[
https://issues.apache.org/jira/browse/CASSANDRA-10681?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15016841#comment-15016841
]
Sam Tunnicliffe commented on CASSANDRA-10681:
---------------------------------------------
My solution is almost identical to your first one, the only real difference
being to group the tasks by the class of the task itself, rather than by the
index class. That way, all existing index implementations can continue to share
a single task without any modifications. To avoid constructing some task
instances which are immediately discarded, I did consider having the new method
on {{Index}} return the class of the build task instead of an instance and
handling construction via reflection, but the added complexity doesn't seem
worth it right now. Patch on top of your initial commit
[here|https://github.com/beobal/cassandra/commits/10681-trunk]
> make index building pluggable via IndexBuildTask
> ------------------------------------------------
>
> Key: CASSANDRA-10681
> URL: https://issues.apache.org/jira/browse/CASSANDRA-10681
> Project: Cassandra
> Issue Type: Sub-task
> Components: Local Write-Read Paths
> Reporter: Pavel Yaskevich
> Assignee: Pavel Yaskevich
> Priority: Minor
> Labels: sasi
> Fix For: 3.x
>
> Attachments: 0001-add-table-support-for-multi-table-builds.patch,
> 0001-make-index-building-pluggable-via-IndexBuildTask.patch
>
>
> Currently index building assumes one and only way to build all of the indexes
> - through SecondaryIndexBuilder - which merges all of the sstables together,
> collates columns etc. Such works fine for built-in indexes but not for SASI
> since it's attaches to every SSTable individually. We need a "IndexBuildTask"
> interface (based on CompactionInfo.Holder) to be returned from Index on
> demand to give power to SI interface implementers to decide how build should
> work. This might be less effective for CassandraIndex, since this effectively
> means that collation will have to be done multiple times on the same data,
> but nevertheless is a good compromise for clean interface to outside world.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
