You are viewing a plain text version of this content. The canonical link for it is here.
Posted to oak-issues@jackrabbit.apache.org by "Alex Parvulescu (JIRA)" <ji...@apache.org> on 2014/10/21 13:53:33 UTC
[jira] [Comment Edited] (OAK-2203) Full reindexing is triggered
when the IndexEditor is missing
[ https://issues.apache.org/jira/browse/OAK-2203?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14178317#comment-14178317 ]
Alex Parvulescu edited comment on OAK-2203 at 10/21/14 11:52 AM:
-----------------------------------------------------------------
bq. async
There's no need for waiting. I like #1: just loop again without applying the index updates, treat the missing index like a commit exception. the issue is with indexes that are missing permanently/are disabled. We should at least add an ignored type (or list of types) to make sure we don't block the index updates indefinitely (like a type="disabled").
bq. sync case
I wouldn't resort to waiting, just keep the current behavior, mark for reindex.
was (Author: alex.parvulescu):
bs. async
There's no need for waiting. I like #1: just loop again without applying the index updates, treat the missing index like a commit exception. the issue is with indexes that are missing permanently/are disabled. We should at least add an ignored type (or list of types) to make sure we don't block the index updates indefinitely (like a type="disabled").
bq. sync case
I wouldn't resort to waiting, just keep the current behavior, mark for reindex.
> Full reindexing is triggered when the IndexEditor is missing
> ------------------------------------------------------------
>
> Key: OAK-2203
> URL: https://issues.apache.org/jira/browse/OAK-2203
> Project: Jackrabbit Oak
> Issue Type: Bug
> Components: core, oak-lucene, oak-solr
> Affects Versions: 1.0, 1.0.1, 1.0.2, 1.0.3, 1.0.4, 1.0.5, 1.0.6, 1.1.0, 1.0.7, 1.1.1
> Reporter: Alex Parvulescu
> Assignee: Alex Parvulescu
> Fix For: 1.0.8, 1.1.2
>
>
> Currently the IndexUpdate class responsible for handling the index editors is a bit eager in triggering a reindex in the case where is doesn't find the editor for a specific index type [0].
> This has a few drawbacks:
> - if you update a bundle with an index impl (ex: oak-lucene) IndexUpdate will catch that downtime, see the index as offline and ask for a reindex. not cool.
> - it will eagerly set 'reindex=true' even on nodes that are not index definitions, causing mild annoyances (OAK-2024).
> This begs the question: when should we automatically mark for reindex an index definition?
> [0] https://github.com/apache/jackrabbit-oak/blob/trunk/oak-core/src/main/java/org/apache/jackrabbit/oak/plugins/index/IndexUpdate.java#L162
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)