[
https://issues.apache.org/jira/browse/OAK-3534?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14968763#comment-14968763
]
Davide Giannella commented on OAK-3534:
---------------------------------------
I think it should try X times and then set some properties. The
{{IndexEditorProvider}} could come and go as of OSGi and it could be
that at the first run it fails but it could succeed in the second or
third.
Probably a decent approach is retrying for 5 times. If keeps failing
set {{reindex=failed}} so that it will be visible in the repository as
well as tracking it in the logs.
> Endless reindexing of async if no provider
> ------------------------------------------
>
> Key: OAK-3534
> URL: https://issues.apache.org/jira/browse/OAK-3534
> Project: Jackrabbit Oak
> Issue Type: Bug
> Components: core, query
> Affects Versions: 1.3.8
> Reporter: Davide Giannella
>
> Placeholder issue for the moment. Require more investigations and a
> test case.
> It seems that if an index definition has {{reindex=true}} but no index
> provider is able to serve the index type, the re-indexing process will
> retry endlessly.
> Possible way to reproduce
> - create an initial content with an index definition of type lucene
> and reindex=true
> - start the repository without the lucene IndexProvider or having it
> disabled.
> - see in the logs that the reindex will always re-index from scratch
> as there are no checkpoints.
> Should the reindex fail and stop after some attempts?
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
