[
https://issues.apache.org/jira/browse/HBASE-14004?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15044308#comment-15044308
]
Heng Chen commented on HBASE-14004:
-----------------------------------
{quote}
To be clear, this means we will hold the WAL.sync request if there are some
entries have already been written out but not acked and never return until we
successfully write them out and get ack back. And if WAL.sync or WAL.write
fails(maybe due to queue full), we will still rollback MemStore since we can
confirm that the WAL entries have not been written out. Right?
{quote}
I have a big concern about it. If we not configure hsync every time( hsync
periodically), it means there are always some entries we make hflush but not
hsync. And as our logical designed, when one hflush failed, we close old wal
and open a new one, the entries which not hsync will be written into new WAL.
If RS crashed at this time, what will happen? Is it means some entries may be
already in place (you have told to client your mutation successed and data was
really in place on DN already) will lost. I think it is a regression. Because
one failed mutation may cause more mutations inconsistency.
I think it is also [~carp84] concern as his problem.
> [Replication] Inconsistency between Memstore and WAL may result in data in
> remote cluster that is not in the origin
> -------------------------------------------------------------------------------------------------------------------
>
> Key: HBASE-14004
> URL: https://issues.apache.org/jira/browse/HBASE-14004
> Project: HBase
> Issue Type: Bug
> Components: regionserver
> Reporter: He Liangliang
> Priority: Critical
> Labels: replication, wal
>
> Looks like the current write path can cause inconsistency between
> memstore/hfile and WAL which cause the slave cluster has more data than the
> master cluster.
> The simplified write path looks like:
> 1. insert record into Memstore
> 2. write record to WAL
> 3. sync WAL
> 4. rollback Memstore if 3 fails
> It's possible that the HDFS sync RPC call fails, but the data is already
> (may partially) transported to the DNs which finally get persisted. As a
> result, the handler will rollback the Memstore and the later flushed HFile
> will also skip this record.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
