Meeting Notes 2009

20091214Video

Agenda/Minutes

Attendees: Maxim, Ezra, Aaron, Eric, Andy, Jeff, Joe, Brian, Martin, Marcos
Developer Updates * Maxim: Resolution for pingER problem is not finished. Have been unable to duplicate. Working primarily on eCenter. Looking at ways to do remote administration of pS services on hosts. Need to work on functional and design specifications. * Ezra: Testing emulab architure (virtual switches), testing ciruit monitoring. Currently working on making sure emulab images have the same pS versions installed. Ezra is interested in Deb packages of pS. * Aaron: Iperf3 integration into bwctl. Prliminary version working. Working on NTP security patch. * Eric: Nothing to report. * Andy: USAtlas found processes showing processes running when they are not. Has a patch ready for testing. Working on adding value buckets into owamp MA. Database interaction of MA is slow. * Jeff: Roadmap, OSG report, Training/Workshops in the future. * Joe: LHCOPN report. Monitoring tutorial. Map of LHCOPN (graphical e2emon system). Seemed to be discontent from regionals. For circuit monitoring - (good, good, unknown, good) == good? Don't like that model. DEISA has rejected MDM because it is not multi-domain, they also do not like the GPS clock req. * Brian: Nothing to report * Martin: Nothing to report * Marcos: Put out more results from gLS testing. Please look at the wiki page.
Actions from last call: * ACTION: Everyone - please provide feedback on cfengine config. * ACTION: Maxim will work with Andy and Aaron to verify the pingER-MP fix. * ACTION: Andy will propose methods to deal with test point uniqueness. * ACTION: Jeff, Brian and Joe will work on the roadmap. * ACTION: Jon will add Aaron to Iperf project. Aaron will add autoconf tools and begin bwctl/iperf3 integration. * ACTION: Joe/Brian will consider if ESnet has resources they want to put towards the RedHat based pS-NPTk version.
pS-Performance-Toolkit - Security updates * NTP issue
- plan is to include a fix in the 3.1.2 release. Already have a script available that will dynamically upgrade a 3.1 and 3.1.1 host whenever it is rebooted. * Policy for security updates going forward
1. Need a more private list for discussion of issues
2. Publicly announce once we have solution?
3. Automatic security updates?
TestPoint Uniqueness (Andy) * Andy discovered two separate types of duplicates that might appear in serviceTest.cgi. In one case, there were two separate entries that returned the same data set, and in the other, there were two identical looking entries that returned different data. Andy thought that the former problem might be a problem in the pSB MA's SQL generation where requests for two different 'node' elements were getting mapped to only one of them. To handle the latter issue, Andy thought that he could modify the service test CGI script to merge the duplicate entries into a single element that requests multiple keys and overlays both sets of data onto the same plot. There was some discussion about how this might get rearchitected, but it was decided that these kinds of ideas would probably be more appropriate for a future "MA database architecture" meeting. Jeff and Aaron thought that Andy's short term approach seemed sound.
Future Meetings * Suggested Topics:
- MA database architecture
- gLS/hLS profiling/optimizations
- Automated updates for pS-PerfTK
- Circuit Monitoring
- Architecture for web-services based test scheduling
- Nagios on the toolkit * 1/2 day vid-conf in January * 1/2 day f2f @ SLC (Jt Techs - Sunday afternoon)
Future topics: * pSPS:
- Community handling in regular testing screens. If the user doesn't pick a community, others can'y use this mechanism to find the node
  - N.B. the node is still findable via other methods including domain/IP and it does show up in the gLS
  - Does it make sense to set a default?
  - Should there be other methods to find sans communities?
- GUIs
  - BWCTL
  - OWAMP * Google MLab
- Register each deployment (and tools on each) into pSPS
- GUI to find an NDT server. Started by Jason, still have the public vs private routing to these servers
- Make MLab data public - even better expose it via pSPS * NSF perfSONAR Workshop
- How would a researcher use pS to fulfill research goals?
- How do you get tool developers to publish via pS?
- NSF will be creating a steering committee for this workshop.
- Hopefully NSF is interested in funding pS. * Service Monitoring
- Availability (current + historic)
- Nagios Plugins
- Data monitoring vs service liveness * Config management
- Sharing service configuration through IS
- Centralized way to manage configs of similar services. * Status Collection Enhancements
- Overview: DCN Status page uses the TL1 collector to extract and display many of the Cienna counters. Currently only cares about interfaces and interface-like things.
- Similar Work: Jon Dugan's http://code.google.com/p/esxsnmp/|ESxSNMP which is a high performance SNMP polling system.
- Enhancement Ideas
  - Combine efforts to use TL1 collectors in ESxSNMP. Would use same database backend and pS interface by doing this
  - Collecting/Storing network topology. Specifically: how interfaces are connected via the switching fabric/external links and any other tie ins to the discussions in working groups.
  - Gathering and storing alarm information. Is this the same as Status? Could or should this be done in nagios.

20091130Video

Agenda/Minutes

Attendees: Ezra, Martin, Aaron, Andy, Eric, Fahad, Brian, Marcos, Maxim, Jeff, Joe
Developer Updates * Aaron - Iperf3 integration. * Andy - Histogram addition to pS-B, writing clients. Atlas support: Illinois is seeing bwctl crashing. Duplicate test issue. * Eric - Baseline performance measurements of vmware test environment. * Fahad - Testing gridftp pS service. * Brian - Talked with Jason to discuss svn branching for ESnet production environment. * Marcos - Posted results from results time from gLS. * Maxim - pingER admin GUI creates landmarks file with no hostnames which causes pingER db problems and eventual failure. * Jason - Jason is on vacation until January. His email update included pS used in BWC successfully. CMS making progress installing pS-NPTk. USAtlas having problems with:
- pS-B duplicate tests
- daemon.pl killed by kernel
- pingER dying * Jeff - Working on roadmaps. * Joe - started wiki page on cfengine configuration.
Actions from last call: * ACTION: Joe continue to refine wiki about cfengine. * ACTION: Maxim will work with Andy and Aaron to verify the pingER-MP fix. * ACTION: Andy will propose methods to deal with test point uniqueness. * ACTION: Jeff, Brian and Joe will work on the roadmap. * ACTION: Jon will add Aaron to Iperf project. Aaron will add autoconf tools and begin bwctl/iperf3 integration. * ACTION: Joe/Brian will consider if ESnet has resources they want to put towards the RedHat based pS-NPTk version.
SC Postmordem * pS, just worked.
Iperf3 - bwctl integration issues. (Aaron) * Working on integrating into bwctl. Need a library. Aaron will work on adding autotools into iperf3.
Circuit Monitoring - proposed developer meeting at Jt Techs. (FNAL?)
TestPoint Uniqueness * pS-BUOY treats node/host/addr as unique... pinger is slightly different * how do we want to deal with this? Andy will propose solution.
pS-PerfTK * Need to be prepared to do one more bug fix release of current incarnation before Jt Techs. * RedHat migration needs to be high on the priority list.
- Modular build and revision control would help.
Future topics: * pSPS:
- Community handling in regular testing screens. If the user doesn't pick a community, others can'y use this mechanism to find the node
  - N.B. the node is still findable via other methods including domain/IP and it does show up in the gLS
  - Does it make sense to set a default?
  - Should there be other methods to find sans communities?
- GUIs
  - BWCTL
  - OWAMP * Google MLab
- Register each deployment (and tools on each) into pSPS
- GUI to find an NDT server. Started by Jason, still have the public vs private routing to these servers
- Make MLab data public - even better expose it via pSPS * NSF perfSONAR Workshop
- How would a researcher use pS to fulfill research goals?
- How do you get tool developers to publish via pS?
- NSF will be creating a steering committee for this workshop.
- Hopefully NSF is interested in funding pS. * Service Monitoring
- Availability (current + historic)
- Nagios Plugins
- Data monitoring vs service liveness * Config management
- Sharing service configuration through IS
- Centralized way to manage configs of similar services. * Status Collection Enhancements
- Overview: DCN Status page uses the TL1 collector to extract and display many of the Cienna counters. Currently only cares about interfaces and interface-like things.
- Similar Work: Jon Dugan's http://code.google.com/p/esxsnmp/|ESxSNMP which is a high performance SNMP polling system.
- Enhancement Ideas
  - Combine efforts to use TL1 collectors in ESxSNMP. Would use same database backend and pS interface by doing this
  - Collecting/Storing network topology. Specifically: how interfaces are connected via the switching fabric/external links and any other tie ins to the discussions in working groups.
  - Gathering and storing alarm information. Is this the same as Status? Could or should this be done in nagios.

20091102Video

Agenda/Minutes

Attendees: Eric, Andy, Jason, Fahad, Joe, Ezra, Brian, Marcos, Jeff, Maxim
Developer Updates * Eric: Working on deployment of gLS/hLS nodes in test deployment. Nearly ready for testing. * Andy: Helping with Atlas support issues. Also looking at owamp database issues. Have a proposal for how to do owamp data. * Jason: Involved with SC, BWC, publications. Working with Atlas. Would like testing this week. Seeing pingER issues. * Fahad: Nothing to report * Joe: Nothing to report * Ezra: Preparing for SC and BWC. * Brian: Testing the latest ISO. * Marcos: Working on IS testbed. Finding performance issues due to planet-lab limitations. Attempting to set database parameters to keep planet-lab from killing jobs. * Maxim: Working on pingER issue. Suspects the issue is that the landmarks file is being zero'd out somehow. Aaron mentioned that it is actually the MP that is not running. * Jeff: Totally on SCinet.
Actions from last call: * ACTION: Joe will add something to wiki about cfengine. * ACTION: Andy will send a mail to the mailing list outlining what he was tasked with developing. Aaron/Jason will add to this to see if there is overlap. * ACTION: Aaron to organize a separate meeting with people interested in this
mailing lists * psps-dev, psps-announce, psps-users?
pS-PTK RC testing update. * How close are we to a 'production' bug-fix release?
- Plan is to attempt to get a production ISO out tomorrow for internal testing only. As long as that works in testing, it will be made available to Atlas Friday.
SC * Coordination work
- Joe/Aaron working on owamp/bwctl test coordination.
- If interesting utilization wanted across backbones, aaron/jeff will ask eli for help finding in esnet.
Future topics: * pSPS:
- Community handling in regular testing screens. If the user doesn't pick a community, others can'y use this mechanism to find the node
  - N.B. the node is still findable via other methods including domain/IP and it does show up in the gLS
  - Does it make sense to set a default?
  - Should there be other methods to find sans communities?
- GUIs
  - BWCTL
  - OWAMP * Google MLab
- Register each deployment (and tools on each) into pSPS
- GUI to find an NDT server. Started by Jason, still have the public vs private routing to these servers
- Make MLab data public - even better expose it via pSPS * NSF perfSONAR Workshop
- How would a researcher use pS to fulfill research goals?
- How do you get tool developers to publish via pS?
- NSF will be creating a steering committee for this workshop.
- Hopefully NSF is interested in funding pS. * Service Monitoring
- Availability (current + historic)
- Nagios Plugins
- Data monitoring vs service liveness * Config management
- Sharing service configuration through IS
- Centralized way to manage configs of similar services. * Status Collection Enhancements
- Overview: DCN Status page uses the TL1 collector to extract and display many of the Cienna counters. Currently only cares about interfaces and interface-like things.
- Similar Work: Jon Dugan's http://code.google.com/p/esxsnmp/|ESxSNMP which is a high performance SNMP polling system.
- Enhancement Ideas
  - Combine efforts to use TL1 collectors in ESxSNMP. Would use same database backend and pS interface by doing this
  - Collecting/Storing network topology. Specifically: how interfaces are connected via the switching fabric/external links and any other tie ins to the discussions in working groups.
  - Gathering and storing alarm information. Is this the same as Status? Could or should this be done in nagios.

ACTIONS

ACTION: Joe will add something to wiki about cfengine.
ACTION: Marcos and Eric will work together to get Marcos access to the ESnet test servers for doing LS performance testing.
ACTION: Maxim will work with Andy and Aaron to debug the pingER-MP problem.
ACTION: Jason will send the psps related publications Internet2 has been working on.

20091026Video

Agenda/Minutes

Attendees: * Marcos * Maxim * Jason * Brian * Fahad * Jeff * Joe * Inder * Aaron * Andy * Ezra
Developer Updates * Marcos
- Working on LSScalabilityInvestigation right now. See wiki page for updates.
- Talking with Martin about involving Ahmed in the process - he may start to help. * Maxim
- Nothing Major
- Will talk with Aaron about the service restarting work he is doing.
- Mini-discussion on what Aaron is doing and the other possible approaches (e.g. low level scripts to restart/watch daemons vs something more sophisticated like Nagios). Will push out to mailing list. * Jason
- BWC - Consuming most time. Will be uploading some GUIs pretty soon that are being contributed to REDDnet
- LHC Outreahch
  - USAtlas - Keeping close contact on the open critical bug. Whole group was a little ticked, but they are happy we are fixing it.
  - USCMS - Got an invite to the Tier2 Briefing 10/27 - will give a brief talk on the disk, what it is used for, and some recommendations. * Brian
- Nothing * Fahad
- Working with the service in branch
- Installing SLAC USAtlas servers * Jeff
- At SCinet Staging
- Deploying perfSONAR servers
- Note there will be jobs posted for new positions at Internet2 * Joe
- Adding 'tasks' to the Issue tracker * Inder
- Still coming up to speed, nothing new to report * Aaron
- Working w/ Andy on the USAtlas bugs
- SC development (weathermaps) * Andy
- Working w/ Aaron on the USAtlas bugs
- Setting up a pSB testing and development environment
- Looking at what it will take to implement the pieces to plot jitter * Ezra
- BWC Prep w/ Jason
- Giving a talk at SC on Phoebus and Grid FTP
Actions from last call: * ~~ESnet: add 'current' issues to issue tracker~~
pS-Ptk stability update (Andy/Aaron) * Andy: Setting up test instances at LBL. Trying to mimic the setup of USAtlas (in throughput and latency). Set up on Friday, was hoping to hang it over the weekend, but nothing happened yet. Have noticed some data points missing * Andy: Aaron ha suggested starting 2 master instances to see the effects. Both are running right now and have not done anything strange. * Aaron: Released 3.1.1a today to Philippe/Shawn at AGLT2. Hoping to get feedback on:
- If the fix solves the full ramdisk problem
- If the fix restarts down services
- If the fix does not blow away old configurations on disk * Will be ready for a release next week if this fix proves ok (need testers). * Mini discussion on what the error is vs how to address it. Aaron explains its related to how knoppix works (it needs a disk, so it mounts the ram disk). All of our stuff runs through the ramdisk to write the underlying storage (we write to a /scratch file on the real disk). Joe ponders if this fix was harder or took more time than converting to LiveCD. Aaron notes that LiveCD will do something similar. * Brian mentions that /var/mail/error is filling up with small files - mostly cron errors. Aaron thinks another cleanup script would work, Brian says we could just stop cron from emailing too. Joe mentions cfengine scripts as being a solution.
- ACTION: Joe will add something to wiki about cfengine. * Jason will relay information to USAtlas on status.
LHC Community Update (Jason) * USAtlas - Cheesed off at the error, vented to me last week. I think they just want to make sure we are are moving on it. Keeping them updated. * USCMS - Giving a mini talk on pS to this group 10/27.
Coordination of short term GUI development * ACTION: Andy will send a mail to the mailing list outlining what he was tasked with developing. Aaron/Jason will add to this to see if there is overlap.
What should the ISO do with data longer term? (Requested by Brian) * Brian: noticed that the disks are filling up (FAST) with owamp data. No ability to delete this data (yet). * Jeff: can clear out data on many different levels and solutions can be complex or simple
- Scripts to backup up
- Scripts to hot-drop and re-start services
- Databases vs files * ACTION: Aaron to organize a separate meeting with people interested in this
Terena conference submission? * Extended abstract (1-2 pages) due Nov 30th * Conference is June 1 in Lithuania * Brian wants to submit an ESnet deployment paper * Marcos has a paper on IP summarization too.
Future topics: * Google MLab
- It would be great if we could start registering all these nodes in pS.
- LS archive for finding an NDT server. How do you know what things are available 'publicly' or not.
- Pushing to make the MLab data available. It would be great if we could provide a pS interface for them. * NSF request to fund a workshop for perfSONAR.
- How would a researcher use pS to fulfill research goals?
- How do you get tool developers to publish via pS?
- NSF will be creating a steering committee for this workshop.
- Hopefully NSF is interested in funding pS. * Service Monitoring and Config management
- gLS 'monitoring' * Status collection for low-level device PMs
- The DCN status page that was put together for the monitoring is a relatively simple collector that grabs various performance counters and other data about elements on the Cienas. Right now, it confines itself to things that look like interfaces. There's a similar project that Jon Dugan has been working on for ESnet called ESxSNMP that grabs performance counters for SNMP hosts. It would be good to have a discussion about how these two monitoring infrastructures might work together. At minimum, it might be feasible to get the TL1 monitoring components to use the same metadata database and counter databases, and, implicitly, perfsonar interface.
- Another possible area of discussion include how the monitoring above can be extended to support collection and storage of topological information. The topology information could include how interfaces are connected together internally in the switch/router (via cross-connects, adaptation or whatever). It might also be good to think about how to collect information about how switches/routers are wired together.
- It'd be feasible to grab and store TL1/SNMP alarms as part of the collection framework. Though, that might be more reasonably done via nagios plugins.
- Currently in the regular testing setup screen, if a user doesn't specify a 'community', the node can not be found. We need to figure out a way to 'find' these nodes. Is there a reasonable way to 'find' them? (Perhaps if there is no 'community' specificied, it should get a 'default' of some kind?)

ACTIONS

ACTION: Joe will add something to wiki about cfengine.
ACTION: Andy will send a mail to the mailing list outlining what he was tasked with developing. Aaron/Jason will add to this to see if there is overlap.
ACTION: Aaron to organize a separate meeting with people interested in this

20091019Video

Agenda/Minutes

Attendees: Brian, Andy, Fahad, Marcos, Eric, Aaron, Inder, Joe, Jason, Ezra, Jeff, Maxim
Developer Updates

Brian: New hires Andy: Still ramping up Fahad: Working on grid-ftp service. Have not commited into svn yet. SLAC has two new pS servers. Marcos: Working on testing the LS. Eric: Open DevNet - plateform to deploy machines and routers on the internet. The purpose is to test pS protocols and scaling and use as a Q/A mechanism. Can open it to anyone in the community. If anyone wants access, contact Eric. Aaron: Working on a service watcher script for the pS-PerformanceTK. Aiming for another release for the pS-PerfTK before SC (new kernel, watcher script). Inder: Still ramping up. NML WG is progressing. Will be looking for volunteers for OGF docs. Joe: Working on CFengine scripts to run tests. Trying to understand owamp data on the ESnet backbone. Looking at adding 20 more machines. Jason: OGF, no NMC meeting but docs were not updated anyway. Infrastructure area update. He would like the documents updated more frequently. NML was not as contentious as it could have been. Finally decided on the definition of a network. Went over Martin's proposal for the service element. Need more outside participation for NML. NSI - progress is being made. Focusing on SC into the future - bandwidth challenge, Internet2 booth, SCinet. Ezra: No updates. Working on bandwidth challenge for SC. Noticing performance differences between iperf and nuttcp. iperf versioning seems to be an issue. Maxim: Not much since last. Jeff: JET pS regular tests. Review OGF documents.

Coordination of short term GUI development
What should the ISO do with data longer term? (Requested by Brian)
ESnet OpenDevNet status update
How to best use Issue tracker for "current work" that is not yet mapped to any specific release * Brian will add some things here
Terena conference submission?
Future topics: * Google MLab
- It would be great if we could start registering all these nodes in pS.
- LS archive for finding an NDT server. How do you know what things are available 'publicly' or not.
- Pushing to make the MLab data available. It would be great if we could provide a pS interface for them. * NSF request to fund a workshop for perfSONAR.
- How would a researcher use pS to fulfill research goals?
- How do you get tool developers to publish via pS?
- NSF will be creating a steering committee for this workshop.
- Hopefully NSF is interested in funding pS. * Service Monitoring and Config management
- gLS 'monitoring' * Status collection for low-level device PMs * The DCN status page that was put together for the monitoring is a relatively simple collector that grabs various performance counters and other data about elements on the Cienas. Right now, it confines itself to things that look like interfaces. There's a similar project that Jon Dugan has been working on for ESnet called ESxSNMP that grabs performance counters for SNMP hosts. It would be good to have a discussion about how these two monitoring infrastructures might work together. At minimum, it might be feasible to get the TL1 monitoring components to use the same metadata database and counter databases, and, implicitly, perfsonar interface. * Another possible area of discussion include how the monitoring above can be extended to support collection and storage of topological information. The topology information could include how interfaces are connected together internally in the switch/router (via cross-connects, adaptation or whatever). It might also be good to think about how to collect information about how switches/routers are wired together. * It'd be feasible to grab and store TL1/SNMP alarms as part of the collection framework. Though, that might be more reasonably done via nagios plugins. * Currently in the regular testing setup screen, if a user doesn't specify a 'community', the node can not be found. We need to figure out a way to 'find' these nodes. Is there a reasonable way to 'find' them? (Perhaps if there is no 'community' specificied, it should get a 'default' of some kind?)

20090921Video

Agenda/Minutes

Attendees: Jason, Rich, Aaron, Joe, Brian, Jeff, Maxim
Team Updates: * Jason: Focusing on OGF documents. Final release candidate going out this week. Performance problems with RC4 bwctl. * Rich: Trying to close out JET/LSN perfSONAR deployment plan. Rich's last day will be at the end of FMM week, he is moving on to DOE. Good luck Rich! * Aaron: Looking at performance issues with bandwidth tests. It looks likely that this is a network driver issue (potentially a VMI issue as well). Also working on performance monitoring in the ION interface. * Joe: DICE meeting update. * Brian: Helping look at RC's. Fully deployed OWAMP on ESnet. Testing of cache.pl script. * Maxim: Finishing up RPMs with 2 binary rpms. One 32-bit, one 64-bit. * Jeff: dice update and roadmap things added to the agenda.
Previous Actions * Joe will send request for implementation specific web sites in perfsonar.net space. * Joe will send email to EU participants to find out who will be at DICE meeting. * Brian will work with Marcos, Martin and perhaps Joe on scalability testing of the hLS/gLS.
Release Activities * Next RC coming out Friday. Looking at throughput performance problem. Need to discuss bug-release schedules.
Updating of roadmap: http://code.google.com/p/perfsonar-ps/wiki/PPStrawMan * Jeff asked that everyone review the roadmap and open issues so we can create a new roadmap for the next 6-12 months.
Review of DICE meeting * Joe: Overall went well, still concerned about methodology/goals across atlantic.
Update from the ecenter project * FNAL/ESnet working together on portal to pS data. Working on 1st year milestones. Main task for ESnet will be to add traceroute data to all bwctl/iperf results so you know the path for the test. FNAL milestones will be on creating prototypes to show multiple paths with traceroute and will likely include analysis. It will allow collaboration as well.
LSScalability testing update
LHC/Atlas deployments (status update) * Waiting for the final release. Will install at the testing sites soon. Will have others install within a week if all goes well.
Future topics: * Google MLab
- It would be great if we could start registering all these nodes in pS.
- LS archive for finding an NDT server. How do you know what things are available 'publicly' or not.
- Pushing to make the MLab data available. It would be great if we could provide a pS interface for them. * NSF request to fund a workshop for perfSONAR.
- How would a researcher use pS to fulfill research goals?
- How do you get tool developers to publish via pS?
- NSF will be creating a steering committee for this workshop.
- Hopefully NSF is interested in funding pS. * Service Monitoring and Config management
- gLS 'monitoring' * Status collection for low-level device PMs

The DCN status page that was put together for the monitoring is a relatively simple collector that grabs various performance counters and other data about elements on the Cienas. Right now, it confines itself to things that look like interfaces. There's a similar project that Jon Dugan has been working on for ESnet called ESxSNMP that grabs performance counters for SNMP hosts. It would be good to have a discussion about how these two monitoring infrastructures might work together. At minimum, it might be feasible to get the TL1 monitoring components to use the same metadata database and counter databases, and, implicitly, perfsonar interface. * Another possible area of discussion include how the monitoring above can be extended to support collection and storage of topological information. The topology information could include how interfaces are connected together internally in the switch/router (via cross-connects, adaptation or whatever). It might also be good to think about how to collect information about how switches/routers are wired together. * It'd be feasible to grab and store TL1/SNMP alarms as part of the collection framework. Though, that might be more reasonably done via nagios plugins. * Currently in the regular testing setup screen, if a user doesn't specify a 'community', the node can not be found. We need to figure out a way to 'find' these nodes. Is there a reasonable way to 'find' them? (Perhaps if there is no 'community' specificied, it should get a 'default' of some kind?)

Next VC will be 9/28 2:30 EDT.

ACTIONS

Brian will check on bug/svn procedures.
Jeff will update the 'java' issue tracker.

20090914Video

Agenda/Minutes

Attendees:
Review previous actions * Joe will send request for implementation specific web sites in perfsonar.net space. * Joe will send email to EU participants to find out who will be at DICE meeting. * Brian will work with Marcos, Martin and perhaps Joe on scalability testing of the hLS/gLS.
Team Updates:
Release Activities * pS-PerformanceTK * pS-PS 3.1
Testing Progress hLS/gLS scaling
LHC/Atlas deployments (status update)
Future topics: * Service Monitoring and Config management
- gLS 'monitoring' * Status collection for low-level device PMs * The DCN status page that was put together for the monitoring is a relatively simple collector that grabs various performance counters and other data about elements on the Cienas. Right now, it confines itself to things that look like interfaces. There's a similar project that Jon Dugan has been working on for ESnet called ESxSNMP that grabs performance counters for SNMP hosts. It would be good to have a discussion about how these two monitoring infrastructures might work together. At minimum, it might be feasible to get the TL1 monitoring components to use the same metadata database and counter databases, and, implicitly, perfsonar interface. * Another possible area of discussion include how the monitoring above can be extended to support collection and storage of topological information. The topology information could include how interfaces are connected together internally in the switch/router (via cross-connects, adaptation or whatever). It might also be good to think about how to collect information about how switches/routers are wired together. * It'd be feasible to grab and store TL1/SNMP alarms as part of the collection framework. Though, that might be more reasonably done via nagios plugins.
Next VC will be 9/21 2:30 EDT.

20090910Call

Agenda

Getting Started * OGF Docs (in progress)
- NM-WG
  - Base document
  - iperf instance
- NML-WG
  - pS Topology
- NMC-WG
  - Base document
  - MA Specific document
  - IS Specific document * Caveats to the perl code, etc.
Project Plans * Client libraries first? * Server stuff second?
Ideas * Wiki page of stuff * WSDL? (EU has been using for a while: see here and here) * Others?

Minutes

Attendees: * Jason - Internet2 * Aaron - Internet2 * Dan - LBL * Monte - LBL * Maxim - FNAL * Brian ESnet * Jon - ESnet

Maxim mailed out links to old PingER WSDL/SOAP UI Testing

Dan started us out: The goal of the work that LBL (e.g. Dan and Monte) are doing is to collect schemas/WSDLs for pS services to figure out the level of interoperabilty they will need with the service they are developing (GridFTP MA). Service is in python.

Jason asks if this is the same thing that SLAC is doing. Dan says the backend is the same thing, but SLAC will do a frontend in perl, LBL will be python. Code will also be useful for building other services in python.

Jason asks if this will be related to the USATLAS request/work (BNL, UMich) for GridFTP exposure in pS. Brian/Dan talk about how it fits their needs, they are not collaborating.

Dan also mentions work that someone at LBL (Josh?) is doing to make monitoring widgets using the google APIs. He may use python servers to get there.

Dan asks about testing in the framework(s). Jason points out work that was done in the EU, and Brian points out his work (here). The pSPS test harness is not about interoperability - more of a functional test.

Dan/Maxim discuss Maxim's SOAPUI WSDLs/Schemas and how they fit in with the EU testing. Maxim points out that the WSDL file (at least how he is used to using it) is pretty basic (e.g. describes the functionality and the input - a doc/lit xml message). The validation of the message is done via the schema (at least in SOAPUI). pSPS tools do not validate against a schema currently (perl support is poor - libxml can validate against an XSD, but it is slow). Dan mentions that it is possible to validate more than just a doc/lit glob in WSDL. Each service shoul be offering this to encourage interop (and to easily generate parsing libs, etc.).

Dan asks about what the expected request/response structure is in pS. Jason responds that its all related to the original service (the RRDMA) in that there is some paritally constructed md element that drives the interaction, that is matched to the backend, and then things happen w/ regards to what the message type may be. The LS is a special case (e.g. xpath/xquery). Dan sends along examples he worked on:

#
# Schema for binary range queries
#
# Author: Dan Gunter <dkgunter@lbl.gov>
# Created: 8/30/2009
#
default namespace = "http://acs.lbl.gov/~dang/brq/"

start = query

query = element brq { ns*, expr }
# namespace
ns = element ns {
   attribute prefix { xsd:QName },
   attribute value { xsd:token }
}
# expression
expr = element expr {
   ( attribute op { "and" | "or" }, expr, expr ) |
   ( attribute op { "not" }, expr ) |
   leaf
}
leaf =
   ( attribute op { "cmp" }, leaf_content )
leaf_content = ( rel, var, const )
# comparison relational operators (rel)
rel = rel_eq | rel_range | rel_re
rel_eq = attribute rel { "=" | "!=" }
rel_range = attribute rel { "<" | "<=" | ">" | ">=" }
rel_re = attribute rel { "=~" }
text_rel = ( rel_eq | rel_re )
num_rel = ( rel_eq | rel_range )
# variable
var = attribute var { text }
# const
const = attribute const { text }
num_const = attribute const { xsd:decimal }

and

#
# Example derived binary range query schema that restricts
# the elements to 'foo', which is text, and 'bar',
# which is numeric.
#
# Author: Dan Gunter <dkgunter@lbl.gov>
# Created: 8/30/2009
#

include "binary_range_query.rnc" {
   leaf_content = (
      ( text_rel, foo_var, const ) |
      ( num_rel, bar_var, num_const )
   )
}

foo_var = attribute var { "foo" }
bar_var = attribute var { "bar" }

Jason thinks these are good examples of what a query/request could be. Dan/Monte will refine more.

Dan discusses current work plan:

Branch from trunk
Jon/Monte/Dan will work from there

Jason mentions that they should expect to release the end-product as source, should make an RPM package (will be easier to include on pSPT).

Jason asks some followup questions on how the service works (to see if this would be a valuable addition to the pSPT). Service consumes gridftp logs and then exposes an MA. Does not need to be on the same machine as gridftp software (exporting via netlogger currently?).

Group decides further meetings will be on-demand.

20090824Video

Agenda/Minutes

Attendees: * Maxim - Fermilab * Fahad - SLAC * Marcos - UDel * Rich - Internet2 * Aaron - Internet2 * Jason - Internet2
Release Activities for RC3 * Open Bugs (See Jason's Note from 8/20)
- Of the bugs open for RC3 in the spreadsheet, 1/2 to 3/4 are ready to be closed pending testing. * Release Schedule
- Bugs Addressed/Tested by Weds (8/26)
  - Hard stop weds night - unfinished items will be migrated to RC4
- Disk Generated/Tested by Thurs (8/27)
  - Aaron will be testing disks locally all day Thursday
- Release to Testers by Fri (8/28)
  - Released to website/testers by morning.
Release Schedule Post RC3 * RC4 - 9/11
- Before DICE Meeting * Release - 9/23
- May slip up to Friday (9/25) of this week if needed. * Dates may slide depending on bugs/staff availability
SVN Update (Aaron) * Move slated for 8/27 - 8/28 (Aaron will announce). * Move trunk to 'oldtrunk' branch * Move RELEASE branch to new trunk * Will follow svn usage rules from now on

20090817Video

Agenda/Minutes

Attendees: Marcos, Brian, Aaron, Jason, Joe, Jeff, Priscilla
Team Updates: * Marcos: Researching fault analysis. * Brian: Testing new releases. Found a few minor things. Strange things on ESnet measurements. hLS registration bug. Wants more testing before final release. * Aaron: Working on the toolkit - adding in dojo for web-gui to add cross-browser compatibility and tooltips. * Jason: Release activities - bug tracking. Late next week. * Joe: Gave feedback on an early RC and nothing since. Joe will send a request for psps.perfsonar.net and mdm.perfsonar.net. * Jeff: Release/bug activities. Mostly working on proposals. * Priscilla: Joining to hear how things are going.
Release Activities * pS-PerformanceTK
- Getting good feedback, lots of bugs to followup on.
- Will likely have at least two more RCs given the number of bugs. (Next one late next week.)
Testing * Brian will work with Marcos, Martin and perhaps Joe on scalability testing of the gLS/hLS.
LHC/Atlas deployments (status update) * Giving good feedback. * GridFTP archive service from SLAC is on the horizon.
Future topics: * Service Monitoring and Config management
- gLS 'monitoring' * Status collection for low-level device PMs * The DCN status page that was put together for the monitoring is a relatively simple collector that grabs various performance counters and other data about elements on the Cienas. Right now, it confines itself to things that look like interfaces. There's a similar project that Jon Dugan has been working on for ESnet called ESxSNMP that grabs performance counters for SNMP hosts. It would be good to have a discussion about how these two monitoring infrastructures might work together. At minimum, it might be feasible to get the TL1 monitoring components to use the same metadata database and counter databases, and, implicitly, perfsonar interface. * Another possible area of discussion include how the monitoring above can be extended to support collection and storage of topological information. The topology information could include how interfaces are connected together internally in the switch/router (via cross-connects, adaptation or whatever). It might also be good to think about how to collect information about how switches/routers are wired together. * It'd be feasible to grab and store TL1/SNMP alarms as part of the collection framework. Though, that might be more reasonably done via nagios plugins.
Next VC will be 8/24 2:30 EDT.

ACTIONS

Joe will send request for implementation specific web sites in perfsonar.net space.
Joe will send email to EU participants to find out who will be at DICE meeting.
Brian will work with Marcos, Martin and perhaps Joe on scalability testing of the hLS/gLS.

20090803Video

Agenda/Minutes

Attendees: Rich, Maxim, Jason, Aaron, Jeff, Marcos
Team Updates: * Jason: Working on proposals. First round of testers - lots of bug reports. RC2 released this morning. Major fixes for networking scripts and backing-store fixes. Added link to pingER GUI, added fix for pS-B. * Aaron: Working on propoosals. Working on bugs as they come up. * Maxim: Finished admin-gui for pingER. Unfortunately not unified like aaron's version. Feature set is different. It would be nice to have both. Would like to discuss GUI architecture going forward, and how full functionality can be implemented for all services. Also testing new toolkit on KOI boxes. * Rich: Email from Bigrow - has RC2 in. * Jeff: Working on proposals and toolkit bugs. * Marcos: Working on statistical anomoly detection that he hopes can grow into an eventual pS service.
Review Actions and Issues * Jason - ~~will update release pages as needed~~ * Jason - ~~will contact Marcos about IP Summarization~~ * Jason/Martin/Jeff - Finish editing OGF documents * Jason/Martin/Jeff - Complete/Upload NML document * Aaron/Maxim - Review admin GUIs and unify as much as possible. * Joe - Suggest interim steps to improve LS user experience, that don't require new software deployments. * All - Review release pages
Toolkit 'branding' * Maxim is unhappy about any organizational branding. Current 'plan' is in issue 205 (on Google Code). (Jeff - my intent is to slightly brand the toolkit as Internet2 just as the NPToolkit has always been. Likewise, individual tools on the toolkit have had some organizational branding: NPAD from PSC, pingER from SLAC/FNAL. If there is serious concern about this I'm happy to discuss it with anyone and I'm willing to concider changes. But please consider the fact that the only way some of the organizations can stay involved in developing these tools is if they can publicly get some credit for having done something for the communities they serve. It is extremely important to me that no groups are slighted here - so please review and speak up when public announcements are sent out for review. Thanks.)
Release Activities * pS-PerformanceTK
- RC2 out this morning
- Testing issues - anything huge we need to worry about? Or normal churn? * Update Your Bug Status!
LHC/Atlas deployments (status update)
Future topics: * Service Monitoring and Config management
- gLS 'monitoring' * Status collection for low-level device PMs * The DCN status page that was put together for the monitoring is a relatively simple collector that grabs various performance counters and other data about elements on the Cienas. Right now, it confines itself to things that look like interfaces. There's a similar project that Jon Dugan has been working on for ESnet called ESxSNMP that grabs performance counters for SNMP hosts. It would be good to have a discussion about how these two monitoring infrastructures might work together. At minimum, it might be feasible to get the TL1 monitoring components to use the same metadata database and counter databases, and, implicitly, perfsonar interface. * Another possible area of discussion include how the monitoring above can be extended to support collection and storage of topological information. The topology information could include how interfaces are connected together internally in the switch/router (via cross-connects, adaptation or whatever). It might also be good to think about how to collect information about how switches/routers are wired together. * It'd be feasible to grab and store TL1/SNMP alarms as part of the collection framework. Though, that might be more reasonably done via nagios plugins.
Next VC will be 8/17 2:30 EDT.

20090720Meet

Where

Summer Joint Techs 2009, Indianapolis

When

July 20, 2009 - Dinner

Attendees

Agenda/Minutes

Status collection for low-level device PMs
The DCN status page that was put together for the monitoring is a relatively simple collector that grabs various performance counters and other data about elements on the Cienas. Right now, it confines itself to things that look like interfaces. There's a similar project that Jon Dugan has been working on for ESnet called ESxSNMP that grabs performance counters for SNMP hosts. It would be good to have a discussion about how these two monitoring infrastructures might work together. At minimum, it might be feasible to get the TL1 monitoring components to use the same metadata database and counter databases, and, implicitly, perfsonar interface.
Another possible area of discussion include how the monitoring above can be extended to support collection and storage of topological information. The topology information could include how interfaces are connected together internally in the switch/router (via cross-connects, adaptation or whatever). It might also be good to think about how to collect information about how switches/routers are wired together.
It'd be feasible to grab and store TL1/SNMP alarms as part of the collection framework. Though, that might be more reasonably done via nagios plugins.

20090713Video

Agenda/Minutes

Attendees:
Team Updates * Joe: Have not looked at LS issues yet. East coast servers should be running in the next week. Working on a perfSONAR deployment white paper and jt techs talks.
- Jt Techs: Deployment experiences track session and perfSONAR deployment BOF. * Aaron: Working on DCN weather-map/circuit monitoring page. Functionally complete for this release. Also working on pS-PerfTK - specifically have fixed an issue where it would not boot on the Dell 1750. Focus going forward is making DCN monitoring pretty and working on talks.
- Jt Techs: Circuit Monitoring demonstration in perfWG, scribe DCN-WG. * Jason: Release pain continues. pS-B not yet done, new iso's out thursday.
- Jt Techs: Supporting role. * Rich: Bigrow is looking at the pre-release.
- Jt Techs: Mini-Perf workshop, JET, pS deployment BOF * Jeff: Working on pS-B/owamp. Specifically, the signal handling in perl completely changed in version 5.8, and the master/collector process loops were crafted with the old model in mind (and now go crazy sometimes). Also preparing for Jt Techs track session on the pS-Performance Toolkit.
- Jt Techs: PerfWG (staff update), JET, pS-PerfTk track-session, DCN-WG
Review Actions and Issues * Jason - ~~will update release pages as needed~~ * Jason - ~~will contact Marcos about IP Summarization~~ * Jason/Martin/Jeff - Finish editing OGF documents * Jason/Martin/Jeff - Complete/Upload NML document * Aaron/Maxim - Review admin GUIs and unify as much as possible. * Joe - Suggest interim steps to improve LS user experience, that don't require new software deployments. * Jeff - Start collecting agenda topics for a Jt Techs meeting. (First up is LS convergence time assumptions, goals.) * All - Review release pages
Release Activities * New Dates
- pSPS 3.1 Release (7/10)
  - pSB still in progress
  - RCs for other services if necessary
    - LS fixes
- pS-PerformanceTK (7/15)
  - Consider this a beta, will go out to testers Weds/Thurs/Fri * Update Your Bug Status!
Discuss agenda for F2F in Indianapolis * If more topics are not sent in, we will likely just touch base but not do anything as coordinated as last time. Still, if there are topics you would like to discuss - please send to Jeff.
LHC/Atlas deployments (status update)
Next Meeting will be Face to Face in Indianapolis.

20090706Video

Agenda/Minutes

Attendees: Aaron, Joe, Maxim, Marcos, Rich, Jeff, Jason, Martin
Team Updates * Maxim: Investigating administrative interfaces for pingER. No more bug fixes. No integration with Aaron's administration GUIs yet. Catalyst supports many auth plugins and lots of other goodness. Maxim thinks we should do catalyst. * Joe: weathermap.es.net prototype. Topology is pS driven, will do utilization soon. SNMP-MA will likely use xml-db for store in the future to deal with the dynamic changes to topology. Aaron is looking at the same issue. Joe will be developing a risk analysis of deploying a pS-PerfTK host. * Marcos: Test sets started for the IPTree implementation. * Martin: NML schema update before Thursday NML call. * Rich: Just waiting for the new disks. * Aaron: Working on monitoring of the Cienas. Attempting to have that by jt techs. Hoping to leverage ESnet weathermap. * Jason: Working on submission for NIST proposal. Closed out bugs on GUI issues. Temporary version of new pS-B GUIs. Fixing LS bugs. Bug fix on rpm packages. pS-B development is finishing up. Call with Marcos/Martin on summarization work. Need more test cases for IP summarization. * Jeff: NIST proposal, pS-PerfTK splash screen, publicity, use cases. Working on other proposals as well.
Review Actions and Issues * Jason - ~~will update release pages as needed~~ * Jason - ~~will contact Marcos about IP Summarization~~ * Jason/Martin/Jeff - Finish editing OGF documents * Jason/Martin/Jeff - Complete/Upload NML document * Aaron/Maxim - Review admin GUIs and unify as much as possible. * Joe - Suggest interim steps to improve LS user experience, that don't require new software deployments. * Jeff - Start collecting agenda topics for a Jt Techs meeting. (First up is LS convergence time assumptions, goals.) * All - Review release pages
New email lists (ps-perfsonar-users, ps-perfsonar-announce, ps-perfsonar-dev)
Release Activities * New Dates
- pSPS 3.1 Release (7/10)
  - pSB still in progress
  - RCs for other services if necessary
    - LS fixes
- pS-PerformanceTK (7/15)
  - Consider this a beta, will go out to testers Weds/Thurs/Fri * Update Your Bug Status!
LHC/Atlas deployments (status update)
Next VC will be 7/13.

20090622Video

Agenda/Minutes

Attendees: Jason, Brian, Maxim, Marcos, Joe, Rich, Aaron, Jeff
Team Updates * Jason: NIST proposal. Hope to get bugs done over the next week. Will update release plan in the next week. * Aaron: IDC developers meeting. Document clean up pS-PerfTK. Throughput tests seem to be more stable. Testing hardware for good owamp hosts. Will be focusing on circuit monitoring going forward. * Rich: Nothing new. All of Atlas is up except for SLAC. * Jeff: Focus on pS-B owamp from here on out. * Brian: Bug fixes in test harness. Tested nptoolkit. * Joe: Working on throughput-google-maps GUIs. Having problems interacting with gLS infrastructure. Working on presentations - Jt Techs. Also working on recommendations for ESnet sites for deploying pS. * Marcos: Preparing paper on IPsummarization. Perl is done, but not integrated into LS yet. * Maxim: Admin interface of pingER is now in catalyst. AJAX interaction is a bit strange. Anticipate finishing this week.
Review Actions and Issues * Jason - ~~will update release pages as needed~~ * Jason - will contact Marcos about IP Summarization * Jason/Martin/Jeff - Finish editing OGF documents * Jason/Martin/Jeff - Complete/Upload NML document * All - Review release pages * Aaron - Update CGIs based on suggestions today. * Aaron - Consider the switch to catalyst - weigh positives and negatives * All - Review Aaron's CGIs, provide more feedback
Release Activities 1. Development Update
- pSB
- Admin GUIs 1. Open Bugs
LHC/Atlas deployments (status update)
Next VC will be 6/29.

ACTIONS

Jason - ~~will update release pages as needed~~
Jason - will contact Marcos about IP Summarization
Aaron/Maxim - Review admin GUIs and unify as much as possible.
Joe - Suggest interim steps to improve LS user experience, that don't require new software deployments.
Jeff - Start collecting agenda topics for a Jt Techs meeting. (First up is LS convergence time assumptions, goals.)

20090608Video

Agenda/Minutes

Attendees: * Jason * Rich * Maxim * Marcos * Brian * Aaron
Team Updates * Internet2
- Aaron - Most time spend on NPToolkit work, particularly the configuration GUIs. See section later in meeting for more info. Other bugs related to toolkit services were found/fixed/pending fixing as well.
- Jason - NPW last week at Vanderbilt. Lots of interest from Vandy (both REDDnet and OSG participation) in performance tools and perfSONAR. OGF documents entering a slowdown after some more tweaking (see update later). Release activities starting. See pages for more information.
- Jeff - At TERENA this week, was at SCinet meeting last. Trying to finish pSB.
- Rich - NPW @ Vandy last week. * ESnet
- Joe - Unable to make call today. Brian gave minor update that he has been involved in many sysadmin activities.
- Brian - Has summer students working on visualization. Jason provided with old WMap code. 2 students for WMap development, maybe one more for SNMP visualization. Asks if we are aware of visualisation activities for geni (GRNOC is involved). Jason will ask Matt Z/Jeff (N.B. Jason did ask and both are aware). * Fermi
- Maxim - PingER GUI completed, used catalyst and liked it. Would like to see other GUIs use it as well to keep look/feel the same and to speed up development. Fixing bugs (e.g. 181) and will give RPMs to Jason for uploading.
  - Minor conversation between Maxim/Brian on the speed of PingER GUI. Brian states it is slow for large chunks of data, is there some sort of status spinner to indicate progress? Maxim says yes, Jason/Aaron confirm. Maxim explains how it works in relation to MA and that slowness may be MA related but also tweaked GUI parameters to need less data. * UDel
- Marcos - IPSummarization work is progressing, converted to perl and awaiting Jason/Martin/Jeff to start some form of evaluation. Jason will get the ball rolling this week. Next project is anomaly detection.
Review Actions and Issues * No actions posted from last week * Jason has uploaded new RC (not posted but was a holdover) * OGF docs are ready for Review * Add more if anyone thinks of any...
KOI performance update - Aaron * Aaron reports nothing new beyond Brian's observation that new kernel/driver software helped. Seems to be problems w/ perfSONAR-BUOY now (database related). Jason thinks it may be related to the new package structure in SVN: perfSONAR-BUOY was never converted so may not work correctly. Jason/Aaron will look into this. Jason recommends sticking with trunk version of pSB.
Aaron's GUIs * Aaron presented some GUIs:
- Scheduled Tests - https://lab246.internet2.edu/admin/regular_testing.cgi
- BWCTL - https://lab246.internet2.edu/admin/bwctl.cgi
- OWAMP - https://lab246.internet2.edu/admin/owamp.cgi
- Running Services - https://lab246.internet2.edu/admin/services.cgi
- NTP - https://lab246.internet2.edu/admin/ntp.cgi
- Administrative Info - https://lab246.internet2.edu/admin/administrative_info.cgi * Maxim: See comments from email list regarding GUI vs. Web GUI. * Aaron: Explains that the dojo stuff he is using does what maxim wants. CGI::AJAX may make something not look as Maxim would like to see it. * Maxim/Aaron: Discussion regarding look feel + architecture. Maxim thinks we should use catalyst as the framework (at a minimum we should all use the same thing - catalyst is easy to use and fully featured). * Aaron: discusses why he choose dojo (IDC folks use it, also fully featured, CSS etc. built right in). * Jason: Can discussion architecture details later - should focus on layout/content for this discussion * Jason has some comments on layout/structure/content:
- Finds some of the longer configuration screens confusing. E.g. if you have to scroll to see all content, there may be too much content on the page. Suggests a Wizard interface that gives a small number of choices per screen - less confusion as to what to do (example: regular setup has too many options near the bottom, need to know what each 'click' is going to do). Brian remarks that a wizard will frustrate power users. Two interfaces is hard to maintain, so need to weigh the positives and negatives of each.
- Some of the choices for checkboxes and similar have a mouseover hover property. These look like hyperlinks but don't click. Could we make them a pop-up menu with more information?
- Names of things are not conistent (e.g. throughput testing vs BWCTL testing vs perfSONAR-BUOY testing). Aaron will suggest alternatives. * Brian's last suggestion is for more advanced options (e.g. TCP cong. alg)
Release Activities - Jason 1. perfSONAR-PS 3.1 status
- New Bugs - these have been identified and are working to be closed.
- RC Generation - still generating on demand. 1. perfSONAR-PS 3.1+ status
- See pSPS200907Release
- Tasks Assigned - read the web page to see if you are on the hook for anything. Note that some tasks (testing/doc/code reviews) may be stalled on development. Developers should communicate when they are ready for these subsequent steps.
- Bugs Updated - keep this list up to date: cross off things when you know they are done. 1. pS-NPToolkit 3.1 status
- See PerformanceToolkit200907Release
- Tasks Assigned - read the web page to see if you are on the hook for anything. Note that some tasks (testing/doc/code reviews) may be stalled on development. Developers should communicate when they are ready for these subsequent steps.
- Bugs Updated - keep this list up to date: cross off things when you know they are done.
OGF Document status - Jason * NM-WG
- Base Uploaded to GF - https://forge.gridforum.org/sf/go/doc15649?nav=1
- Outstanding Edits - Jeff/Martin/Brian
- New version to upload by end of this week * NML-WG
- Nothing uploaded (yet)
- Outstanding Edits - Jeff/Martin
- Expected upload in 1/2 weeks * NMC-WG
- Base Uploaded to GF - https://forge.gridforum.org/sf/go/doc15644?nav=1
- MA Uploaded to GF - https://forge.gridforum.org/sf/go/doc15646?nav=1
- MA still incomplete
- Changes/upload in 1/2 weeks
LHC/Atlas deployments (status update) - Rich * Nothing to add
Next VC will be 6/22. * Will be short and focused on release due to DCN developers conference.

ACTIONS

Jason - will update release pages as needed
Jason - will contact Marcos about IP Summarization
Jason/Martin/Jeff - Finish editing OGF documents
Jason/Martin/Jeff - Complete/Upload NML document
All - Review release pages
Aaron - Update CGIs based on suggestions today.
Aaron - Consider the switch to catalyst - weigh positives and negatives
All - Review Aaron's CGIs, provide more feedback

20090601Video

Agenda/Minutes

Attendees: Joe, Brian, Evangelis, Marcos, Jeff, Maxim
Team Updates * Brian: Nothing new, planning to work with some students to work on a weathermap application. Will contact Jason for latest weathermap developments. Also have a new student working on iperf. Anyone interested in specific features should add things to the iperf tracker. * Joe: Have 2 topology servers for ESnet installed now. Updated automatically - data should be current within 2 hours. Unsure how compliant it is. Joe would like to discuss svn externals. * Evangelis: Showed google-earth visualizations at OGF. * Marcos: Working on Perl version of IPTree. * Maxim: Fixing statistical bug in pingER. Have new RPM's. Explored catalyst for GUIs. Can handle large loads. Moved pingER data UI under catalyst under trunk. Will start a wiki page to start evaluation of moving pS-PS GUIs into the same framework. * Jeff: OGF last week. NM-WG is progressing as expected. NMC is on track, but more input is needed from GN3 participants. NML was contentious as expected, but progress was made. Specifically, lifetime concept was agreed to. And, the 'relation' element in UNIS makes the UNIS model much closer to NDL. Watch for notes/docuements.
Review Actions and Issues * http://perfsonar-ps.googlecode.com/svn/wiki/20090518Video/20090601perfsonar-ps-issues.xls
- Brian: will look at the data since it has been running a while now.
- Joe: We should be tracking progress of issues week by week.
Next VC will be 6/8.

20090518Video

Agenda/Minutes

Attendees: Brian, Jason, Marcos, Aaron, Jeff, Martin, Fahad
Team Updates * Brian: KOI network drivers. Solid if within around 15 ms. * Jason: OGF docs, looking for reviews this week. Wants to keep this going. * Marcos: Summarization looks reasonable, will work on perl version. Joe agreed this version looked reasonable. * Aaron: Working on next pS-NPToolkit release. Updating configuration GUIs. Working on an interaction with 'keywords'/communities to allow better community related things across multiple pages. Discussion on how to optimize the HTTP connects. * Martin: Looking at topology, and focusing on OGF. * Jeff: OGF/ pS-B-owamp. * Joe: no update. * Fahad: Working on integrating gridFTP with pS.
Review Actions and Issues * http://perfsonar-ps.googlecode.com/svn/wiki/20090518Video/20090518perfsonar-ps-issues.xls
KOI performance update
perfSONAR-PS v3.1 Release Update * New Bugs * Release Candidate Generation
- LS, SNMP-MA, LS-reg, status to be updated this week (by friday).
- pingER may need one for docs and hLS registration issue. * Suggestions on Testing
- Jason will be trying to get Penn st/ TAMU to test.
- Fahad will talk with Les/Yee about testing the latest pS-pingER RC.
pS-NPToolkit 3.1 status * Updated the roadmap with the atlas requests: http://code.google.com/p/perfsonar-ps/wiki/NPToolkitPlans * Plans to update the roadmap with feature status.
OGF Document status (Jason) * Discussed reviewers.
Circuit Monitoring Development Status
SNMP-MA ESnet new database integration update * Still waiting on release branching issues
LHC/Atlas deployments (status update) * MIT/FL/UNL all pending. * Jeff attended the Atlas call last week and updated that community on the 'roadmap'. They seemed ok with the schedule as communicated. (Next pS-NPToolkit in late June.)
Next VC will be 6/1.

20090511Video

Agenda/Minutes

Attendees: Brian, Xi, Aaron, Maxim, Rich, Martin, Joe, Jason, Jeff, Marcos, Evangalis
Team Updates 1. Internet2
- Aaron: pS-NPToolkit (web admin update)
- Jason: OGF26 docs.
- Jeff: AA, OWAMP, planning.
- Rich: nothing 1. ESnet
- Brian: SNMP-MA back up - registering gLS. Looking at traceroute visualizer.
- Joe: Working on hardening deployments (monitoring etc...). 1. UDel
- Marcos: Updates to summarization. Updates on the wiki.
- Martin: Preparing for OGF. 1. FNAL
- Maxim: Working on catalyst/modperl/tt. (KOI perf problems still exist.)
KOI performance * network issues still exist. Aaron will make sure new driver is on a test release of the ISO. Brian will test the hardware more.
DOE funding issues * FNAL GUI was funded (Bucknell? Statistics on pS.) * ESnet stimulus $$$ for pS * Total of 2-3 more FTEs.
perfSONAR-PS v3.1 Release Update * New Bugs
- Missing extension script to the configure daemon, new RC will correct. * Release Candidate Generation
- LS, SNMP-MA, LS-reg, status to be updated this week (by friday).
- pingER may need one for docs and hLS registration issue. * Suggestions on Testing
- Jason will be trying to get Penn st/ TAMU to test.
Review Actions from last call: April 29
Review open issue tracker items * Jeff will dump the table before calls and bring out specific ones for discussing at future calls.
Discuss priorities for next pS-NPToolkit Release (see Shawn email for ref) * Brian/Rich like Shawns list, Jeff is not crazy about the GUI. * Maxim would like an option to 'reset' any service. * Brian (self monitoring?) * Aaron reviewed the list from Shawn, and will attempt to address all but the GUI question.
OGF Documents (Jason) * 3 main thrusts: Base, Topo, NMC * Brian, Jon, Joe will be asked to review pieces.
Circuit Monitoring * Aaron/Andy working on overlap issues. * Joe asked about an overall architecture document.
SNMP-MA ESnet new database integration update * Waiting on release branching issues
LHC/Atlas deployments (status update) * MIT/FL/UNL all pending. * Rich needs good estimate for next pS-NPToolkit Release. Jeff suggested Jt Techs as the absolute limit, but hopes late June.
Next VC will be 5/18.

20090501IDCAgenda

IDC Agenda

8:30 (0:30) NSI Summary (John) * http://perfsonar-ps.googlecode.com/svn/wiki/20090429Meet/NSI.IDC.05012008.ppt * Coordinate IDC planning with NSI architecture definition
9:00 (0:15) Circuit Status (E2EMon lives on...): Current Arch (Aaron) * http://perfsonar-ps.googlecode.com/svn/wiki/20090429Meet/e2emon_howto.ppt * Architecture: GUI/MA/agent * Configuration requirements * Limitations: CurrentE2EMonLimitations
9:15 (2:00) Circuit Status: Planned Architecture Enhancements (Aaron) * Use-case clarification collection
- What needs to be 'monitored'
- Look at chrobb/IU use-case * Circuit Monitoring Roadmap * pS DCN Monitoring Architecture
- Picture of the pieces
- Proposal for interfaces
- Topology examples
- Evaluation
  - How does 0.5 fit this model? 0.6?
  - Where does the hardware specific content get added?
11:15 (0:45) pS/IDC Overlap Issues (Aaron) * 0.6 module decomposition brings out overlap in
- AAA
- IS (Topology and Service Discovery)
12:00 Lunch
12:30 (4:00) 0.6 Design Decisions (Chin) * Technologies to support modules' WS interface (e.g. Tomcat, Jetty) * Distinction between modules' exposed interface (e.g. to Coordinator) and internal module * Symmetric/asymmetric behavior for module functions

20090429PSAgenda

pS Agenda

Wednesday

2:00 (0:30) Release Process (Jason) * Release Process
- Overall comments
  - Maxim: More testing needed, shorter release cycle.
  - - Distribute testing among more people.
  - Martin: DAMSL students. action:Martin - Build-Bot
  - Jason: Need to automate
  - Jon: Unit tests?
  - - Suggests requiring the unit test before fixing a bug.
  - Aaron: Unit tests are difficult due to the current software arch.
  - Decision: Q/A contact will not close issues until determing if a test should be developed before closing an issue.
  - Decision: Add standing agenda topic to review issues/prios in issue tracker.
- Potential changes * 3.1 Release
- RC Status
- Testing Status
- Next RC?
- Post-mortum planned for VC in May. * Future Releases
- 3.1+
- 3.2
2:30 (1:00) Software Architecture (Jeff) * One goal is to unify more of the code base. * Another goal is to make it easy for new developers to create a service to provide a new kind of performance data. * Common config file interfaces * Common daemon (mod-perl/Apache)
- else needs to be common - please add into agenda here
- Action: Maxim will explore making Status Service available via mod-perl.
- Clear need to refactor.
  - Need to ensure functionality is not lost.
  - Action: Aaron will take the metadatakey request handling in status-ma and snmp-ma and write up something to describe how it could be implemented by common modules.
3:30 (1:00) pS-Performance node (Aaron) * pS-NPToolkit roadmap
- Change of defaults (iperf buffer/duration)
- What functions must be in the next release
- Fedora transition * software.internet2.edu yum repo
- Mirrors? (Regionals, ESnet?) * pS-Performance Grid
- How do we make configuring a 'group' of pS Perf Nodes easier?
- Action: Martin will take this.

Thursday

8:30 (1:00) Project Roadmap (Jeff) * Political landscape (IRNC, GENI, CTP, Regionals, Labs, LHC)
1:15 (0:30) Auth use-case discussion (Andy)
9:30 (1:00) pS protocols (Martin) * Strategy for publishing standards (IRNC etc...)
- Start with de-facto (as-builts)
- Engage OGF process
1:45 (0:30) MA protocol enhancements (Brian) * Addition of 'limit' option to MA/MP queries (on by default) * Addition of metadata queries to MA (i.e.: what interfaces have new data in past T?) * Addition of TTL negotiation to LS registration
10:30 (0:30) gLS: Current System Arch (Jason) * Slides
- http://perfsonar-ps.googlecode.com/svn/wiki/20090429Meet/dev_ls.ppt
- http://perfsonar-ps.googlecode.com/svn/wiki/20090429Meet/dev_gls.ppt
- http://perfsonar-ps.googlecode.com/svn/wiki/20090429Meet/dev_issues.ppt * hLS mode * gLS overlay/root servers * Summary of current Summarization * Distribution - discovery phase * Query * Timing issues
- Synchronization of root gLSs
- Refresh of registrations
- Refresh of summaries * API
- Bootstrapping * Percieved limitations
- Stability (XMLDB dies...)
- Performance (XMLDB/Summarization)
- IPTree (v4 only)
- Inflexible discovery for new data/event types * hLS/MetadataQuery Action Items:
1. Specific Use Cases for Metadata Queries
2. Metadata Query Language Proposal
3. Architecture Proposal for where different queries are implemented
4. Summarization 'hints'
- An MA might be able to say what the IP summarization should get...
1. Consider time dim to metadata Martin
2. Add netlogger into hLS/gLS
- time - summarization issue
- time - request/responses
1. perf comparison for hLS/gLS on ramdisk. Brian/Joe
11:00 (1:00) gLS: Transition to IS (Martin/Marcos) * Summarization notes
- http://perfsonar-ps.googlecode.com/svn/wiki/20090429Meet/summ.pdf * Address current limitations as much as possible * Topology distribution within summarization * Authoritative data * New IPTree summaries * Action:
- How do we do time dimension?
  - Add timestamps to topo-elements.
  - Need to think about queries/discovery/storage for this.
- Question on how to summarize L2/L1 for gLS distribution. * Use cases for topology service. Evangelis * Action:
- Basic element relationships in combined IS. Martin
12:00 (0:30) Lunch
12:30 (0:15) gLS: Test Plan (Brian) * Scalability testing (See: http://code.google.com/p/perfsonar-ps/wiki/TestPlan#Service_Specific) * Resolution of overlapping summaries Martin/Joe - deathcar
12:45 (0:30) gLS: Roadmap (Jeff) * Specific enhancements for 3.2
2:15 (1:00) Configuration Management (Joe) * Problem statement
- use-cases * Current range of solutions * List of effected software packages
3:15 (1:00) Meta-scheduler (???) * Single node scheduling among tools * Shared link/resource scheduling
4:15 (0:30) GUI Development (???) * What APIs should we be providing?
- Current messages good enough?
X:XX (1:00) Project Roadmap (Jeff) * Project scalability * Community Outreach Milestones (including demos) * Documentation Milestones
- Web-site cleanup/orgainization
- Availability of protocols/schemas * Software Milestones

20090429Meet

RSVP

It is critical that you send a note to Jason (zurawski at internet2 dot edu) if you are planning on attending. Please indicate if this will be in person (for food purposes) or video (for MCU purposes). Please let us know as soon as possible.

Where

Post Spring Member Meeting 2009, Washington, DC

1150 18th Street, NW
Suite 1020
Washington, DC 20036
Phone: (202) 872-9119
Fax: (202) 872-6648

When

The meetings will span 3 days:

perfSONAR - Weds April 29, 2009 - 2PM to 6PM
perfSONAR - Thurs April 30, 2009 - 8AM to 5PM
perfSONAR and IDC (Circuit Monitoring) - Friday May 1, 2009 - 8AM to 5PM

The meeting will focus on development activities within and between two groups:

perfSONAR Developers (4/29-30)

The perfSONAR developers meeting will focus on pending issues within the perfSONAR-PS framework. Topics will span the interests of IDC developers as well, and all are encouraged to arrive early to attend and participate.

pS Attendees

In Person
1. Jason Zurawski - Internet2
2. Jeff Boote - Internet2
3. Aaron Brown - Internet2
4. Andrew Lake - Internet2
5. Maxim Grigoriev - Fermilab
6. Chris Tracy - MAX
7. Jarda Flidr - MAX
8. Xi Yang - ISI
9. Tom Lehman - ISI
10. Evangelos Chaniotakis - ESnet
11. Joe Meztger - ESnet
12. Jon Dugan - ESnet
13. Martin Swany - UDel
14. Ezra Kissel - UDel
15. Marcos Portnoi - UDel
16. Guilherme Fernandes - UDel
17. John Vollbrecht - Internet2
Remote
1. Brian Tierney - ESnet
2. TBD - ESnet
3. Chris Robb - Internet2

pS Agenda and Notes

See the Agenda here: 20090429PSAgenda.

pS/IDC Developers (5/1)

This meeting will focus primarily on the topics that related to circuit monitoring. This will touch on global identifiers, topology, technology/layer abstraction and monitoring architecture. We will have phone conferences in the weeks leading up to this to articulate use cases and general architecture. The plan is for the meeting to be used to fully describe that architecture to everyone involved, and break it down into action items for development.

If IDC developers like, a portion of the day can be set aside for other current topics.

pS/IDC Attendees

In Person
1. Jason Zurawski - Internet2
2. Jeff Boote - Internet2
3. Aaron Brown - Internet2
4. Andrew Lake - Internet2
5. Maxim Grigoriev - Fermilab
6. Tom Lehman - ISI
7. Xi Yang - ISI
8. Chris Tracy - MAX
9. Jarda Flidr - MAX
10. Evangelos Chaniotakis - ESnet
11. Joe Meztger - ESnet
12. Jon Dugan - ESnet
13. Chin Guok - ESnet
14. Martin Swany - UDel
15. Ezra Kissel - UDel
16. Marcos Portnoi - UDel
17. Guilherme Fernandes - UDel
18. John Vollbrecht
Remote
1. Brian Tierney - ESnet
2. David Robertson - LBL
3. Chris Robb - Internet2
4. TBD - ESnet

IDC Agenda and Notes

See the Agenda here: 20090501IDCAgenda.

Local Information

Some compiled local information from a semi-local. Will add more over time, requests for specific information would be great.

DC Office

Located in the Golden Triangle neighborhood, borders Dupont/Scott Circle and George Washington University. Not known for great food or restaurants (walk somewhere north or east for the good stuff). Office Address:

1150 18th Street, NW
Suite 1020
Washington, DC 20036

This is between L and M streets on 18th (left side of street when facing North, white building between 2 parking garages). The office is on the 10th floor (see the lobby attendant when entering).

From Orange/Blue Metro Line:

Get off at the Farragut West stop. There are two exits, please exit on 18th Street. Walk North along 18th street (Straight as you exit the station) and cross over K and L Streets. N.B. There is construction on the West (left) side of 18th Street between K and L. 1150 is midway between L and M on the left hand side, it is a white building (there is a construction fence outside of our entrance).

From Red Metro Line:

Get off at the Farragut North stop. There are three exits, please exit on one of the L Street and Conn. Ave exits. Walk West along L street (towards oncoming traffic) and cross over 18th Streets. Make a right (North) onto 18th Street, 1150 is midway between L and M on the left hand side, it is a white building (there is a construction fence outside of our entrance).

Remote Participation

Video access is available, we will be using the ESnet conference bridge:

GDS: 00113498255555
Phone: 1-510-883-7860 then 8255555# at the prompt.

Transportation (Airports)

Reagan National Airport

DCA is located along the Potomac river in Arlington VA. Due to flight restrictions there are limited cross country flights each day, so finding a variety of arrival departure times is sometimes challenging. DCA is directly connected to the Washington DC metro system via the Yellow and Blue lines making it easy to get to Crystal City (for the SMM) and into the city (where the Internet2 Office is located). Allow 20 minutes to travel on the Metro between the DC office and the airport. Security lines may be long during peak periods (Afternoons) - allow at least 45 minutes for security.

Dulles International Airport

IAD is located in Chantilly VA, which is fairly far out (and very close to were Jason lives). There is a wider variety of airlines and schedules available here. Unfortunately there is limited access to transportation. The Metro does not extend to IAD (yet) so using the public bus (Route 5A, drops off at Rosslyn station on the Orange and Blue lines) or a private coach service (Washington Flyer picks up/drops off at the West Falls Church station on the Orange Line) would be necessary. It may take upwards of an hour to get from the airport to the DC office. Security lines may be very long in the mornings and evenings - allow 45 minutes to an hour for security and 10 minutes for transportation to gates (shuttle buses).

Transportation (Ground)

Using the Metro is by far the easiest way to get around. I would suggest buying a SmartTrip electronic card (comes with $10 loaded on it) and using this for your stay. The card can be loaded with more value as you need it in any station. This is preferred to attempting to purchase paper tickets for each ride, or trying to figure out rules regarding entry vs exit fare, etc.

Shuttle services such as Washington Flyer and Super Shuttle can get you to the airports.

Taxis are also common, note some are un-licensed and no longer follow the zone fare rules.

Suggested Hotel

We recommend staying at the Crystal Gateway Marriott when attending these meetings. (Everyone attending the SMM will already be there, so this is where most people will already be staying.) This hotel offers a competitive rate compared to hotels in the DC business district. The information page is here. The hotel address is:

Crystal Gateway Marriott
1700 Jefferson Davis Highway
Arlington, Virginia 22202

From the hotel we will take the Blue Line in the morning to get to the office. Note it may take about a half hour and trains will be crowded in the morning.

Alternative Hotels

There are several other hotels available within walking distance of the DC office, prices will vary (check a travel site for best deals - note that the listed prices were captured in late February and will adjust with the local market). April is the start of tourist season, so prices may begin to be expensive. Most hotels offer government rates.

The Quincy Hotel - $169 (on the Quincy Web Site)

Phone: 202-223-4320 or 1-800-424-2970
Fax: 202-293-4977
1823 L Street NW (around the corner from the office - between 18th and 19th Streets)
Washington, DC 20036

Holiday Inn Central - $189 (on the Holiday Inn Web Site)

Phone: 202-483-2000 or 1-800-248-0016
Fax: 202-797-1078
1501 Rhode Island Avenue NW (off of Scott Circle)
Washington, DC 20036

Crowne Plaza Hamilton - $144

Phone: 202-682-0111
Fax: 202-682-9525 
14th and K Street, NW, 
Washington, DC 20005 USA

Four Points by Sheraton - $148

Phone: 202-289-7600
201 K Street N.W. 
Washington, DC 20005

Renaissance on M Street - $159

Phone: 202-775-0800
Fax: 202-331-9491
1143 New Hampshire Avenue NW
Washington, District Of Columbia 20037 USA

Westin Hotel - $179

Phone: 202-429-1700
1400 M Street NW
Washington, DC 20005

Restaurants

Caribou Coffee‎

1800 M St (Corner of 18th and M - close to office)

Starbucks

1734 L St NW (On L between 18th and 17th - close to the office)

McCormick and Schmick - Seafood

Phone: 202-861-2233
1625 K Street NW ( K Street between 16th and 17th street)

Daily Grill - American

Phone - 202-822-5282
1200 18th Street NW (M Street and 18th street)

Georgia Brown - Low Country Southern

Phone - 202-393-4499
950 15th Street NW (corner of 15 th and K Street)

DC Coas - Seafood

Phone - 202-216-5988
1401 K Street (corner of 14 th and K Street)

Vidalia - Southern

Phone- 202-659-1990
1990 M Street NW (bet. 19th and 20th streets)

Penang - Malaysian

Phone - 202-822-8770
1837 M Street NW (corner of M and 19 th Street)

Spezie - Modern Italian

Phone - 202-467-0777
1736 L Street NW (between 18th Street and Connecticut Avenue )

Fin - Modern American Seafood

Phone - 202-530-4430
1200 19th Street NW (between M and N Streets)

Smith and Wollensky - Steakhouse

Phone - 202-466-1100
1112 19th Street NW (between L and M Streets)

Legal Seafood - Seafood

Phone: 202-496-1111
2020 K Street NW (between 20th and 21st Streets)

Mackeys Public House - Irish Pub

Phone 202-331-7667
1823 L Street NW

Mai Thai - Thai

Phone: 202-452-6870
1200 19th Street. N.W.

Thai Phoon Restaurant - Thai

Phone: 202-667-3505
2011 S Street, N.W. (Past Dupont Circle)

20090420Video

Agenda/Minutes

Attendees * Aaron, Maxim, Jason, Rich, Jeff, Marcos, Martin
Team Updates 1. Internet2
- Aaron: Cisco teleprescence work. pingER testing.
- Jason: Cisco teleprescence work - nagios and remote hands at hotel. Also help pingER testing.
- Rich: iperf testing is showing a lot of variations.
- Jeff: CTP, auth, architecture. 1. ESnet 1. UDel
- Martin: Seeking funds...
- Marcos: IPSummarization, Protocol docs. 1. FNAL
- Maxim: pingER testing, and RPM preparation. Work with MIT to get bwctl tests going - test results have large variations and Maxim will be working more with them. Maxim will try and reproduce a problem with zombie processes.
perfSONAR-PS v3.1 Release Update * New Bugs
- 132 - ~~PingER packaging is non standard~~
  - Fixed and closed
- 133 - ~~PingER Libs are still in Shared/ area~~
  - Fixed and closed
- 134 - ~~PingER RPM depends on Text::CSV::Simple~~
  - Fixed and closed
- 137 - ~~RRD.pm cannot read RRD File~~
  - Fixed and closed
- 138 - PingER request fail with Can't coerce GLOB to string...
  - Fixed and closed
- 139 - ~~PingER is giving off warnings~~
  - Fixed and closed * Software/Performance Web Pages will be updated this week * Release Candidate Generation
- Waiting on PingER testing. Will regenerate PingER/SNMP MA. * Suggestions on Testing
Review Actions from last call: April 06
Spring Member Meeting * Group Activities
- Perf WG
- IS-WG
- Demos
  - DCN
  - CTP
- Others? * Post Meeting Topics
- pS Agenda
- IDC Agenda
Circuit Monitoring
SNMP-MA ESnet new database integration update
LHC/Atlas deployments (status update)
Next meeting will be 4/29 in D.C.

ACTIONS

MP: Update IPSummarization based on email/conference call suggestions
JZ: Coordinate RC testing
AB: Update Circuit Examples
JB: Prepare agenda page from ideas talked about (not putting them in the meeting notes).

20090406Video

Agenda/Minutes

Attendees: * Jason * Aaron * Joe * Rich * Marcos
perfSONAR-PS v3.1 Release Update * See pSPS20090327Release for details
- Check individual progress.
  - JZ: All software to be released has been reviewed. Betas are being generated and tested internally
- Documentation Reviews
  - Developers Finish (3/27)
    - JZ: All developers have completed this. User guides are still outstanding, but will have to be a work in progress for the next couple of weeks.
  - Reviews Due (4/3)
    - JZ: Did not receive all reviews on time. AB/JZ reviewed the remaining items to make the deadline. Will still need reviewers after the user documentation is complete
- Uploading to repository
  - AB/JZ have access, Maxim is still on vacation (will ask him if he wants access to rpm/source repo).
- Testing
  - Testing Procedure
    - JZ has a list of interested parties to send testers to. Should test locally first (on as many different OS/Arch combinations as possible). Testing should focus on:
      1. Install Procedure vs instructions
      2. Install success on specific OS/Architectures
      3. Configuration vs instructions
      4. WS Test cases included in the package
      5. Other factors (firewall configuration, etc.)
Other Issues 1. Wiki-reorg
- There were no objections to cleaning the wiki 1. IPSummarization Comments
- JM/MP talked about recent comments on IS-WG mailing list.
- JM would like to see goals/metrics listed that are used to evaluate one summarization alg. vs another. Has suggestions:
  1. ratio of entries in result over entries in initial set (e.g. compression)
  2. ratio of unique IP addresses in initial set over unique IP addresses in result set (e.g. if we go from a /32 to a /31, we have increased the number we are claiming by 255)
- MP understands, and will try to describe the definition of good summarization on the wiki for conversation.
- AB cautions about hLSs advertising false ranges (over summarizing) especially when dealing with data sets that are not continuous. JZ explains that MP should be testing with domain data sets (e.g. ESnet) but then combining 2 data sets (e.g. ESnet and APAN for example) to see how the alg. behaves.
- JM closes with a statement about accuracy when collapsing IP blocks (adjacent blocks = good, non = troublesome) 1. Circuit Monitoring update
- Status of exmples (aaron)
  - AB reports no change since last time 1. SNMP-MA ESnet new database integration update
- JM is not aware of progress, but that doesn't mean there isn't any.
- JM is pretty sure Jon will make it to DC, we can discuss this topic during the meeting.
- JM wants to ensure we deal with diet constraints for the meeting. JZ has action of sending out something. 1. LHC/Atlas deployments (status update)
- RC reports that things are going about the same as before, still waiting on BNL to be fixed.
- JZ notes that REDDnet folks are happy Ultralight is in working order again.
Next meeting will be 4/13

ACTIONS

MP: Update IPSummarization based on email/conference call suggestions
JZ: Coordinate testing
JZ: Send out lunch email for F2F meeting for diet constraints
AB: Update Circuit Examples
All: Review documentation

20090330Video

Agenda/Minutes

Attendees: Aaron, Marcos, Jason, Brian, Jeff
Team Updates 1. Internet2
- Jeff: pS-B owamp database code.
- Aaron: Packaging/documentation for 3 services. Still needs minor doc on status. Working on examples.
- Jason: Finished yum repo. Finished doc/testing on services. Worked with PSU on deployment issues. 1. ESnet
- Brian: Working on pS-B deployment. Mostly done. Lots of suggestions for improved logging. BNL debugging. Many network problems. 1. UDel
- Marcos: Prototypes for IPSummarization. Examples on wiki, wants comments on what is good. (We want to encourage everyone interested in this to attend IS-WG meetings.) 1. FNAL
Review Actions from last call
perfSONAR-PS v3.1 Release Update * See pSPS20090327Release for details
- Check individual progress.
- Documentation Reviews
  - Developers Finish (3/27)
  - Reviews Due (4/3)
- Alpha Testing
- Beta Testing
RPM package question * Since we are making available RPMs of software that we don't produce and may be available elsewhere, should we tag our RPMs with a special version identifier (e.g. -1.ps or -1.internet2)?
Circuit Monitoring
SNMP-MA ESnet new database integration update
LHC/Atlas deployments (status update)
Next meeting will be 4/6

ACTIONS

Joe/Others: Please look at the IPSummarization page and comment. http://code.google.com/p/perfsonar-ps/wiki/IPSummarization.
Jeff: Porting/installation testbed. (UWisc thing)
Jason/Aaron/Maxim: Find beta testers.
Aaron: Circuit Monitoring examples.
All:
- Documentation Reviews
- Beta testing

20090323Video

Agenda/Minutes

Attendees: * Jason * Marcos * Joe * Brian * Rich * Maxim * Aaron * Martin
Circuit Monitoring * Joe/Aaron/Jeff have been discussing this issue on the mailing list. * Joe: Wants to be careful we do not try to build from scratch, should be trying to use as much of what the EU has contributed as possible. Need to worry about interoperating especially on legacy systems that would not upgrade. * Aaron: Old topology in use for E2Emon systems was not flexible (especially for what we want to do down the road with IS activities and more efficient queries). Hard to go between old and new. * Joe: Agrees there are problems, but we need to show how any new method is an evolution of the old (not starting new). We need to move forward by making better - not re-inventing. Most looks ok. Has a concern about the demarc issue (e.g. which side monitors something that is shared). * Aaron: Sent out email talking about this. Will always be some overlap to worry about, have point of view for circuits. * Joe and All would like more examples. Aaron will provide.
perfSONAR-PS v3.1 Release Update * See pSPS20090327Release for details
- Check individual progress.
  - Jason - Writing tests cases and documentation
  - Aaron - Same
  - Maxim - same
  - All three will alert when doc is ready for reviewers
  - Alpha testing this week
  - Beta testing next week
  - Maxim can wait to implement changes suggested in aarons code review
- Code reviews by the 20th.
  - See the reviews here: pSPS20090327Release#Code_Review_Reviews
  - COMPLETED
    - Maxim
    - Guilherme
    - Brian
    - Joe
    - Marcos
    - Jason
    - Aaron
  - MISSING BUT EXCUSED
    - Jeff
- Documentation Reviews
  - Developers Finish (3/27)
  - Reviews Due (4/3)
- Open Bugs are all now closed or held over
  - 19 - Aaron/Maxim - ~~PingER GUI gLS pingable hosts~~
  - 73 - Jason - ~~Client::LS Registration Logic Re-arrange~~
    - Addressed, but held over till 3.1+
  - 79 - Jason - ~~Empty gLS/hLS Messages~~
  - 84 - Jason - ~~Exit MA service if no data is present~~
  - 85 - Jason - ~~add verbose logging on errors for ma/ls~~
  - 105 - Jason - ~~hLS Summarization of DCN Data~~
    - Held over for 3.1+
  - 109 - Jason - ~~gLS Discovery of IP Addresses~~
  - 110 - Jason - ~~perfSONAR-BUOY Data Query Error~~
  - 112 - Jason - ~~supportedEventType Examples for SNMP/pSB~~
  - 116 - Jason - ~~Migrate getRRDData to SNMP.pm~~
  - 62 - Maxim - ~~pingER-MA requires 'hostname" type attribute in endPointPair~~
  - 66 - Maxim - ~~PingER Collection Errors~~
  - 81 - Maxim - ~~PingER Resolution of IPaddress/Hostnames~~
Code Review Topics * Jason reads major areas to address for next time * Maxim points out that the params validate doesn't handle validation errors well (uses Carp, etc.). Could try to extend work Yee did on error classes (use call backs). Can investigate after release for next time. * Aaron/Maxim talk about script to generate POD (so that everyone keeps up with doc) * Maxim/Aaron discuss moving modules out of root of perfSONAR-PS lib. Held over till SMM, but we should take action this time. * Segways into removing unnecessary code in general. Need to be mindful of the shared space. Will require someone to physically do the cleaning.
Other Topics * Maxim brings up that the documentation regarding the NPToolkit is scattered and also not geared to the proper audience at times. * Would like to see things better organized. Currently there are many individual pages that answer some questions, but need a true step by step guide for the absolute beginner. * Jason Points out NPToolkitQuickStart * Rich/Jason/Maxim discuss what some groups (like USAtlas) do. These groups usually maintain their own specific documentation and suggestions and point to ours when necessary. * Joe points out that it will be up to communities to keep up, we can only do so much.
Next meeting will be 3/30

ACTIONS

Jason
- ~~Packaging Finalization~~
- ~~Documentation Finalization~~
- Beta Testing
Marcos
- IP Summarization
- Doc Reviews
- Beta Testing
Guilherme
- Doc Reviews
- Beta Testing
Joe
- Doc Reviews
- Beta Testing
Brian
- Doc Reviews
- Beta Testing
Maxim
- Figure out changes to make NPToolkit documentation easier to follow
- ~~Packaging Finalization~~
- ~~Documentation Finalization~~
- Doc Reviews
- Beta Testing
Aaron
- ~~Update Examples for circuit status~~
- ~~Packaging Finalization~~
- ~~Documentation Finalization~~
- Doc Reviews
- Beta Testing
Martin
Jeff
- pSB Development
- Doc Reviews
- ~~Read CHEP - Give Maxim Feedback~~

20090316Video

Agenda/Minutes

Attendees: * Martin * Guilherme * Marcos * Aaron * Brian * Jason * Maxim * Rich * Joe
Team Updates 1. Internet2
- Jeff - On his way to DICE
- Aaron - Working on release activities (packaging, testing, documentation). Spent much time on preparing new release of Performance Node for Cisco TP Activity.
- Jason - Release management (documentation, bugfixing, packaging, organizing).
- Rich - Worked with BNL to debug installation issues with Performance Node. Bugs fixed with pSB, now working on trying to develop rules and instructions for deploying certain number of nodes to test with for bandwith (and eventually latency). 1. UDEL
- Martin - Speant time working w/ Marcos on IP Summariation code. Started thinking about changes to hLS/gLS architecture
- Guilherme - Code Review
- Marcos - IP Summarization changes. Will do code review when Jason gives info on what and where. 1. ESNet
- Joe - Code reviews of LSReg Daemon. Would like to talk about some issues sometime soon (next call?). Hardened pS installation on ESnet backbone.
- Brian - Code reviews of SNMP MA. Stuck on SNMP MA development of ESnet extension. Will look to create a branch, and work with Jon D. to get further. Expects to work with Jason/Aaron F2F at the SMM. 1. FNAL
- Maxim - Code Review of pSB. Closing bugs for PingER. CHEP Paper. Needs everyone to read the paper, contribute information that may be missing (deployments, data). Deadline weds (will be talking with Phil M Thurs about presentation).
perfSONAR-PS v3.1 Release Update * See pSPS20090327Release for details
- Check individual progress.
- Code reviews by the 20th.
  - COMPLETED
    - Maxim
    - Guilherme
    - Brian
    - Joe
    - Marcos
    - Jason
    - Aaron
  - MISSING
    - Jeff
- Documentation Reviews
  - Developers Finish (3/27)
  - Reviews Due (4/3)
- Open Bugs
  - 19 - Aaron/Maxim - ~~PingER GUI gLS pingable hosts~~
  - 73 - Jason - Client::LS Registration Logic Re-arrange
  - 79 - Jason - ~~Empty gLS/hLS Messages~~
  - 84 - Jason - ~~Exit MA service if no data is present~~
  - 85 - Jason - ~~add verbose logging on errors for ma/ls~~
  - 105 - Jason - hLS Summarization of DCN Data
  - 109 - Jason - ~~gLS Discovery of IP Addresses~~
  - 110 - Jason - ~~perfSONAR-BUOY Data Query Error~~
  - 112 - Jason - ~~supportedEventType Examples for SNMP/pSB~~
  - 116 - Jason - ~~Migrate getRRDData to SNMP.pm~~
  - 62 - Maxim - ~~pingER-MA requires 'hostname" type attribute in endPointPair~~
  - 66 - Maxim - ~~PingER Collection Errors~~
  - 81 - Maxim - ~~PingER Resolution of IPaddress/Hostnames~~ * Code slush status
Circuit Monitoring * Discuss the path forward for circuit monitoring including adding additional performance metrics and handling dynamic circuits. Please see:
- CircuitMonitoring
- CircuitMonitoringMoreDetails * Aaron has some things to correct based on a discussion with Jason/Martin/Jeff * Will send out notice to the list to get more feedback this week.
SNMP-MA ESnet new database integration update * See Brian update. Stalled pending help at the SMM.
CHEP * See Maxim update.
LHC/Atlas deployments (status update) * See Rich update. * Jason adds that Jay Packard is re-engaging to install/develop tools.
GUI followup * Stalled pending release.
GSoC * Application submitted. * Join the mailing lists if interested * Expect to get results of application the end of this week.
Next meeting will be 3/23 * Release Focused.

ACTIONS

Jason
- ~~Get Marcos info on code reviews~~
- ~~Close bugs~~
- ~~Finish packaging~~
- ~~Finish documentation~~
- ~~Code Reviews~~
- Doc Reviews
- ~~Read CHEP - Give Maxim Feedback~~
Aaron
- ~~Address Martin/Jason/Jeff feedback on circuits~~
- ~~Send email regarding circuits for feedback~~
- ~~Finish packaging~~
- ~~Finish documentation~~
- ~~Code Reviews~~
- Doc Reviews
- ~~Read CHEP - Give Maxim Feedback~~
Jeff
- ~~DICE Meeting~~ - Report to Group
- ~~Code Reviews~~
- Doc Reviews
- ~~Read CHEP - Give Maxim Feedback~~
Marcos
- IP Summarization
- ~~Code Reviews~~
- Doc Reviews
- ~~Read CHEP - Give Maxim Feedback~~
Martin
- ~~Read CHEP - Give Maxim Feedback~~
Guilherme
- Doc Reviews
- ~~Read CHEP - Give Maxim Feedback~~
Brian
- Doc Reviews
- ~~Read CHEP - Give Maxim Feedback~~
Maxim
- Doc Reviews
- ~~Read CHEP - Give Maxim Feedback~~
Rich
- Continue to work with USAtlas - bring questions to this group when applicable.
- ~~Read CHEP - Give Maxim Feedback~~
Joe
- Doc Reviews
- ~~Read CHEP - Give Maxim Feedback~~

20090309Video

Agenda/Minutes

Attendees: Brian, Marcos, Jason, Aaron, Jeff, Maxim, Guilherme, Rich, Joe
Team Updates 1. Internet2
- Aaron: Working on next generation pS-NPToolkit. RPMs etc...
- Jason: Release stuff, documentation stuff, gLS summarization issues. GUI progress, bug fixes.
- Jeff: pS-B, logistics for f2f meeting. 1. IU 1. UDEL
- Martin: OGF, identifiers, IS-WG summarization, beginning to look at 'authority' issues in IS (who registered particular information). 1. ESNet
- Joe: Working on nagios monitoring of pS services on ESnet.
- Brian: Working with Joe to integrate Jon's SNMP database backend. Also working on pS-B deployments. Debugging distributed services is hard, we should do as much as we can to make that better. 1. FNAL
- Maxim: CHEP paper, took PAM paper for much content. (Due March 20) Still needs deployment information. MDM of pingER-PS. Looking at pingER-PS GUI. Would like numbers on current and planned deployments. Paper should end up about 10 pages.
perfSONAR-PS v3.1 Release Update * See pSPS20090327Release for details
- Check individual progress.
- Code reviews by the 13th.
- Open Bugs
  - 19 - Aaron/Maxim - ~~PingER GUI gLS pingable hosts~~
  - 73 - Jason - Client::LS Registration Logic Re-arrange
  - 79 - Jason - ~~Empty gLS/hLS Messages~~
  - 84 - Jason - ~~Exit MA service if no data is present~~
  - 85 - Jason - ~~add verbose logging on errors for ma/ls~~
  - 105 - Jason - hLS Summarization of DCN Data
  - 109 - Jason - ~~gLS Discovery of IP Addresses~~
  - 110 - Jason - ~~perfSONAR-BUOY Data Query Error~~
  - 62 - Maxim - ~~pingER-MA requires 'hostname" type attribute in endPointPair~~
  - 66 - Maxim - ~~PingER Collection Errors~~
  - 81 - Maxim - ~~PingER Resolution of IPaddress/Hostnames~~ * Code slush status * Take pS-B out of the 3.1 release schedule, and release it one month later (with same version number)
- Jeff: will put a project planning page up with milestones, and verify with Jason that the milestones are understandable represented in the issue-tracker.
- All: Code reviews in trunk, accepted changes will be copied to the 3.1 branch.
Circuit Monitoring * Discuss the path forward for circuit monitoring including adding additional performance metrics and handling dynamic circuits. Please see:
- CircuitMonitoring
- CircuitMonitoringMoreDetails * Anything we need to discuss here before a joint meeting with IDC developers? * Jeff: Will contact Martin to see if he is ready to use these links to engage the IDC developers on the topic of circuit monitoring.
SNMP-MA ESnet new database integration update * Jason: Will respond to the SNMP-MA questions on-list. Brian: Expects this will only take a couple more days to integrate.
CHEP * status - due date is 20th. * All: Review/add to doc this week.
LHC/Atlas deployments (status update) * BNL bwctl's still having problems.
GUI followup 1. wmap status 1. gLS related GUIs - status
GSoC * Please send ideas.
DICE next week * Jeff is going, please let him know of any topics that need to be discussed with our EU compatriots.
Next meeting will March 16.
Other topics?

20090302Video

Agenda/Minutes

Attendees: Jason, Aaron, Brian, Guilherme, Marcos, Jeff, Maxim
Team Updates 1. Internet2
- Jason: Release management, API's for J Packard, Deployments on observatory hosts. GUIs for NDT and word-cloud.
- Aaron: Knoppix work for future releases - no boot questions. dhcp-client instead of pump, clean-up of how things get mounted to more closely match fedora, write-up for tl1-querying schema.
- Jeff: Project planning, owamp database schema. 1. IU 1. UDEL
- Marcos: Research security/authentication. 1. ESNet
- Joe: Working on Nagios infrastructure for monitoring pS services.
- Brian: pS-B/bwctl on ESnet 1. FNAL
- Maxim: CHEP paper this week. Still have 3 weeks. Code review of pS-B on his list. Coming up with RPM for pingER-PS for MDM. Perhaps integrate keywords into the pingER GUI or pingER admin interface.
perfSONAR-PS v0.10 Release Update * SVN Branching tonight - Update from Aaron.
- Old modules will be removed with the exception of pingER. A release branch will be created. * See pSPS20090327Release for details
- Check individual progress. * Code slush is scheduled for March 6!
Circuit Monitoring * Discuss the path forward for circuit monitoring including adding additional performance metrics and handling dynamic circuits. Please see:
- CircuitMonitoring
- CircuitMonitoringMoreDetails
SNMP-MA ESnet new database integration update * Jon/Brian working on it this week.
CHEP * working on it this week, 3 weeks to finish - due 20th.
LHC/Atlas deployments (status update) * no updates.
GUI followup 1. wmap still on the list - currently stalled. 1. gLS related GUIs under development
- probably adjusting the API.
pS-NPToolkit Administrative GUI * Any feedback now that everyone has had time to see this?
- Brain: Do we really want opt-in system?
- Brian: Default ntp servers all in US?
- Brian: Should we have standard keywords?
GSoC * Please send ideas.
Next meeting will March 9.
Other topics?

20090223Video

Agenda/Minutes

Attendees: Martin, Jason, Joe, Aaron, Guilherme, Jeff
Team Updates 1. Internet2
- Aaron: tl1 alarm/PM monitoring (pulling out internal topology and determining how to represent it). Testing MySQL to see if it effects OWAMP performance.
- Jason: bwctl deployment issues, APAN talk, USATLAS GUI work, Yum repo work, wmap packaging, nagios development, release stuff.
- Jeff: Cisoc telepresence whitepaper, project plan. 1. IU 1. UDEL
- Martin: Geni proposal pS measurements to protogeni, Welcome Guilherme. 1. ESNet
- Joe: L2status monitoring working (really lame still). Working on hardening owamp/bwctl infrastructure. 1. FNAL
perfSONAR-PS v0.10 Release Update * Goal is to get all services out by March 6 (Date Negotiable - Must factor in time to release NPToolkit after) * See pSPS20090327Release for details * See YumRepository for details on the Yum/APT repository
SNMP-MA ESnet new database integration update * Future work.
CHEP
LHC/Atlas deployments * Metadata keyword issues.
GUI followup 1. wmap still on the list
- Jason has worked on packaging. 1. gLS related GUIs under development
- Making progress.
pS-NPToolkit Administrative GUI * Add ability to see what keywords are available. (Might need to add more information into the summary sets to help indicate the 'weight' of a given keyword.) * Aaron update on use cases, and how we would fulfill them.
1. diagnostic host - all on domand, no config.
2. measurement host - long term, hold data, configuration needed.
- Came up with a user-interface definition described at the above link. Basically most configuration is done via HTTP interface with absolutely minimal configuration done on command-line TTY.
GSoC * Please send ideas - we need to apply in the next two weeks.
Next meeting will March 2.
Other topics?

20090217Video

Agenda/Minutes

Attendees: Joe, Maxim, Rich, Aaron, Jason, Jeff, Matt, Brian, Martin
Team Updates 1. Internet2
- Aaron: Worked on UDP testing. Loss happens even by just starting mysql, without even using it. (second thing)
- Jason: bwctl GUIs. LS issue. Release Docs. USAtlas GUIs - sent strawman. Only cares about USAtlas, doesn't want to have to dig. CGI script for SCinet data. Yum/apt repo work (w/PSU & MCNC).
- Rich: JETnet grid of potential services and tools.
- Jeff: teleprescence deployment plan, dice-aa. 1. IU 1. UDEL
- Martin: Working on perfSONAR proposal for GENI Spiral II effort. NML will spin up more, Freek is new co-chair. LSN talk Feb 10, went well. 1. ESNet
- Joe: Trying to get status-ma working again.
- Brian: Getting bwctlds working on ESnet hosts. hLS registration problems. 1. FNAL
- Maxim: KOI box install.
perfSONAR-PS v0.10 Release * Goal is to get all services out by March 6 (Date Negotiable - Must factor in time to release NPToolkit after) * See pSPS20090306Release for details * See YumRepository for details on the Yum/APT repository
New 'type' for SNMP MA. A couple of issues need to be discussed
UDP testing * do we want to support high levels? If so, do we switch what services are run where? * Brian asked "why"? A long discussion boiled down to the decision not to automatically run UDP tests, and not to automatically enable UDP testing. But, if UDP testing is enabled, to limit it to something around 75 Mbps. (This is fast enough to look for many problems, and slow enough to not be a problem with other things running on the host. It is roughly a factor of 10 less than where we saw most problems.)
CHEP
LHC/Atlas deployments
GUI followup 1. wmap still on the list 1. gLS related GUIs under development
GSoC
Next meeting will Feb 23.
Other topics?

20090203Meet

Where

Winter Joint Techs 2009, College Station, TX

When

February 3, 2009 - 6 to ? Central

Attendees

Jason (Internet2)
Aaron (Internet2)
Eric (Internet2)
Jeff (Internet2)
Martin (UDel)
Joe (ESnet)
Brian (ESnet)
Andy Lake (Internet2)
Tom Lehman (ISI East)

Agenda/Minutes

UDP Status on the KOI systems * Brian and Aaron each reported on some testing they had done, but there was nothing definitive yet. Brian will put FreeBSD on two back-to-back KOI systems. Aaron will put the Knoppix on two back-to-back Dell systems. With that information we can hopefully determine if this is OS related, or KOI hardware related.
Tech Report Status * NSF should be credited on the tech-report * It is no problem to liberally take from the tech-report for the CHEP paper.
Test Plan * From memory review of the Test Plan. No changes requested.
RequirementsForCGI * Light discussion on what would be desirable for IS related GUIs. * hLS/gLS issues discussion
- Some discussion of how new measurements will go into the hLS/gLS cloud along with current technical limitations that require all namespaces to be known in advance.
Release Management * Need to set timeline.
- Aiming for a final release of the Knoppix before SMM.
- Jeff/Jason will work on a timeline with that aim. * Features for next release
- pS-B with OWAMP support
- gLS/hLS performance updates
- gLS/hLS GUIs
- CircuitMon (GLIF version)
- minimal-config knoppix * Will attempt to organize a 'regionals' as betatesters * Will likely need a version to support telepresence

20090126Video

Agenda/Minutes

Attendees:
Issue: Release Management * Review Documents:
1. ReleaseManagement
- Does this basic approach work for people?
1. TestPlan
- Suggestions/Comments/Additions
1. RequirementsForDeploymentsCGI
- Suggestions/Comments/Additions * Next perfSONAR-PS (not perf Node yet) Release
- Target Date
- Services Involved
  - Key Functionalities missing?
  - Bugfixes outstanding?
- People Involved
- Outstanding issues
Next meeting will be at Jt Techs * Proposed: Sunday, Feb 1 * Informal Poll of who is available when
Other topics?

20090112Video

Actions

who	task
All	review/comment on gLS GUI: http://code.google.com/p/perfsonar-ps/wiki/RequirementsForDeploymentsCGI
All	review/comment on testing: http://code.google.com/p/perfsonar-ps/wiki/TestPlan
All	(at jt techs) discuss/determine development priorities
aaron	UDP testing with back 2 back dell servers
brian	UDP testing with FreeBSD KOI servers
jason	wmap package
joe/maxim	try wmap package when available and provide feedback
aaron	look into packaging alarm/tl1 MA functionality (and see if JG is interested in trying)
jeff	writeup highlevel roadmap strawman post jt techs
martin	Determine 4 OGF WG status and actions we need to take
martin/brian/maxim	Paper publishing goals/strategy

Agenda/Minutes

Attendees: Jared, John Hicks, Jason, Aaron, Maxim, Joe, Brian, Martin, Rich, Marcos
Team Updates 1. Internet2
- Aaron: UDP testing, owamp & bwctl releases. helping pS-B
- Jason: pS-B packaging and disting out on Internet2 observatory hosts. Jason has been working on release documentation. Also working on making SCinet data available.
- Rich: Met with Koreans - they are interested. It would be good to contact them for APAN meeting.
- Jeff: Working on 1. IU
- John: starting on a visual-backbone MA. Followup on APAN pS workshop. 1. UDEL
- Martin: nothing to update. OGF - should push forward the LS in the OGF-NSI group. Action: Figure out where we are in all 4 groups and what we need to do for next week. 1. Ga Tech 1. ESnet
- Joe: nearly done with security review
- Brian: UDP testing 1. SLAC
- Jarad: nothing to update. Interested in feedback on gmaps. 1. FNAL
- Maxim: helping Tier-2's deploy performance nodes. 1. MCNC
Prioritize brain-storm goals from last meeting. (Time frame is next 6-12 mo) * Community defined measurement meshes. * Community defined dashboards/analysis. * ML or Spectrum gateways or RealTime alarms.
- Persistent queries * Test framework
- scalability
- reliability * traceroute MA/MP
- BGP MA? * How do multiple meshes interact? (BCP?) * Focus on deployments/production
- GUIs
- How do you use these tools to actually debug problems? * GUIs - easy to use, not demoware! * Installation usability * Need more user requirements?
- Need to bring users the first good data GUIs, then we can iterate with users.
- What specific visualizations do the users need? (First data must be there...) * First priority to work on until Jt Techs is 'usability' (stability and installation). Other priorities will be worked on in person at Jt Techs. Jeff will put a high-level road map up for discussion after this meeting.
SC * How much of the one-off coding can be turned into 'real' stuff?
- wmap: Jason has something others can try. Joe and Maxim will try it. Joe will be looking at trying to do something similar to the visible-backbone but perfsonar driven.
- wan-monitoring: Aaron said this is a couple of weeks away. John Graham might be a good user to attempt to work with on this.
Questions regarding gLS data displays - Brian 1. Should we advertise things that are only available to certain groups (e.g. NDT servers on Interent2 backbone are not reachable to home users - do we note this? How?)
- How do we manage expectations?
- We will need ways to discern what data is available to write good GUIs, this issue is not just about the 'publicity' pages.
- (Jeff suggestion: attack this problem by creating a GUI for actually addressing a specific performance problem - that GUI will need to figure this out...)
- Action: Everyone please review RequirementsForDeploymentsCGI
PAM/CHEP paper status 1. Where to resubmit PAM paper and when - Brian/Martin
- Brian: Need something with a results section. Previous paper can be an LBL tech report.
- Lots of discussion on places to publish, and reasons for publishing. 1. Important Dates for CHEP - Maxim
- Maxim: Chep paper should be about usefulness of perfSONAR for LHC community.
Next meeting will be Jan 26, 2:30 pm Eastern Time (Issue call on release mgmt)

Last Updated

$Id$

Useful Links

Reporting an Issue

Using the Issue Tracker

Project Planning

Weekly Meeting Agenda

Support Rotation

Current Support Schedule