[
https://issues.apache.org/jira/browse/PHOENIX-4021?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16086476#comment-16086476
]
Lars Hofhansl commented on PHOENIX-4021:
----------------------------------------
Yep. Cache the (H)Connection and create HTable from that. (I spent a lot of
time on HBase making it work that way).
There is no need to ever cache HTables, nor is it recommended (due to the all
the reasons Geoffrey lists).
> Remove CachingHTableFactory
> ---------------------------
>
> Key: PHOENIX-4021
> URL: https://issues.apache.org/jira/browse/PHOENIX-4021
> Project: Phoenix
> Issue Type: Bug
> Affects Versions: 4.11.0
> Reporter: Geoffrey Jacoby
> Assignee: Geoffrey Jacoby
> Fix For: 4.12.0
>
>
> CachingHTableFactory is used as a performance optimization when writing to
> global indexes so that HTable instances are cached and later automatically
> cleaned up, rather than instantiated each time we write to an index.
> This should be removed for two reasons:
> 1. It opens us up to race conditions, because HTables aren't threadsafe, but
> CachingHTableFactory doesn't guard against two threads both grabbing the same
> HTable and using it simultaneously. Since all ops going through a region
> share the same IndexWriter and ParallelWriterIndexCommitter, and hence the
> same CachingHTableFactory, that means separate operations can both be holding
> the same HTable.
> 2. According to discussion on PHOENIX-3159, and offline discussions I've had
> with [~apurtell], HBase 1.x and above make creating throwaway HTable
> instances cheap so the caching is no longer needed.
> For 4.x-HBase-1.x and master, we should remove CachingHTableFactory, and for
> 4.x-HBase-0.98, we should either get rid of it (if it's not too much of a
> perf hit) or at least make it threadsafe.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
