2021-07-23 - Sys Ops & Management SIG Agenda and Notes
Date
Attendees
- Ingolf Kuss
- Wayne Schneider
- Tod Olson
- James Fuller
- Chris Rutledge
- Axel Dörrer
- jpnelson
- Brandon Tharp
- Steffen Köhler
- Anton Emelianov (Deactivated)
- Philip Robinson
Meeting Link
- https://zoom.us/j/591934220
- Password: folio-lsp
Goal
- Discuss the TC Reference Data Upgrade Recommendation
- As a reminder: we had discussed this topic several times in SysOps. For example, also look at this (apart from the new Recommendation!)
- Upgrades with Reference data - System Operations and Management SIG - FOLIO Wiki
- Defining data types in FOLIO for automatic upgrades - Technical Council - FOLIO Wiki
- 2020-05-22 - System Operations and Management SIG Agenda and Notes - System Operations and Management SIG - FOLIO Wiki
- Reference Data and Upgrades - Special Interest Groups / Sysops and Management - FOLIO Discussions
- 2020-05-15 - System Operations and Management SIG Agenda & Notes - System Operations and Management SIG - FOLIO Wiki
- ... and probably has been discussed at many other places, in addition
- As a reminder: we had discussed this topic several times in SysOps. For example, also look at this (apart from the new Recommendation!)
Discussion items
Time | Item | Who | Notes |
---|---|---|---|
Find a note taker | Ingolf | ||
The TC Reference Data Upgrade Recommandation | Wayne, Tod, all | Please look at the TC Reference Data Upgrade Recommendation, posted July 1st: Improved Handling of Reference Data During Upgrades - Technical Council - FOLIO Wiki Wayne Schneiderand Tod Olson were in the working group who write the recommendation, and they will be there today. They would like to give an opportunity to respond and ask questions. |
Meeting Notes
Discussion of data migration between upgrades
Concern that system was inconsistent after an upgrade, this has been stablized
In the future Reference Data will overwrite any customization during an upgrade. Technical Council has a recommendation for this
Wayne presents slides (hopefully he can share his deck with the SIG)
- Terminology is very important to understand in this situation
- Reference Data
- Default Data (comes with each FOLIO release, new data added during an update like contributor or material type
- Operational Data
- Question: How large is a reference data entity? More than a label?
- Generally a UUID, potentially a Label or a Code. Some codes are used for two modules two build on each other (circulation rules built on other metadata)
- In inventory search, ISBN comes from IdentifierType of ISBN (IdentifierTypes are reference data)
- Records should be quite contained, but could mean a large number of things operationally
- Working Group came up with a solution
- Split data into two groups Default and Operational
- Default comes from FOLIO and is never changed by the user
- Operational data contains any modifications
- The two sets are resolved during an update with a 3 way comparison
- Old Default vs New Default
- Old Default vs Old Operational
- New Default vs Old Operational
- Diagram of update process using Operator and SME roles
- Two JIRA issues
- UXPROD-3111
- FOLIO-3215 (spike)
- Links
Original Discuss post: https://discuss.folio.org/t/reference-data-and-upgrades/2858
Tech Council working group recommendation: https://folio-org.atlassian.net/wiki/display/TC/Improved+Handling+of+Reference+Data+During+Upgrades
FOLIO-3215: - FOLIO-3215Getting issue details... STATUS
- POC spike is scheduled for R3 2021
- Will be made Production ready in subsequent releases, SIG members that are interested in this should follow these two issues
- Questions
- How does sample data fit into this?
- It is considered throw-away data and is not in scope
- Should reference data always be loaded?
- It depends on your use case
- How does sample data fit into this?
- Remarks
- The main development item will be on creating reports, reconciliation will be done by the users. This process creates labels/tag (deprecated, custom, modified)
- A shortcut could be to modify the "Record Label" but that is a little hacky
- Where the label resides has not been determined yet
- Could be in the database schema
- If you have many reference data, and delete most of the operational data b/c it is not different; then after an upgrade it would need to be removed again
- There are varying opinions
- A lot of reference data customization is for internationalization/multi-lingual, there is another effort to improve internationalization/multi-lingual in general with translation files
- Could more of the process be automated
- A lot of risk with automated deprecation/deletion or modification (see new permissions API purge command as an example)
- The main development item will be on creating reports, reconciliation will be done by the users. This process creates labels/tag (deprecated, custom, modified)