.. meta::
:keywords: DataONE, CCIT, 20120509, VTC
Developers Call - 2012-05-09
============================
:attendees: Dave, Matt, Andy, Ben, Bob S., Bruce, Jeff, Mark, Rob, Ryan, Skye, Chris, Rebecca, Roger, Robert
Call Info:
1. 09May, 2012 at 16:00 EDT.
https://www1.gotomeeting.com/join/727082329
2. Use your microphone and speakers (VoIP) - a headset is
recommended. Or, call in using your telephone.
Dial 1 (805) 309-0014
Access Code: 727-082-329
Audio PIN: Shown after joining the meeting
Meeting ID: 727-082-329
Agenda and Notes
----------------
* General update on DataONE CI development and overal schedule
- APIs
- Schema
- API definitions
- Java and python implementations
- Implementations
- CN Stack
- Metacat MN
- Generic MN
- Mercury MN
- Dryad MN
- ITK:
- client libraries in java and python
- Command Line Client (CLI)
- (later) R-client
- (later) DataONE Drive
- Environments
- Development: highly volatile, used for development purposes, running trunk
- Sandbox: more stable than the development environment, running trunk
https://cn-sandbox-unm-1.dataone.org/onemercury/ (current one-mercury version)
- Staging: more stable than sandbox, running (primarily) stable packages
- https://cn-stage.dataone.org/
- Production: operational infrastructure. Tight management, only stable packages.
- Timeline for deployment (approximate)
- Final testing of unstable / trunk development stream
- Generate stable packages
- Install production CNs
- Start bringing on MNs
* Discussion on the process for bringing online initial set of member nodes
Member node tester: http://mncheck.test.dataone.org:8080/d1_web_test_site-v1/list
- tests all tiers 1-4, except those methods that require interaction with CNs
- would be helpful to have this integrated into development stream
- These are all junit tests so should be fairly easy to integrate with build environment
Member Nodes to bring online in first round:
http://mule1.dataone.org/OperationDocs/membernodes.html
Three or more "replication target" nodes co-located with CNs
- Metacat and GMN based
- Approx 2TB space available for each (license and OS limitation), goal is to have 16TB each.
- KNB (Metacat, tier 4)
- PISCO (Metacat, tier 4)
- LTER (Metacat, tier 4)
- SANParks (Metacat, tier 4)
- ESA (Metcat, tier 4)
- Merritt (Metacat, tier 1)
- Dryad (Dryad, tier 1)
- DAAC (Mercury, tier 1)
- Clearing House (Mercury, tier 1)
General process will be to bring on the nodes with already replicated content (i.e. KNB, PISCO, LTER, SAN, ESA) first, followed by other nodes.
- Bring on native content first, followed by content that has been replicated to the node
- Synchronize first, then turn on replication
- Synchronization runs at about 1-2 seconds per object (< 100,000 / day)
Metacat V2
- some issues with existing data (e.g. metadata validation issues due to new parser libraries etc)
- What are the expectations for successful replication + sync? Ans: All valid content synced and replicated
- Verification of replication, synchronization will be challenging due to the large numbers of objects to be checked
- Need an audit mechanism e.g. align identifiers recorded by CN with identifiers reported by MN
- Test using valid content, large numbers (10's of thousands of objects that are known to be good)
Dryad MN
- major changes in development process, led to delays on implementation
- working on bringing things inline fairly soon
- Some review of packaging / ORE generation required
* Scheduling for possible CCIT meeting sometime within the next three months.
- Venue options
- Dates to avoid
- Possible dates: