[
https://issues.apache.org/jira/browse/HBASE-27775?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Duo Zhang resolved HBASE-27775.
-------------------------------
Fix Version/s: HBASE-27109
Hadoop Flags: Reviewed
Resolution: Fixed
Merged to branch HBASE-27109.
Thanks [~heliangjun] for reviewing!
> Use a separate WAL provider for hbase:replication table
> -------------------------------------------------------
>
> Key: HBASE-27775
> URL: https://issues.apache.org/jira/browse/HBASE-27775
> Project: HBase
> Issue Type: Sub-task
> Components: Replication
> Reporter: Duo Zhang
> Assignee: Duo Zhang
> Priority: Major
> Fix For: HBASE-27109
>
>
> TestMultiSlaveReplication fails when using table based replication queue
> storage
> The test uses a small wal rolling size for testing more operations
> {code}
> // smaller block size and capacity to trigger more operations
> // and test them
> conf1.setInt("hbase.regionserver.hlog.blocksize", 1024 * 20);
> conf1.setInt("replication.source.size.capacity", 1024);
> {code}
> But after moving to table based replication queue storage, the update to
> hbase:replication tablr will also be append to the same WAL file, so it will
> make the replicaton worker move forward, and then it will update the
> hbase:replication table since the offset has been changed, so then it will
> make the replication work again, and so on...
> This will make the wal file size reach the configured size easily and trigger
> unexpected wal rolling, and then fail the test.
> For fixing the test, we can increase the wal rolling size, or we can use
> multi wal, to separate the wal files of hbase:replication table along. so it
> will not mess up the normal wal files.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
