hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Pankaj Kumar (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-16509) Add option to LoadIncrementalHFiles which allows skipping unmatched column families
Date Sat, 27 Aug 2016 04:33:20 GMT

    [ https://issues.apache.org/jira/browse/HBASE-16509?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15440666#comment-15440666

Pankaj Kumar commented on HBASE-16509:

I meant, whether "private List<String> unmatchedFamilies" will have the unmatched column
families because in validateFamiliesInHFiles(...) it is added locally,
    List<String> unmatchedFamilies = new ArrayList<String>();
    Iterator<LoadQueueItem> queueIter = queue.iterator();
    while (queueIter.hasNext()) {
      LoadQueueItem lqi = queueIter.next();
      String familyNameInHFile = Bytes.toString(lqi.family);
      if (!familyNames.contains(familyNameInHFile)) {

> Add option to LoadIncrementalHFiles which allows skipping unmatched column families
> -----------------------------------------------------------------------------------
>                 Key: HBASE-16509
>                 URL: https://issues.apache.org/jira/browse/HBASE-16509
>             Project: HBase
>          Issue Type: Improvement
>            Reporter: Ted Yu
>            Assignee: Ted Yu
>         Attachments: 16509.v1.txt, 16509.v2.txt
> During my work on HBASE-15449, there came the need for LoadIncrementalHFiles not to bail
out on unmatched column families.
> The use case is like this:
> * table has two column families, A and B
> * full backup is taken on this table
> * column family B is dropped
> * more data is written to the table
> * incremental backup is performed on the table
> * user performs incremental restore using the latest backup
> Since column family B is gone, LoadIncrementalHFiles would fail with IOException for
family mismatch.
> This issue is to add an option to LoadIncrementalHFiles so that bulk load can load hfiles
for the existing column families.

This message was sent by Atlassian JIRA

View raw message