Skip to content

Meeting Notes 2017

Andrew R. Lake edited this page Nov 2, 2017 · 32 revisions

November 2, 2017 Conference Call 3:00PM UTC/GMT/11:00AM EDT/8:00AM PDT/5:00PM CEST

  • Attendees: Andy, Mark, Erik, Antoine, Dan, Scott, Ivan, Ed, Szymon
  • Apologies:

Agenda Topics

Notes

  • 4.0.2
    • A few open issues. Mark is working on IPv6 issues and context issue. Michael is out but will look at txdrops/rxdrops reporting issue.
  • 4.1 Planning
    • pSConfig work started, mostly just infrastructure at this point, nothing to show yet
    • Will look at merging TWAMP code into OWAMP master after 4.0.2
  • Lead Update
    • Looking at putting together requirements workshop
    • Trying to plan next dev meeting, possibly in Europe
  • QA Effort - Planning Document
    • No update

October 12, 2017 Conference Call 3:00PM UTC/GMT/11:00AM EDT/8:00AM PDT/5:00PM CEST

  • Attendees: Ivana, Hakan, Ed, Ivan, Erik, Michale, Dan, Mark, Szymon
  • Apologies:

Agenda Topics

  • 4.0.2
  • 4.1 Planning
  • Lead Update
  • QA Effort - Planning Document
  • TechEx 2017 - Discussion of comments on slides
  • GlobalSummit 2018 - CFP is out; do we want to do anything next year? (Mark)
  • Schedule for this meeting going forward (Andy)

Notes

  • 4.0.2
    • On track, will finalize release schedule after TechEx
  • 4.1 Planning
    • No major updates
  • Lead Update
    • Talked about techex, Cisco testing, centos7 testing, and face-to-face
  • QA Effort - Planning Document
    • No updates
  • TechEx 2017 - Discussion of comments on slides
    • Mark taking pass today
  • GlobalSummit 2018 - CFP is out; do we want to do anything next year? (Mark)
    • agreed we should at least get a slot. Mark will submit. Group of us can work on description if needed at TechEx.
  • Schedule for this meeting going forward (Andy)
    • Next meeting on 11/2, will be every 2 weeks thereafter

September 28, 2017 Conference Call 3:00PM UTC/GMT/11:00AM EDT/8:00AM PDT/5:00PM CEST

  • Attendees: Andy, Mark, Szymon, Valentin, Erik, Shawn, Dan, Michael, Antoine, Ivan, Scott C., Ed
  • Apologies:

Agenda Topics

Notes

  • 4.0.2
    • On track for Monday announcement of beta 1
  • 4.1 Planning
    • No updates
  • Lead Update
    • No updates
  • QA Effort - Planning Document
    • No updates
  • MeshConfig GUI issue
    • Michael will create issue on missing interfaces in meshconfig gui

September 21, 2017 Conference Call 3:00PM UTC/GMT/11:00AM EDT/8:00AM PDT/5:00PM CEST

  • Attendees: Andy, Mark, Valentin, Luke, Michael, Dan, Ivan, Antoine, Matt, Ivana
  • Apologies: Szymon

Agenda Topics

  • 4.0.2
    • On track
  • 4.1 Planning
    • On track
  • Lead Update
    • Looking at doing a requirements workshop, extremely early stages
  • QA Effort - Planning Document
    • No update
  • Workshop on Tuesday
    • Went well, had 41 people even with issues of having only 25 slots for the first part
    • Will likely do some more workshops on different topics

September 14, 2017 Conference Call 3:00PM UTC/GMT/11:00AM EDT/8:00AM PDT/5:00PM CEST

  • Attendees: Ivana, Andy, Mark, Erik, Valentin, Ivan, Matt Z., Hakan, Szymon, Dan, Michael, Ed, Antoine, Luke, Scott C.
  • Apologies:

Agenda Topics

Notes

  • 4.0.2
    • On-track for beta in early October. Last sprint ends 9/25.
  • 4.1 Planning
    • Some work started
  • Lead Update
    • Ivana showed some cool work GEANT is doing with Splunk
  • QA Effort - Planning Document
    • no updates

September 7, 2017 Conference Call 3:00PM UTC/GMT/11:00AM EDT/8:00AM PDT/5:00PM CEST

  • Attendees: Andy, Mark, Shawn, Matt Z., Erik, Scott C., Szymon, Ivan, Ed, Antoine
  • Apologies:

Agenda Topics

Notes

  • 4.0.2
    • On track
    • Ed got Ryan back, hopes to have SNMP issues resolved next week
  • 4.1 Planning
    • TWAMP effotr started
    • pSConfig libs progressing
  • Lead Update
    • Look at slides for 9/22. Want to add more why, fill-in details and maybe add some stuff from Mark's API talk
  • TWAMP
    • Working on implementing powstream version, most of talk centered around on how we need to fit into pscheduler
    • TWAMP has three sets of timestamps, with each set containing a host and network time: when sender host sent, when reflecter received when sender received. Can use these three to get an idea of unidirectional delays as well as a rtt (which is a bit different than the addition of the one-ways)
    • Most important part for plug-in is figuring out how it fits into rtt pscheduler test. Will want protocol but also will likely need some extra fields to capture unidirectional stuff
    • Might also make sense to put in latency/latencybg test and just capture one way stats. Could do similar and add section for two-way stats
    • There will likely be a rttbg test to be added as well.

August 10, 2017 Conference Call 3:00PM UTC/GMT/11:00AM EDT/8:00AM PDT/5:00PM CEST

  • Attendees: Andy, Ryan, Adilah, Scott C., Erik, Mark, Ed, Ivan, Luke, Michael, Szymon
  • Apologies: Antoine, Valentin, Eric B., Matt Z., Ivana, Dan

Agenda Topics

Notes

  • UMich perfSONAR summer wrap-up presentation (Reminder to Andy: Record this part of meeting)
  • 4.0.1 Status
    • On track for Monday or Tuesday release
  • 4.0.2
    • Officially going to do 4.0.2 in November (beta before TechEx)
    • Action: Mark stuff in ZenHub you want in 4.0.2
  • 4.1 Planning
    • Will talk more next week
  • Lead Update
    • Agreed 4.0.2 a good idea
    • NRP meeting had lots of perfSONAR mentions
    • We will give a webinar on 9/22 to CI group on upcoming 4.1. Coordinated by Jason Z.
  • QA Effort - Planning Document
    • Action: We should review 4.1 QA tasks so we are sure everything makes sense.

August 3, 2017 Conference Call 3:00PM UTC/GMT/11:00AM EDT/8:00AM PDT/5:00PM CEST

  • Attendees: Dan, Mark, Ivan, Scott, Szymon, Erik, Matt, Shawn, Ed
  • Apologies: Luke, Michael

Agenda Topics

  • 4.0.1 Status

    • 4.0.1 ZenHub board
      • About 10% of Shawn's host upgraded, seems to be working, Mark to help debug run state issues possibly unrelated to upgrade
      • Several hosts upgraded according to LS, haven't heard complaints on list
      • Should be on track for official 4.0.1 release next week
  • 4.1 Planning

  • Lead Update

    • Meeting canceled, no update
  • QA Effort - Planning Document

    • Skipped

Notes

For future meeting:

  • Testbed Next Steps - Andy, Michael

    • Skipped
  • UMich interns wrapping up their work for Ed. Would like to do a presentation up to 30m at the start of next developer meeting Aug 10th.

July 27, 2017 Conference Call 3:00PM UTC/GMT/11:00AM EDT/8:00AM PDT/5:00PM CEST

  • Attendees: Eric K., Michael, Dan, Mark, Matt, Marcos, Ivan, Ivana, Szymon, Scott C.
  • Apologies: Antoine

Agenda Topics

Notes

  • 4.0.1 Status
    • Email going out today after Andy confirms status of Debian packages with Antoine and Valentin
  • 4.1
    • In two weeks start going through epic issues
  • Lead update
    • Not much to report, confirmed 4.0.1 is ready
  • QA effort
    • No updates
  • Meetings next week
    • Dan will run on Thursday unless not much to discuss
    • Plan on having Monday and Wednesday meetings, but Dan and Andy out so cancel if not much to discuss

July 13, 2017 Conference Call 3:00PM UTC/GMT/11:00AM EDT/8:00AM PDT/5:00PM CEST

  • Attendees: Andy, Valentin, Ivana, Michael, Luke, Ed, Erik, Scott C., Scott H., Mark, Antoine, Hakan, Ivan, Shawn
  • Apologies: Szymon, Eric

Agenda Topics

Notes

  • 4.0.1 Status
    • Sprint finishes monday, probably announce first rc1 within the next couple weeks
  • 4.1 Planning
    • No updates
  • Lead Update
    • No updates
  • QA Effort - Planning Document
    • Graphs and toolkit stuff being tested
    • Libs of all pscheduler plug-in type. latency, latencybg test plugins and tools now all have unit tests
    • Mark will pick back up on remaining plug-ins next sprint

June 29, 2017 Conference Call 3:00PM UTC/GMT/11:00AM EDT/8:00AM PDT/5:00PM CEST

  • Attendees: Luke, Andy, Mark, Michael, Dan, Hakan, Scott H, Erik, Ivan, Marcos
  • Apologies: Matt Z.

Agenda Topics

Notes

  • 4.0.1 Status
    • Sprint ends Monday, last sprint starts after that
  • 4.1 Planning
    • Leads still need to review
  • Lead Update
    • Mostly focused on Friday's steering committee call
  • QA Effort - [Planning Document]
    • More tests added this week

June 22, 2017 Conference Call 3:00PM UTC/GMT/11:00AM EDT/8:00AM PDT/5:00PM CEST

  • Attendees: Andy, Michael, Scott H., Dan, Luke, Mark, Hakan, Ivan
  • Apologies:

Agenda Topics

Notes

  • 4.0.1 Status
    • Started new sprint Monday
    • Decided to do one more sprint after this, then will start packaging and getting ready for RC1
  • 4.1 Planning
    • Everything input, too much tasks as expected and people (especially Michael) way oversubscribed
    • Andy went through today and moved tasks to icebox that he thinks we should drop due to lack of resources. Cut Michael's load in half.
    • Andy also handed a few thing to Valentin
    • Project plan is just about there, but need to review with leads
  • Lead Update
    • Working on steering committee report and 4.1 project plan
  • QA Effort - Planning Document
    • Michael added some graphs and toolkit tests
    • Andy added some esmond archiver tests

June 15, 2017 Conference Call 3:00PM UTC/GMT/11:00AM EDT/8:00AM PDT/5:00PM CEST

  • Attendees: Andy, Mark, Shawn, Ryan, Adilah, Antoine, Erik, Ed, Luke, Michael, Dan
  • Apologies: Matt Z.

Agenda Topics

  • 4.0.1 Status
  • 4.1 Planning
  • Lead Update
  • QA Effort - Planning Document
  • MaDDash 2.1 - Andy
  • Graph Enhancements - Michael
  • Definition of requirements per-requester limits feature.

Notes

  • 4.0.1 Status
    • Current sprint ends Monday
    • Next sprint should be last, or we need to push 4.0.1 out. Will likely need to make decisions about what to cut.
    • No pending changes that need a new rpm to be cut pre-4.0.1
    • Update went out Monday, python-lru of RPMs got missed though and went out Wed.
  • 4.1 Planning
    • Almost all issues plugged-in to github/zenhub from face-rto-face
    • Leads will review tasks again next week.
  • Lead Update
    • Working on steering committee report and 4.1 planning
  • QA Effort - Planning Document
    • QA items plugged-in for 4.1
  • MaDDash 2.1 - Andy
    • Two features: Historical dashboards and Filtering by host
    • Shawn requested to be able to diff two dashboards at moment in time. Right now diff code takes URL, so may just be able to give diff the URLs
    • Dan mentioned it would be nice to see worst state or average state (i.e. something to help see when things went wrong)
  • Graph Enhancements - Michael
    • See epic issue for list of features
    • Antoine mentioned need to think about how we add TWAMP
  • Definition of requirements per-requester limits feature.
    • Agreed that limiting total duration over a time window is desirable
    • Want ability to specify limit to specify both to group as a whole as well as any individual in the group

June 1, 2017 Conference Call 3:00PM UTC/GMT/11:00AM EDT/8:00AM PDT/5:00PM CEST

  • Attendees: Mark, Andy, Scott H., Michael, Dan, Ivan, Valentin, Matt Z., Szymon, Brian, Eric, Ed., Marcos
  • Apologies: Antoine

Agenda Topics

  • 4.0.1 Status
    • ZenHub board
    • Anything from last sprint we want to release immediately?
  • QA Effort - Planning Document
  • Face-to-Face Planning
  • Measurement Archive Performance Improvements - Mark, Andy, Valentin
  • TSDS Archiver - Dan, Andy
  • Stable Tesh Mesh vs Nightly Test Mesh - Andy/Brian

Notes

  • 4.0.1 Status
    • New RPMs going out with pscheduler-server fix
    • May do another set with traceroute and IPv6 fixes
  • QA Effort - Planning Document
    • Toolkit, esmond archiver and throughput plugins progressing
    • Dan will look at writing a script to exercise pscheduler + iperf3 options
  • Face-to-Face Planning
    • Today is last day for hotel block
    • Items to add to agenda:
      • Improved traceroute visualization
      • Possibly an item about campus deployment
  • Measurement Archive Performance Improvements - Mark, Andy, Valentin
    • testing shows python initialization primary cause of load
    • agrrement on adding a streaming mode where archiver brought up once and fed jobs
  • TSDS Archiver - Dan, Andy
    • Main goal is to get feet wet with TSDS, not ready to migrate anything in 4.1 timeframe
    • Issues created to write archiver, update nagios plugins and graphs
  • Stable Tesh Mesh vs Nightly Test Mesh - Andy/Brian
    • Action: Michael/Andy will prepare something ina couple weeks to talk about testing

May 25, 2017 Conference Call 3:00PM UTC/GMT/11:00AM EDT/8:00AM PDT/5:00PM CEST

  • Attendees: Scott, Andy, Shawn, Mark, Dan, Michael, Luke, Ed, Valentin, Eric, Matt, Antoine, Adilah, Brian, Szymon
  • Apologies:

Agenda Topics

Notes

  • 4.0.1 Status
    • Will push out updates for URL timeouts and ICMP traceroute immediately
  • QA Effort
    • John Simpkins working on checklist and submission procedure for QA checklist
  • Face-to-Face Planning
    • Be sure to email Linda (see web page Eric shared) so she can get headcount
  • Ubuntu 16/Debian 9 - Antoine and Valentin
    • Ubuntu 16 is pretty straightforward and can be done in 4.0.1 timeframe
    • Debian 9 looks straight-forward except for Django version on esmond. Need to investigate that.
    • Debian 9 not even officialy out until this summer, so definitely 4.1 timeframe.
  • Update GUI and graphs to use LS caches - Michael
    • Shared slides presenting issues
    • Updating clients will be straight-forward, need to test LS caches at scale will probably be biggest challenge
  • Apache security updates
    • Discussed whether it made sense to have things like testpoint pull in apache fixes
    • Group agreed we should pull into perfsonar-toolkit-security package

May 18, 2017 Conference Call 3:00PM UTC/GMT/11:00AM EDT/8:00AM PDT/5:00PM CEST

  • Attendees: Ivan, Luke, Michael, Dan, Brian, Mark, Valentin, Ryan, Scott H., Adilah, Hakan
  • Apologies: Szymon

Agenda Topics

Notes

  • Mailing list problems
    • Internet2 TSG working on fixing issue where intenrnet2 hosted mailing lists are not getting routed to certain domains (e.g. ESnet). Check archives in meantime.
  • 4.0.1 Status
    • Planned second sprint, finishing first
    • Any found doubled-value issue with mesh-config and fixed it
  • QA Effort - Planning Document
    • Tests a big part of next sprint
  • Face-to-Face Planning
    • Agenda looks good
    • Next week we will discuss Ubuntu 16/Debian 9 and LS caches

May 11, 2017 Conference Call 3:00PM UTC/GMT/11:00AM EDT/8:00AM PDT/5:00PM CEST

  • Attendees: Hakan, Andy, Brian, Mark, Ivan, Shawn, Michael, Antoine, Valentin, Marcos, Ryan, Adilah
  • Apologies:

Agenda Topics

Notes

  • 4.0.1 Status
    • ZenHub should be filled in, though may need some more time estimates
    • Andy working on something that broke testbed throughput dashboards
  • QA Effort - Planning Document
    • Action: Michael created and assigned issues, if you were assigned some, come up with time estimate by Monday
  • Face-to-Face Planning
    • Went through document
    • Action: If you are lead on document, put together proposal with team members using project templte. Also try to have issues in tracker (including epic issues) with time estimates.

March 2, 2017 Conference Call 4:00PM UTC/GMT/11:00AM EST/8:00AM PST/5:00PM CET

  • Attendees: Mark, Valentin, Andy, Luke, Michael, Dan, Hakan, Ed, Brian, Sowmya, Matt
  • Apologies: Szymon

Agenda Topics

Notes

  • 4.0 Project Planning
    • Reorganized issue on Monday to be more conservative
  • Dashboard Walkthrough
    • Few minor issues we are investigating
  • Test plan
    • Not adding anymore dashboards until after rc3
    • Dan finishing up ganglia analysis, Mark had good note to user list about hardware
  • Graph Walkthrough
    • Fixed tooltips, showed shared graph working
    • BRian would like more hos info by final
  • Binding and the backward compatibility hack (Mark)
    • Mark's proposal with lead-bind should work

February 23, 2017 Conference Call 4:00PM UTC/GMT/11:00AM EST/8:00AM PST/5:00PM CET

  • Attendees: Scott, Andy, Ivan, Shawn, Marcos, Valentin, Mark, Dan, Michael, Ed, Antoine, Brian, Sowmya
  • Apologies: Szymon

Agenda Topics

Notes

  • 4.0 Project Planning
    • Goal: dashboard is cleaned-up, so hoping can keep stable and do rc mid-next week hopefully
  • Test plan
    • Action: Dan do a write-up on how we analyze performance data from ganglia and go about setting new min requirements
  • Graph Walkthrough
    • Added tool, retransmits, not a lot of huge changes since last shown

February 16, 2017 Conference Call 4:00PM UTC/GMT/11:00AM EST/8:00AM PST/5:00PM CET

  • Attendees: Matt Z., Andy, Marcos, John S., Scott C., Szymon, Ivan, Hakan, Mark, Ed, antoine, Brian, Dan, Michael
  • Apologies:

Agenda Topics

Notes

  • 4.0 Project Planning
    • On non-pscheduler side, Michael working through some graph issues and Andy has a couple issues to verify
    • On pscheduler side main thing is figuring out excessive runs
    • Dashboard Walkthrough
      • Skip
    • Test plan
      • Jetstream / large test meshes (Luke)
        • Skip (no Luke)
      • Update toolkit testing checklist (Szymon)
        • John S and Scott will be testing in near future
      • Define “parameter space” to test for each testing tool (Andy)
        • Added ping, traceroute and nuttcp
      • Small-node test plan (Scott C., Dan)
        • Mesh better, a couple orange stragglers
      • Create plan for performance evaluation (Dan)
        • getting ganglia on small nodes
    • Graph Walkthrough
      • Added tool to description
    • Distribution - I2 wants to get rid of old EL5 systems that are reaching EOL.
      • Might need to upgrade svs3 before 4.0, other still probably fine til after because its el6

February 9, 2017 Conference Call 4:00PM UTC/GMT/11:00AM EST/8:00AM PST/5:00PM CET

  • Attendees: Dan, Michael, Luke, Hakan, Ivan, Antoine, Scott, John, Mark, Brian, Ed, Matt, Syzmon, Sowmya
  • Apologies: Andy

Agenda Topics

Notes

  • 4.0 Project Planning

  • Dashboard Walkthrough

    • added nuttcp mesh, seems to be working as far as running the tests go but is not archiving correctly and so dashboard shows all bad
    • udp tests having issues with German host (Dan to work with Ivan to resolve, think it's declared wrong in test config)
    • udp tests also having issues with ps deb 4.0 to ps deb 3.5, looks like the "could not decode JSON" error from bad version of iperf, Dan + Antoine to follow up
    • trace mesh showing consistent issues with qalab host, looks to be a timeout issue with tracepath, should set up a second mesh that is using traceroute instead of tracepath
    • owamp mesh looking good except for a few issues with german host. Mark to double check on the randslip changes to see if that is impacting, otherwise may need Andy to take a look at test
  • Test plan

    • Jetstream / large test meshes (Luke)
      • Nothing to report on this week
    • Update toolkit testing checklist (Szymon)
      • has been put into the wiki
      • looking for help on the command line tools
      • probably ready to start going through this, maybe next week make another cut of rc3 packages and have some people (John?) start going through them to see how well they're going
    • Define “parameter space” to test for each testing tool (Andy)
    • Small-node test plan (Scott C., Dan)
      • has been working with Mark to go through and fix a few things (hostname issue with IPv6 address, problem with limits file)
      • still need to get hosts into ganglia (waiting on Dan for this)
    • Create plan for performance evaluation (Dan)
      • have ganglia up and running
      • still need to add small nodes, want to start tracking things like package versions
  • Graph Walkthrough

    • fixed negative latency representation
    • next actions:
      • clean up mouse hover "flicker" issue
      • urls not updating when toggling values
      • looking for feedback on latency representation
    • possible issues when changing throughput test parameters

February 2, 2017 Conference Call 4:00PM UTC/GMT/11:00AM EST/8:00AM PST/5:00PM CET

  • Attendees: Scott Andy, Mark, Matt, Marcos, Dan, Luke, Ivan, Antoine, Brian, Szymon, Alex
  • Apologies:

Agenda Topics

Notes

  • 4.0 Project Planning
    • RC3 not going out this week
    • Andy will take out some details of announcement and send to list aboit delay
    • main issue is throughput finding timeslots, Mark is working on some fixes
  • Dashboard Walkthrough
    • Latency looks really good
    • Throughut encountering known issues
    • UDP hitting iperf3 bug
  • Test plan
    • Jetstream (Luke)
    • No updates
    • Define large test meshes (Luke)
    • No updates
    • Update toolkit testing checklist (Szymon)
    • Sowmya reviewed
    • Szymon will move to wiki on Monday
    • Define “parameter space” to test for each testing tool (Andy)
    • Will add traceroute and ping tests to dashboard
    • Small-node test plan (Scott C., Dan)
    • Will add some small nodes to mesh
    • Create plan for performance evaluation (Dan)
    • Ganglia running
  • Graph Walkthrough
  • Will do with michael in the next couple days

January 19, 2017 Conference Call 4:00PM UTC/GMT/11:00AM EST/8:00AM PST/5:00PM CET

  • Attendees: Andy, Mark Dan, Michael, Brian, Scott C.,Hakan, Ivan, Valentin, Matt Z., Antoine, Rade
  • Apologies:

Agenda Topics

Notes

January 5, 2017 Conference Call 4:00PM UTC/GMT/11:00AM EST/8:00AM PST/5:00PM CET

  • Attendees: Andy, Scott, Mark, Eric, Ed, Scott C, Scott H., Szymon, Michael, Matt Z., Marcos
  • Apologies: Dan, Luke

Agenda Topics

Notes

  • Dashboard Walkthrough
    • Things mostly stayed the same over vreak
    • ps-deb2 had issue with update, fixed when got back
    • perfsonar-dev IU host getting rebuilt
    • Some issue that were fixed, but broken repo didn't allow us to get them out

    After repo update need to see what is left

  • Repo distribution system failure
    • The svn repo got too big
    • Mark trimmed about 35000 files
    • Hope to have fixed today
  • Face-to-face agenda bash
    • See google doc
  • Test plan
  • Graphs
  • Fixed agent thing which got rid of duplicates. If you use the toolkit interface no agent filter just like before and probably doesn't make sense to add one.
  • Discussed whether we should split out retransmits and loss. Matt and Brian seemed to think that made sense. Itd be nice if you could collapse graphs or drag them around (latter not a priority)
  • Matt noted that when you its confusing tat the value you see is previous version
  • Consensus is that more useful to put retransmits in pop-up on throughput like it used to be
  • Action: Andy needs to look at owamp data and figure out why summarized rate doesn't match summarized lost and sent
Clone this wiki locally