20180919 Technical Council Meeting Notes
- Mike Gorrell
- Tod Olson
Date
Guests
Discussion items
| Time | Item | Who | Notes |
|---|---|---|---|
| 10 min | Update on FOLIO Stakeholder meeting | Meeting Recording: upcoming FOLIO meetups in Chicago and Columbus Update on stakeholder meeting held in Boston last week 3 new teams joining FOLIO - planning Workflow engine PoC, stakeholders approved PoC and provided funding, will contract with BP3 for some input. Increased support for stable demo environment MV: is demo.folio.org insufficient? From Winkler and others it does not seem to be mature, there were issues with the demo in California. Problems had nothing to do with the FOLIO, but with some integrations that were not yet stable. There have been problems with just sharing the main admin login. Need abetter way to provide full access, rather than sharing admin login broadly. EBSCO plan to fund more dev teams. MV: have interviewed folks from EPAM, read to form 3 teams. in phases First team will work closely with Stacks on Acquisitions Around Oct. 1, other teams will start. Assignments are not yet determined. Initial funding approved for next quarter, maybe reassess over Q (Tod): How do we keep that sustainable? One of EPAM team combination of FSE and Stacks, PO is Anne-Marie. Stacks backlog is large, and they are are down to one developer. The very first EPAM team has Batch Loader, A-M will still be 25% PO on the, Tiana is the EPAM scrum master and will also take PO responsibilities Other areas for EPAM teams may include password management, and OAI-PHM (needed by Chalmers), maybe e-Holdings phase 2. Gorrell will facilitate discussion next week about how to coordinate between GBV and EPAM so work is not duplicated | |
| 20 min | Workflow Engine PoC | Update on approval, PoC schedule, SOW with BP3. Start date is currently set at Oct. 8. Statement of Work document created by BP3 based on the PoC document. Jakub then added some things about providing technical design documents. Clarifications: all work will be done Camunda Community Edition, we will want advice on how to deploy Camunda. SOW: https://docs.google.com/document/d/1nhP8QSBSuthL1ABjdlhhK0ixHLPpAGnj0S9Q25vhjUA/edit?usp=sharing Seb and Peter will travel to Austin next week to get BP3 oriented. TAMU has offered a resource to ride shotgun on this and review reports and code contributions, possibly Jeremy Huff. He's already on the core team, so will change capacity. | |
| 10 min | ERM Discussion | Update on ERM discussions with GBV, EBSCO, SIG subgroup Covered in Stakeholder update above. Scheduling a call next Tuesday to be certain that any efforts by a new team are aligned with what Knowledge Integration and GBV are doing. | |
| 10 min | Update on Q3, Q4 planning | Status/updates/news on Q3 efforts So far according to schedule. Q3 environment up as of early this week but needs some more work. See at http://folio-q3.aws.indexdata.com. Want to have Outstanding issues captured in JIRA with filter. Known Critical Q3 Issues Most issues related to building front-end modules, these will be solved. Also loading data, will likely be solved by workarounds. When all done, will provide a link to the release notes. MV: regarding performance, Anton has listed some APIs with more than 10sec response time. We agreed to focus on mod-inventory. How is ist going with the Postgres full-text searching? Test environment difficult to update (e.g. full new versions of modules with new schemas). In our testing we have seen a factor of 10 improvement in some areas. Right now we cannot reliably run with new modules, we will do some manual testing and provide results. But that won't show up on Anton's dashboard until the test environment is resolved. | |
| 30 min | AES Discussion | Update on document and discussion. Some Major revisions made. Changes are: Filling in some previous ToDo items (filled in with green text) and provided some diagrams. Provided specifica about how to provide message queue and some on Rules (while trying not to be too proscriptive). Under implementation highlight some items of contention. See three components: a stream registration, a stream processor, and a message stream client which would be specific to a particular message queue technology. Status of document: close to becoming an RFC, turn into Markdown and pull into github. Discussion on #tech-council channel. Will need to assign reviewers. From Matt Jones: Yarn uses github for their RFC discussion, here's an example: https://github.com/yarnpkg/rfcs/pull/101 Meanwhile, how should Hongwei Ji and others process on their work to flow data to a data lake? VBar, we will have to take a gamble, but is leaning more towards Kafka, so we could start with that. But we will want to decouple, have a layer of abstraction so the mechanism of generating messages is not as important. We now have the post filter, and can just use that. General agreement that using the post filter will be sufficient to let work proceed. This may also be able to tie into the Camunda PoC so may have enough time to be ready for that. May wish to push BP3 start date Oct. 15. One thing to remember is that this solution will not work with those few cases where work does not originate with with a user request. So not a complete solution, but will let us move on the PoC. | |
| Other business? |