[
https://issues.apache.org/jira/browse/HBASE-27231?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17763825#comment-17763825
]
Andrew Kyle Purtell edited comment on HBASE-27231 at 9/11/23 5:01 PM:
----------------------------------------------------------------------
I think we can. I cherry picked the master commit for this JIRA from master
branch to our internal fork of 2.5.5 with a minor conflict (this remains to be
seen) and only one WAL unit test is now not passing, and I think it is because
the test itself is no longer valid. I will report back when the internal change
is all green.
was (Author: apurtell):
I think we can. I cherry picked the master commit for this JIRA from master
branch to our internal fork of 2.5.5 and only one unit test is not passing, and
I think it is because the test itself is no longer valid. I will report back
when the internal change is all green.
> FSHLog should retry writing WAL entries when syncs to HDFS failed.
> ------------------------------------------------------------------
>
> Key: HBASE-27231
> URL: https://issues.apache.org/jira/browse/HBASE-27231
> Project: HBase
> Issue Type: Improvement
> Components: wal
> Affects Versions: 3.0.0-alpha-4
> Reporter: chenglei
> Assignee: chenglei
> Priority: Major
> Fix For: 3.0.0-beta-1
>
>
> Just as HBASE-27223 said, basically, if the {{WAL}} write to HDFS fails, we
> do not know whether the data has been persistent or not. The implementation
> for {{AsyncFSWAL}}, is to open a new writer and try to write the WAL entries
> again, and then adding logic in WAL split and replay to deal with duplicate
> entries. But for {{FSHLog}}, it does not have the same logic with
> {{AsyncFSWAL}}, when {{ProtobufLogWriter.append}} and
> {{ProtobufLogWriter.sync}} failed, {{FSHLog.sync}} immediately throws the
> exception thrown by {{ProtobufLogWriter.append}} and
> {{ProtobufLogWriter.sync}} , we should implement the same retry logic as
> {{AsyncFSWAL}}, so {{WAL.sync}} could only throw {{TimeoutIOException}} and
> we could uniformly abort the RegionServer when {{WAL.sync}} failed.
> The basic idea is because both {{FSHLog.RingBufferEventHandler}} and
> {{AsyncFSWAL.consumeExecutor}} are single-thread, we could reuse the logic
> in {{AsyncWAL}} and move the most code in {{AsyncWAL}} upward to
> {{AbstractFSWAL}} , and just adapting the {{SyncRunner}} in {{FSHLog}} to the
> logic in {{AsyncWriter.sync}}. Once we do that, most logic in {{AsyncWAL}}
> and {{FSHLog}} are unified, just how to sync the {{writer}} is different.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
