2021-1-19 Meeting Notes

Ian: Koha does not have conflicts in production. 

Jacquie: sees notifications daily that another user has a record open, or that a batch process is running and must wait until the batch is done.

Christie: Sees "errors" in the error log that a batch process collision has happened.

Tod: In Horizon, catalogers would get messages when records open in the editor for a long time.

Christie: Some systems can lock the record while it's being exported so the data doesn't change while the record can be exported

Kelly: Wonders if there's something in Koha where 

Ian: Koha has option to update just fields in a record as opposed to a full record where FOLIO must replace the entire record. Usually if you have people during work hours doing work, one doesn't usually run large batch processes at that time.

Christie: updating holdings, bringing in records, updating bibs - doing one process at a time isn't going to work for her institution.

Kelly: so Jakub's proposal isn't going to work?

Christie: don't know if that's a bug or blocker. A lot of the workarounds won't be feasible - for example taking data out and working on it, and then putting it back. 

Kelly: Jakub's proposal assumes that Optimistic Locking is the solution, not that Pessimistic Locking would be implemented. Rely on logs to report collisions.

Christie: Need to know risks for planning. Won't be able to do certain processes at the same time. It is not an uncommon occurrence at UChicago.

What kind of daily, manual analysis is going to have to be put in place to determine if this has happened? 

Batch APIs throughout the system would help. 

Working on updating individual fields - if that was more robust maybe that would help as well. Storage is that there's two fields - the ID field and the JSONB field with the rest of the data. Will need more detailed logging.

Lehigh - batch updates before staff come in, if conflict has happened, they aren't aware. 

Could make a conflict happen if that would be a useful test. Could look for debris from the collision if we knew what to look for. View the log while testing? The writing, saving and re-indexing of blocks of records (1,000-10,000) can take some time. Perhaps collisions happen during that time?

Steve: sees collisions when people are updating authorities, as well as when two people are editing the same record. Has tested FOLIO with updating notes on item records and has seen where second person saved and the first person's changes that were saved are not gone. 

• New release is out and bugs are being dealt with by the Support SIG
• Support SIG - reviews tickets daily, anything that has a P3 priority or above. Make sure the bugs are being addressed, that the POs are involved. In-production libraries can also review the Support Dashboard to know what's been reported. Can add your name to the list of Affected Institution for issues. 
• If you're not on the list of Affected Institutions - contact Kelly Drake or Anya Arnold
• How to submit tickets: wiki page with instructions
• Don't submit a bug if it's not a production bug. Verify that it's a FOLIO software issue and not a hosting issue. 
• Choose Support project, please include details and screenshots if necessary. Can tag the PO.
• Libraries going live on Iris can report issues once they are on Honeysuckle.
• Also a Support Channel on Slack. More development focused, Support SIG does not take bugs from that channel. Please enter Jira issues.

Community Contributed Resources 

Kristen Martin has folder in Google Drive for Resource Management documentation.

Receiving workflow demos deferred until Honeysuckle