[ https://issues.apache.org/jira/browse/HBASE-5547?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13251750#comment-13251750 ]
Jesse Yates commented on HBASE-5547: ------------------------------------ @Ted - fair enough. Tried to avoid modifying interfaces, but I guess its not avoidable. Plan is to on failure fire off a call (probably synchronous) to the master to cleanup the FS for that table backup; first writes an 'invalid' marker for that table directory and then recursively deletes the regions and then the table directory in the archive. We should be able to return after writing the invalid marker as it would get cleaned up at some point. Associated unit tests would be: general table archive failure, master died mid-archive failure and recovery cleanup. Also thinking about removing the per-table archive in favor of an overall archive directory oriented to tables (and dropping a choice of which directory to archive, in favor of just using the config value). This makes particular sense for snapshots, but is also nice to ensure easier lookups of table values: /hbase/.archive/[table] rather than the current: /hbase/[table]/.archive where .archive could actually be any directory the user specified on creation time, which can easily lead to 'loss by forgetfulness'. > Don't delete HFiles when in "backup mode" > ----------------------------------------- > > Key: HBASE-5547 > URL: https://issues.apache.org/jira/browse/HBASE-5547 > Project: HBase > Issue Type: New Feature > Reporter: Lars Hofhansl > Assignee: Jesse Yates > > This came up in a discussion I had with Stack. > It would be nice if HBase could be notified that a backup is in progress (via > a znode for example) and in that case either: > 1. rename HFiles to be delete to <file>.bck > 2. rename the HFiles into a special directory > 3. rename them to a general trash directory (which would not need to be tied > to backup mode). > That way it should be able to get a consistent backup based on HFiles (HDFS > snapshots or hard links would be better options here, but we do not have > those). > #1 makes cleanup a bit harder. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira