[ https://issues.apache.org/jira/browse/TEPHRA-224?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15877321#comment-15877321 ]
ASF GitHub Bot commented on TEPHRA-224: --------------------------------------- Github user anew commented on a diff in the pull request: https://github.com/apache/incubator-tephra/pull/38#discussion_r102371884 --- Diff: tephra-core/src/main/java/org/apache/tephra/TxConstants.java --- @@ -376,6 +376,11 @@ public static final String PRUNE_FLUSH_INTERVAL = "data.tx.prune.flush.interval"; /** + * The buffer in seconds used to pad transaction max lifetime while pruning. + */ + public static final String PRUNE_BUFFER_INTERVAL = "data.tx.prune.buffer"; --- End diff -- PRUNE_SAFETY_BUFFER? PRUNE_GRACE_PERIOD? > Handle delay between transaction max lifetime check and data writes while > pruning > --------------------------------------------------------------------------------- > > Key: TEPHRA-224 > URL: https://issues.apache.org/jira/browse/TEPHRA-224 > Project: Tephra > Issue Type: Sub-task > Reporter: Poorna Chandra > Assignee: Poorna Chandra > Fix For: 0.12.0-incubating > > > Transaction pruning relies on inactive transaction bound (a bound that says > at a given time there will be no transactions smaller than inactive > transaction bound that can still be writing data). To determine the inactive > transaction bound for a given time, TEPHRA-199 introduced transaction maximum > lifetime. > However there can be a delay between the transaction max lifetime being > enforced and the write reaching HBase. Hence it would be good to use a buffer > to pad the transaction max lifetime when pruning. -- This message was sent by Atlassian JIRA (v6.3.15#6346)