plugin descriptions to be added to the wiki [ACTION] Maryam





Kafka:
Investigating 
this
 kafka support in OpenStack to see what openstack projects and installers 
support it.
The next step is to install it and do some benchmarking with collectd and kafka 
then collectd and rabbitmq... note some benchmarks are included in the link 
above.

Discussion topics wiki:
https://wiki.opnfv.org/display/fastpath/Discussion+Topics

All the best
Maryam


___
opnfv-tech-discuss mailing list
opnfv-tech-discuss@lists.opnfv.org
https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss


Re: [opnfv-tech-discuss] [barometer] Update

2017-03-09 Thread MORTON, ALFRED C (AL)
Hi Maryam,
thanks for your notes for the week, apologies for
delay replying to a couple of items below:

* ETSI document: whether we can make the same measurements on virtual CPUs as 
we do physical CPUs.
   

Maryam's research into libvirt's vcpu measures was very helpful
(see her message of 2/15), and we built a convincing example
(in an Informative Annex) showing how the *Scope* metric
parameter can be used to specify either physical Processor or
Virtual Processor measurements. The libvirt domain is the equivalent
of a Virtual Machine Instance ID, and that's what some want to
measure.  ETSI has moved re-named some of their directories,
and the latest draft (0.0.5) of TST008 (the doc we're talking about)
is available here:  
https://docbox.etsi.org/ISG/NFV/Open/Drafts/TST008_NFVI_Metrics_Spec

So, now that we've got fairly solid (ETSI TST WG-approved) definitions for
key metrics, one question came up pertaining to the supporting
information for the metrics - can we collect the Parameters (input
factors) and the Scope along with the measured results at the VIM?

For example, for Processor Utilization, the parameters are:
- Tick Interval (period of interrupts to read Execution context)
- Set of Execution Contexts (likely to be included with the results)
- Start time for the measurement interval
- Measurement Interval Duration
and the Scope should also be clear in the reported results
(but this needs checking).
It seems as though all Parameters should be available (or derivable)
from most forms of measurement collection. Extracting the Tick Interval
might be a current Gap, but since the Tick is static, it should be
surmountable.

There's some synergy with looking at the measurements Maryam reported:
* [TODO] run some tests on the host and the guest side to determine accuracy of 
CPU metrics - In progress 
https://wiki.opnfv.org/display/fastpath/CPU+Utilisation+Case+Study<https://urldefense.proofpoint.com/v2/url?u=https-3A__wiki.opnfv.org_display_fastpath_CPU-26-2343-3BUtilisation-26-2343-3BCase-26-2343-3BStudy=DQMFAg=LFYZ-o9_HUMeMTSQicvjIg=OrbtGCluczz9awEKz9Fv7g=Hz3DosIUyqZlG30zAFpfeH09_SF9q0D6Ke5BpWXTlhY=qdVk_9tMe78fFjQpkECscj3ywYNZPcq1jSuoU3ZTL4E=>
 I would appreciate any input/feedback as I'm finding some unusual anomalies. :)
so I plan to have a look at the Wiki next,
Al

From: opnfv-tech-discuss-boun...@lists.opnfv.org 
[mailto:opnfv-tech-discuss-boun...@lists.opnfv.org] On Behalf Of SULLIVAN, 
BRYAN L
Sent: Tuesday, March 07, 2017 12:03 PM
To: Tahhan, Maryam; 'opnfv-tech-discuss@lists.opnfv.org'; Mcmahon, Tony B; 
Power, Damien
Subject: Re: [opnfv-tech-discuss] [barometer] Update 

***Security Advisory: This Message Originated Outside of AT ***
Reference http://cso.att.com/EmailSecurity/IDSP.html for more information.
FYI, I submitted a patch to move the ves_plugin to the v3 of the VES API.
https://gerrit.opnfv.org/gerrit/#/c/29959/<https://urldefense.proofpoint.com/v2/url?u=https-3A__gerrit.opnfv.org_gerrit_-23_c_29959_=DQMFAg=LFYZ-o9_HUMeMTSQicvjIg=OfsSu8kTIltVyD1oL72cBw=xTS8NKGdsdp_MVl9Z-CpskVrP51XG5Nwe0-Xnt0D93o=FW7cEy27YppnzreMj7a_wppjRi98XUUDQeRudm6C9BU=>

We have rebased the VES code for Danube on the latest AT release on github 
and integrated the current version of the ves_plugin. But we need the plugin to 
use the "v3" API endpoint instead of "v1".

Thanks,
Bryan Sullivan | AT

From: 
opnfv-tech-discuss-boun...@lists.opnfv.org<mailto:opnfv-tech-discuss-boun...@lists.opnfv.org>
 [mailto:opnfv-tech-discuss-boun...@lists.opnfv.org] On Behalf Of Tahhan, Maryam
Sent: Tuesday, March 07, 2017 8:56 AM
To: 'opnfv-tech-discuss@lists.opnfv.org' 
<opnfv-tech-discuss@lists.opnfv.org<mailto:opnfv-tech-discuss@lists.opnfv.org>>;
 Mcmahon, Tony B <tony.b.mcma...@intel.com<mailto:tony.b.mcma...@intel.com>>; 
Power, Damien <damien.po...@intel.com<mailto:damien.po...@intel.com>>
Subject: [opnfv-tech-discuss] [barometer] Update 

Hi folks

No call today due to cancellations...

[barometer] Update 

[ACTION]: Maryam to add collectd usecase diagrams to wiki. --> done 
https://wiki.opnfv.org/display/fastpath/Collectd+Abstraction+Layer+usecases<https://urldefense.proofpoint.com/v2/url?u=https-3A__wiki.opnfv.org_display_fastpath_Collectd-26-2343-3BAbstraction-26-2343-3BLayer-26-2343-3Busecases=DQMFAg=LFYZ-o9_HUMeMTSQicvjIg=OrbtGCluczz9awEKz9Fv7g=Hz3DosIUyqZlG30zAFpfeH09_SF9q0D6Ke5BpWXTlhY=M9peyt6gAA02DGMKlA7GjRDTYlk9LGw2dIMcjew85Zo=>
 there are some more diagrams to be added in based on the agent discussion
[ACTION]: Maryam to follow up on kafka bus support--> see below
[ACTION]: All - to review and update documentation
Node up/down status work --> [ACTION] Maryam to send out more info on this 
separately --> done, please see the email on host health

* Dev update: 
https://wiki.opnfv.org/display/fastpath/Barometer+Danube+Development+Updates<https://urldefense.proofpoint.com/v2/url?u=https-3A__wiki.opnfv.org_

Re: [opnfv-tech-discuss] [barometer] Update

2017-03-07 Thread SULLIVAN, BRYAN L
FYI, I submitted a patch to move the ves_plugin to the v3 of the VES API.
https://gerrit.opnfv.org/gerrit/#/c/29959/

We have rebased the VES code for Danube on the latest AT release on github 
and integrated the current version of the ves_plugin. But we need the plugin to 
use the "v3" API endpoint instead of "v1".

Thanks,
Bryan Sullivan | AT

From: opnfv-tech-discuss-boun...@lists.opnfv.org 
[mailto:opnfv-tech-discuss-boun...@lists.opnfv.org] On Behalf Of Tahhan, Maryam
Sent: Tuesday, March 07, 2017 8:56 AM
To: 'opnfv-tech-discuss@lists.opnfv.org' <opnfv-tech-discuss@lists.opnfv.org>; 
Mcmahon, Tony B <tony.b.mcma...@intel.com>; Power, Damien 
<damien.po...@intel.com>
Subject: [opnfv-tech-discuss] [barometer] Update 

Hi folks

No call today due to cancellations...

[barometer] Update 

[ACTION]: Maryam to add collectd usecase diagrams to wiki. --> done 
https://wiki.opnfv.org/display/fastpath/Collectd+Abstraction+Layer+usecases<https://urldefense.proofpoint.com/v2/url?u=https-3A__wiki.opnfv.org_display_fastpath_Collectd-26-2343-3BAbstraction-26-2343-3BLayer-26-2343-3Busecases=DQMFAg=LFYZ-o9_HUMeMTSQicvjIg=OrbtGCluczz9awEKz9Fv7g=Hz3DosIUyqZlG30zAFpfeH09_SF9q0D6Ke5BpWXTlhY=M9peyt6gAA02DGMKlA7GjRDTYlk9LGw2dIMcjew85Zo=>
 there are some more diagrams to be added in based on the agent discussion
[ACTION]: Maryam to follow up on kafka bus support--> see below
[ACTION]: All - to review and update documentation
Node up/down status work --> [ACTION] Maryam to send out more info on this 
separately --> done, please see the email on host health

* Dev update: 
https://wiki.opnfv.org/display/fastpath/Barometer+Danube+Development+Updates<https://urldefense.proofpoint.com/v2/url?u=https-3A__wiki.opnfv.org_display_fastpath_Barometer-26-2343-3BDanube-26-2343-3BDevelopment-26-2343-3BUpdates=DQMFAg=LFYZ-o9_HUMeMTSQicvjIg=OrbtGCluczz9awEKz9Fv7g=Hz3DosIUyqZlG30zAFpfeH09_SF9q0D6Ke5BpWXTlhY=wpqINCtEbpolJ_r0D5VAE5NMgRRcc0LpLTI58EnmLJY=>

* Test Update: 

* ETSI document: whether we can make the same measurements on virtual CPUs as 
we do physical CPUs. 

* [TODO] run some tests on the host and the guest side to determine accuracy of 
CPU metrics - In progress 
https://wiki.opnfv.org/display/fastpath/CPU+Utilisation+Case+Study<https://urldefense.proofpoint.com/v2/url?u=https-3A__wiki.opnfv.org_display_fastpath_CPU-26-2343-3BUtilisation-26-2343-3BCase-26-2343-3BStudy=DQMFAg=LFYZ-o9_HUMeMTSQicvjIg=OrbtGCluczz9awEKz9Fv7g=Hz3DosIUyqZlG30zAFpfeH09_SF9q0D6Ke5BpWXTlhY=qdVk_9tMe78fFjQpkECscj3ywYNZPcq1jSuoU3ZTL4E=>
 I would appreciate any input/feedback as I'm finding some unusual anomalies. :)

* Monitoring Agents Comparative study  
https://wiki.opnfv.org/display/fastpath/Monitoring+Agents+Comparative+Study<https://urldefense.proofpoint.com/v2/url?u=https-3A__wiki.opnfv.org_display_fastpath_Monitoring-26-2343-3BAgents-26-2343-3BComparative-26-2343-3BStudy=DQMFAg=LFYZ-o9_HUMeMTSQicvjIg=OrbtGCluczz9awEKz9Fv7g=Hz3DosIUyqZlG30zAFpfeH09_SF9q0D6Ke5BpWXTlhY=pNnvJNztxzlmqs0HLjRAu4aXl5kJIlfbvXuIEAvjpqw=>
  Update from Sridhar (thank you Sridhar)

1.   completed (well, almost - except for runtime parameters) for 7 OSS 
agents.

2.   added 4 more rows of parameters.

3.   started with SNAP now.

  Queries:

1.   Should we consider including Intel PCM metrics collection in Collectd?

2.  Will anybody be interested in sharing their inputs to the related work 
in VSPERF - event correlation 
(https://wiki.opnfv.org/pages/viewpage.action?pageId=10289941<https://urldefense.proofpoint.com/v2/url?u=https-3A__wiki.opnfv.org_pages_viewpage.action-3FpageId-3D10289941=DQMFAg=LFYZ-o9_HUMeMTSQicvjIg=OrbtGCluczz9awEKz9Fv7g=Hz3DosIUyqZlG30zAFpfeH09_SF9q0D6Ke5BpWXTlhY=SvmKv6nIB2gqsXubJRxs0_YKJZMHZTYbZrPbGkKDEuM=>)
 Looks interesting to me. I will have a look.



* E release planning, please provide your feedback.

  * updated the release page with a proposal: 
https://wiki.opnfv.org/display/fastpath/Barometer+Release+Plan<https://urldefense.proofpoint.com/v2/url?u=https-3A__wiki.opnfv.org_display_fastpath_Barometer-26-2343-3BRelease-26-2343-3BPlan=DQMFAg=LFYZ-o9_HUMeMTSQicvjIg=OrbtGCluczz9awEKz9Fv7g=Hz3DosIUyqZlG30zAFpfeH09_SF9q0D6Ke5BpWXTlhY=Dhif6XJYw7AQ_53Yx49rif-XgwBktBhHME6p781iyxI=>
 but still looking for feedback/input on:

* any other agents anyone is looking at?

* the next integration point for barometer...

* anything folks would like to see more/less of



Kafka:
Investigating 
this<https://urldefense.proofpoint.com/v2/url?u=https-3A__specs.openstack.org_openstack_oslo-2Dspecs_specs_liberty_adding-2Dkafka-2Dsupport.html=DQMFAg=LFYZ-o9_HUMeMTSQicvjIg=OrbtGCluczz9awEKz9Fv7g=Hz3DosIUyqZlG30zAFpfeH09_SF9q0D6Ke5BpWXTlhY=stzGLryBM5P6aiUHE2boOKzC3CGCvRCFoIiLF32uuq4=>
 kafka support in OpenStack to see what openstack projects and installers 
support it.
The next step is to install it and do some benchmarking with collectd a

[opnfv-tech-discuss] [barometer] Update

2017-03-07 Thread Tahhan, Maryam
Hi folks

No call today due to cancellations...

[barometer] Update 

[ACTION]: Maryam to add collectd usecase diagrams to wiki. --> done 
https://wiki.opnfv.org/display/fastpath/Collectd+Abstraction+Layer+usecases 
there are some more diagrams to be added in based on the agent discussion
[ACTION]: Maryam to follow up on kafka bus support--> see below
[ACTION]: All - to review and update documentation
Node up/down status work --> [ACTION] Maryam to send out more info on this 
separately --> done, please see the email on host health

* Dev update: 
https://wiki.opnfv.org/display/fastpath/Barometer+Danube+Development+Updates

* Test Update: 

* ETSI document: whether we can make the same measurements on virtual CPUs as 
we do physical CPUs. 

* [TODO] run some tests on the host and the guest side to determine accuracy of 
CPU metrics - In progress 
https://wiki.opnfv.org/display/fastpath/CPU+Utilisation+Case+Study I would 
appreciate any input/feedback as I'm finding some unusual anomalies. :)

* Monitoring Agents Comparative study  
https://wiki.opnfv.org/display/fastpath/Monitoring+Agents+Comparative+Study
  Update from Sridhar (thank you Sridhar)

1.  completed (well, almost - except for runtime parameters) for 7 OSS 
agents.

2.  added 4 more rows of parameters.

3.  started with SNAP now.

  Queries:

1.  Should we consider including Intel PCM metrics collection in Collectd?

2. Will anybody be interested in sharing their inputs to the related work 
in VSPERF - event correlation 
(https://wiki.opnfv.org/pages/viewpage.action?pageId=10289941)
 Looks interesting to me. I will have a look.



* E release planning, please provide your feedback.

  * updated the release page with a proposal: 
https://wiki.opnfv.org/display/fastpath/Barometer+Release+Plan but still 
looking for feedback/input on:

* any other agents anyone is looking at?

* the next integration point for barometer...

* anything folks would like to see more/less of



Kafka:
Investigating 
this
 kafka support in OpenStack to see what openstack projects and installers 
support it.
The next step is to install it and do some benchmarking with collectd and kafka 
then collectd and rabbitmq... note some benchmarks are included in the link 
above.

Discussion topics wiki:
https://wiki.opnfv.org/display/fastpath/Discussion+Topics

All the best
Maryam


___
opnfv-tech-discuss mailing list
opnfv-tech-discuss@lists.opnfv.org
https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss