[
https://issues.apache.org/jira/browse/HBASE-26658?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17476749#comment-17476749
]
Duo Zhang commented on HBASE-26658:
-----------------------------------
It is all about semantic. Here unackedAppends means the WAL entries which are
sent out but haven't been acked yet, so theoretically the only way to remove an
entry from unackedAppends is to receive the ack.
If an entry is not in unackedAppends, we can make sure that it has not been
written out yet, then we are free to tell upper layer that you operation is
failed, but if it is already in unackedAppends, in the current design, there is
no way to rollback then.
So removing entries from unackedAppends without receiving ack is dangerous. You
need to give it another name and think really carefully about the whole
architecture if you really want to do this. Maybe in the current code base it
works, but you need to consider future, as the code may be changed by others.
Thanks.
> AsyncFSWAL.unackedAppends should clear after transfered to
> AsyncFSWAL.toWriteAppends
> --------------------------------------------------------------------------------------
>
> Key: HBASE-26658
> URL: https://issues.apache.org/jira/browse/HBASE-26658
> Project: HBase
> Issue Type: Bug
> Components: wal
> Affects Versions: 3.0.0-alpha-2, 2.4.9
> Reporter: chenglei
> Assignee: chenglei
> Priority: Major
>
> When {{ASyncFSWAL}} syncing to HDFS failed, {{AsyncFSWAL.unackedAppends}}
> are transfered to {{AsyncFSWAL.toWriteAppends}} to avoid data loss, but
> {{AsyncFSWAL.unackedAppends}} itself is not cleared. I think there is no need
> to continue retain them in {{AsyncFSWAL.unackedAppends}} because we would
> open a new HDFS pipeline to resend the {{AsyncFSWAL.unackedAppends}}.
> BTW : It would also simplify the logic for fixing HBASE-25905, current fix
> for HBASE-25905 is somewhat hard to understand. I think the problem to cause
> HBASE-25905 is that {{AsyncFSWAL.unackedAppends}} could not exactly reflect
> the *unacked* for current HDFS pipeline. If we clear
> {{AsyncFSWAL.unackedAppends}} after transferring them to
> {{AsyncFSWAL.toWriteAppends}}, HBASE-25905 could also avoid.
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
