hbase-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "stack (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HBASE-1249) Rearchitecting of server, client, API, key format, etc for 0.20
Date Tue, 28 Apr 2009 04:49:31 GMT

    [ https://issues.apache.org/jira/browse/HBASE-1249?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12703494#action_12703494

stack commented on HBASE-1249:

Docs are hard to read because no context.

Let me have a go at interpreting them and commenting on them:


+ DeleteRow is in square brackets.  Whats that mean?
+ Looks like KeyValue is same as we currently have.  Is that right?
+ "Deletes apply only to older StoreFiles" means?  Deletes in current file effect older files
because deletes will have been directly applied in memcache?
+ Will sorting puts before deletes work?  How do I delete an item in memcache? I split the
memcache TreeMap at the put record though I've been passed a delete?  And then iterate over
its tail.  If a put, remove it, then enter the delete?
+ Says "DeleteColumn/Delete" are sorted in ascending lexic order?  Whats that mean?  Ain't
type a single byte?  So ain't it just a case of saying what the code for deletecolumn, deletefamily
is?  Currently, we have a maximum type.  We split the memcache on that and iterate over its
tail.  Deletes sort before so they override any Puts that are present.  I like the idea that
we execute on deletes immediately.  Could make savings here.  But need bit more info on how
this'd all work.
+ In "StoreFile Notes", the idea that we deletes only apply to the next file sounds good but
what about the ryan rawson point of what happens when a compaction? What do we do?  In minor
compaction, we are picking up two files from a possible N.  We are putting together the edits.
 The dictum that deletes only apply to the next file is broke.
+ Don't say nothing on how deletes work regards timestamp.  Does a deletefamily delete all
in a family at the given timestamp?  Or is it all in the family at the timestamp and older?
 What happens if I do a get behind the timestamp?   I can find the values that were "deleted"

More comments to follow.

> Rearchitecting of server, client, API, key format, etc for 0.20
> ---------------------------------------------------------------
>                 Key: HBASE-1249
>                 URL: https://issues.apache.org/jira/browse/HBASE-1249
>             Project: Hadoop HBase
>          Issue Type: Improvement
>            Reporter: Jonathan Gray
>            Priority: Blocker
>             Fix For: 0.20.0
>         Attachments: HBASE-1249-Example-v1.pdf, HBASE-1249-Example-v2.pdf, HBASE-1249-GetQuery-v1.pdf,
HBASE-1249-GetQuery-v2.pdf, HBASE-1249-GetQuery-v3.pdf, HBASE-1249-GetQuery-v4.pdf, HBASE-1249-StoreFile-v1.pdf,
> To discuss all the new and potential issues coming out of the change in key format (HBASE-1234):
zero-copy reads, client binary protocol, update of API (HBASE-880), server optimizations,

This message is automatically generated by JIRA.
You can reply to this email to add a comment to the issue online.

View raw message