2020-08-10 - Data Migration Subgroup Agenda and Notes

Date

Attendees

Discussion items

TimeItemWhoNotes
5WelcomeDaleSomeone who hasn't done it in a while should take notes.


Ingolfcalendar events (invitations) for this meeting => Ingolf will send bi-weekly events for one hour each.
55Bulk APIsIan W

We will continue discussion of Ian's document outlining the bulk APIs needed to facilitate data migration.

FOLIO Bulk API Support (Google Docs)

Meeting notes (Ingolf):

...

Ian: It is desired to have direct access to the DB in Folio. Jon : sounds good ; Patty: This is out of scope (for this meeting).

Patty: MODUSERS-3 - Getting issue details... STATUS ; Let's talk to Harry.

Dale: Systematic coverage of the modules. There was a meeting (with Jakub ?). An initial document (this one) is outlining this.

Ian: Need "upsert" for Agreements

Dale: Will this document be flashed back to Jakub and Cate ?

Jon: Wherever Bulk support is added, the C, the R the U and the D should be implemented. Otherwise it's inconsistent and confusing.

Ann-Marie: We need to take caution about "Delete". We don't have deletes for instances. We have to make sure that the holdings and items have been removed, first.

Jon: That should be up to the developers.

Tod: Concerns of migration are different to those in the operational system. We have to acknowledge this.

Ian / Jon: There are notes about order of magnitude (of numbers of rows). Everything which is above a certain threshold needs to receive bulk support.

Charlotte: Epic for Bulk Edit is UXPROD-868. Would it make sense for this group based on Ian Wall’s documentation - then write up/update the misc. bulk edit feature/jiras?

Dale: We should go through the modules and see how these kinds of principles should be applied. Then bring that to the developers.

Dale: Do we agree that we need CRUD operations for all bulk APIs ?

Patty: We need to do this through Jira tickets.

Ian: Is it reasonable to start with Inventory and Users ? (some agreement)

Jon: The current performance is way too slow.

Patty: This is not how development works. Ian and Dale could talk to Harry ... no one would do this by just looking at a document.

Tod: That document is not the actual work request. But it sets a kind of common understanding.

Ian: So, do we need a SPIKE or a document to take to the technical council ?

Jon: Step1 is to build the support into the RAML Module Builder. The RMB people need to add the support in a consistent fashion.

Ann-Marie: We are approaching the freeze for RMB. Ann-Marie: Q3 RMB freeze is 8/21, so the end of this current sprint Honeysuckle (Q3 2020)

Dale: Ian, Jon and I get together away from this meeting and flash this out for a few modules. If it looks successful we will create a path for doing this for everything. We will have something to build on.

Charlotte: What Dale just described might be that "SPIKE" and that could be taken to the Technical Council.

Tod: Do we have a written statement what we need for bulk import ? We need a high-level document, maybe just two or three sentences.

Charlotte: High level document - as a vision document for Data migration on the FOLIO wiki

Patty: Just write it as a Jira.

Jeff: What is an acceptable performance level for bulk import ?

Action items

Dale Arntson will set up a topic backlog page.