2019-06-21 - System Operations and Management SIG Agenda and Notes

• Welcome
• note taker - Harry

All meeting notes: https://docs.google.com/document/d/1uhMMFDMMMckC2r946j89ECb0j0HwssOCCsmKCqMgkIs

SysOps SIG notes:

https://folio-org.atlassian.net/wiki/pages/viewpage.action?pageId=2097539

https://folio-org.atlassian.net/wiki/display/SYSOPS/2019-06-17+F2F+Session+2

Data migration working group notes:

https://folio-org.atlassian.net/wiki/pages/viewpage.action?pageId=5049907

https://folio-org.atlassian.net/wiki/display/SYSOPS/2019-06-18+f2f+data+migration+subgroup+meeting+notes

Notes:

Overview of F2F meeting.

Please see all meeting notes:  https://docs.google.com/document/d/1uhMMFDMMMckC2r946j89ECb0j0HwssOCCsmKCqMgkIs

1. What is a reference production deployment for FOLIO?  A container based reference deployment.  There are other options.  Will use Kuberneties and Rancher for management of containers.  This is where we will focus with documentation and "play books".
2. Dependency resolution and management of it.  What is needed for production deployment dependencies of the front end?  Concluded that OKAPI should provide metadata via YAML for posting to Kuberneties and Rancher.
3. Upgrading and updating.  A big piece of tech debt that needs to be taken on.  Landed on blue/green upgrade patter will meet our requirements sufficiently. Cluster switching.  Needs to happen very soon.  Very high priority.  
4. Security (this did not get covered)

Lowering the barriers to entry.  Make it easier to get a sandbox running.  The Kuberneties setup might be overkill for a sandbox setup.  Is it possible to to make something lighter than what we have today (vagrant boxes)?  A more stripped down version might be needed.  It currently uses too much memory.  This is technical debt that needs to be resolved as well.  Might need to do something about Java. 

Can you run Okapi without containers?  Yes.  

--------

Data Migration session one:

2019-06-17 - Folio F2F: Data Migration Planning & Implementation

3 needs identified:

1. Bulk API's for loading data.
2. MARC data is the most complex that needs to be mapped to the inventory data model.  The mapping tool is currently a show stopper as the mapping can't be easily modified.
3. Matching records and maintaining relationships between barcodes, overlay, deltas, etc. 

Second session focused on tooling:

2019-06-18 f2f data migration subgroup meeting notes

With some work, we could create a migration manual today.  Wayne will propose as a goal for the data migration group to deliver this quarter.

Bulk API's are the only major issue, but are planned to be upgraded ASAP.

---------

Realtime integrations were discussed at the meeting.

Caiasoft discussions are in process.

*** SysOps SIG can spin up temporary sub groups to help inform or provide requirements to organizations that are planning on building out integrations.  If you want input, please come talk to the SysOps SIG and we can provide what is needed.