[ https://issues.apache.org/jira/browse/BOOKKEEPER-68?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13112216#comment-13112216 ]
dhruba borthakur commented on BOOKKEEPER-68: -------------------------------------------- I am more inclined to say that for the use case of NN using BK, it makes sense to make one of the standby namenodes to fail the recover-ledger api call (rather than retrying inside the BK client). The standby namenode will go back to ZK to see if a new namenode master has been selected, if not then the standby namenode can retry. Ivan: is there some other approach that you have in mind? > Conditional setData > ------------------- > > Key: BOOKKEEPER-68 > URL: https://issues.apache.org/jira/browse/BOOKKEEPER-68 > Project: Bookkeeper > Issue Type: Bug > Reporter: Flavio Junqueira > Assignee: Flavio Junqueira > Attachments: BOOKKEEPER-68.patch > > > The write to ZooKeeper to store ledger metadata when closing a ledger must be > conditional, otherwise concurrent clients might end up writing in a way that > the update of a client overwrites the update of the other. -- This message is automatically generated by JIRA. For more information on JIRA, see: http://www.atlassian.com/software/jira