Hernan Gelaf-Romer created HBASE-29147:
------------------------------------------
Summary: Full backups should clear the bulkloads table on success
Key: HBASE-29147
URL: https://issues.apache.org/jira/browse/HBASE-29147
Project: HBase
Issue Type: Improvement
Components: backup&restore
Reporter: Hernan Gelaf-Romer
At my job, we've run into issues where the system table to track bulkloads for
backups can become needlessly large. The bulkload table is only cleared when an
incremental backup succeeds. This can cause a large buildup of bulkload rows,
making subsequent runs of incremental slower, and more prone to failure. This
behavior can pretty quickly get us into a death spiral.
This also makes the full table scans against the system table more expensive,
causing unnecessary load on the cluster.
It would be nice if a full backup could clear these bulkloaded rows, as we can
consider them successfully backed up.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
