[ 
https://issues.apache.org/jira/browse/HBASE-26067?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17468268#comment-17468268
 ] 

Josh Elser edited comment on HBASE-26067 at 1/3/22, 11:20 PM:
--------------------------------------------------------------

{quote}I'm looking to see if I can spot something obvious, but maybe the same 
thing happened here.
{quote}
I think I see one thing in DefaultCompactor. -I'll tag you on a PR-

This is what I thought was a problem, but after looking at the master branch, 
maybe it's not the cause of the IOException above

[https://github.com/apache/hbase/blob/HBASE-26067-branch-2/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/compactions/DefaultCompactor.java#L87]
{code:java}
  protected void abortWriter(StoreFileWriter writer) throws IOException {
    Path leftoverFile = writer.getPath();
    try {
      writer.close();
    } catch (IOException e) {
      LOG.warn("Failed to close the writer after an unfinished compaction.", e);
    } finally {
      //this step signals that the target file is no longer writen and can be 
cleaned up
      writer = null;
    }
    ...
  }{code}
Instead of null'ing out the member {{writer}} on the parent Compactor.java 
class, we're just null'ing out the local variable {{{}writer{}}}. I assume this 
is wrong, but I think we have this wrong on the branch-2 backport and in master.


was (Author: elserj):
{quote}I'm looking to see if I can spot something obvious, but maybe the same 
thing happened here.
{quote}
I think I see one thing in DefaultCompactor. I'll tag you on a PR

> Change the way on how we track store file list
> ----------------------------------------------
>
>                 Key: HBASE-26067
>                 URL: https://issues.apache.org/jira/browse/HBASE-26067
>             Project: HBase
>          Issue Type: Umbrella
>          Components: HFile
>            Reporter: Duo Zhang
>            Assignee: Duo Zhang
>            Priority: Major
>             Fix For: 3.0.0-alpha-3
>
>
> Open a separated jira to track the work since it can not be fully included in 
> HBASE-24749.
> I think this could be a landed prior to HBASE-24749, as if this works, we 
> could have different implementations for tracking store file list.



--
This message was sent by Atlassian Jira
(v8.20.1#820001)

Reply via email to