[
https://issues.apache.org/jira/browse/OAK-6207?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Julian Reschke updated OAK-6207:
--------------------------------
Issue Type: Improvement (was: Technical task)
Parent: (was: OAK-1266)
> RDBDocumentStore: allow schema evolution part 2: record schema version when
> updating/inserting rows
> ---------------------------------------------------------------------------------------------------
>
> Key: OAK-6207
> URL: https://issues.apache.org/jira/browse/OAK-6207
> Project: Jackrabbit Oak
> Issue Type: Improvement
> Components: documentmk, rdbmk
> Reporter: Julian Reschke
> Assignee: Julian Reschke
> Labels: candidate_oak_1_0, candidate_oak_1_2, candidate_oak_1_4,
> candidate_oak_1_6
> Fix For: 1.7.0, 1.8
>
> Attachments: OAK-6207.diff, OAK-6207.diff, OAK-6207.diff
>
>
> Currently, new/extra columns get reported on startup but then are ignored.
> This allows evolving the schema without immediately breaking existing
> installs. However, we won't know whether a value in a new column was put
> there deliberately or just wasn't updated properly by old code.
> Idea:
> - add a new "VERSION" column
> - when inserting/updating rows, record the DB schema version understood by
> the code doing the update
> Advantages:
> - allows relatively easy check what versions of code are writing to the DB
> - inexpensive
> - helps finding rows that need a migration step
> Disadvantage:
> - the row itself will be unreliably until all instances connecting to the DB
> maintain it
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
