[
https://issues.apache.org/jira/browse/HBASE-2353?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12849746#action_12849746
]
Karthik Ranganathan commented on HBASE-2353:
--------------------------------------------
Jumping in late here... just wanted to throw in my opinion.
I feel that having the option to configure the behavior is good. I also feel
that we should make correctness the default - because it takes someone some
amount of working knowledge to differentiate between the two. When I think of
any DB (whose internals I do not know), I always assume that it preserves data.
And I almost always expect to tweak some settings to get better performance if
it does not cut my needs - but do not expect to have to tweak something to get
absolute data correctness.
Another fallout of this "sloppy" option is that there is a possibility of data
changing from underneath the application using it. The memstore may return a
certain value when the application queries it, then region server goes down,
replays the log and now the application may get a different answer (this case
the correct one). While ok for the most part, it may not play nice with some
application not aware of this. And its pretty hard to debug as well :)
Just my 2 cents.
> HBASE-2283 removed bulk sync optimization for multi-row puts
> ------------------------------------------------------------
>
> Key: HBASE-2353
> URL: https://issues.apache.org/jira/browse/HBASE-2353
> Project: Hadoop HBase
> Issue Type: Bug
> Reporter: ryan rawson
> Fix For: 0.21.0
>
> Attachments: HBASE-2353-deferred.txt
>
>
> previously to HBASE-2283 we used to call flush/sync once per put(Put[]) call
> (ie: batch of commits). Now we do for every row.
> This makes bulk uploads slower if you are using WAL. Is there an acceptable
> solution to achieve both safety and performance by bulk-sync'ing puts? Or
> would this not work in face of atomic guarantees?
> discuss!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
