dlmarion opened a new issue, #6374: URL: https://github.com/apache/accumulo/issues/6374
The CleanUp step of the Delete Fate operation configures the [MetaDataTableScanner](https://github.com/apache/accumulo/blob/rel/2.1.4/server/manager/src/main/java/org/apache/accumulo/manager/tableOps/delete/CleanUp.java#L95) so that it can compute each tablets location. When all tablets are unassigned, then it can move on to the next step in the operation. However, [MetaDataTableScanner.configureScanner](https://github.com/apache/accumulo/blob/rel/2.1.4/server/base/src/main/java/org/apache/accumulo/server/manager/state/MetaDataTableScanner.java#L92) calls [Manager.merges](https://github.com/apache/accumulo/blob/rel/2.1.4/server/manager/src/main/java/org/apache/accumulo/manager/Manager.java#L1812) which iterates over every table to read its merge information from ZooKeeper. Manager.getMergeInfo is synchronized, causing Delete Fate Ops to block. [MetaDataTableScanner.createTabletLocationState](https://github.com/apache/accumulo/blob/rel/2.1.4/server/base/src/main/java/org/apache/accumulo/server/manager/state/MetaDataTableScanner.java#L164), called by `CleanUp.isReady`, doesn't use the Merge information. It's also possible that when CleanUp calls MetaDataTableScanner.configureScanner, `null` could be passed for the Manager state object. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
