hbase-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Stack <st...@duboce.net>
Subject Re: Major vs Minor compactions
Date Thu, 30 May 2013 02:56:29 GMT
On Wed, May 29, 2013 at 1:27 PM, Varun Sharma <varun@pinterest.com> wrote:

> Hi,
>
> I am working on some compaction coprocessors for a column family with
> versions set to 1. I am using the preCompact hook to wrap a scanner around
> the compaction scanner.
>
> I wanted to know what to expect from the minor compaction output. Can I
> assume the following:
> 1) Versions for each cell will be 1
> 2) Delete markers will be retained
> 3) Corresponding versions older than the delete markers shall be removed
>
> Am I missing anything here ?
>

How you going to ensure your cp fires ONLY on a minor compaction?

In your version of hbase, minors can be promoted to be majors if all files
make it into the selection; how you going to deal?

Yeah, if minors, deletes (currently) are retained (unless promoted to
major).

At a guess, the versions should be weeded by the time your cp gets a whack
at the content (don't take my word for it though).

St.Ack

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message