[ https://issues.apache.org/jira/browse/OOZIE-1319?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14010674#comment-14010674 ]
Robert Kanter commented on OOZIE-1319: -------------------------------------- {quote}When oozie server is back up after more than an hour of downtime, you may only skip the first hour of actions, but you still will materialize and run all actions with nominal times in the past one or several hours during downtime since CoordActionInputCheckXCommand runs before CoordMaterializationXCommand.{quote} By CoordMaterializationXCommand, did you mean CoordMaterializeTransitionXCommand? Doesn't CoordActionInputCheckXCommand get run periodically to check if an action's dependencies are met? {quote}Second, you CoordActionInputCheckXCommand relies on a big assumption that the latest action will get called first by CoordActionInputCheckXCommand since if the second to the latest one gets called first in CoordActionInputCheckXCommand, it will mark the latest action which is supposed to run as "SKIPPED".{quote} I don't think that's the case. CoordActionInputCheckXCommand looks at the next action, but it only edits the "current' action. That is, if it's checking action 5, it will look at action 6's nominal time, but it would set action 5's status only, not action 6's; that happens in a different call to CoordActionInputCheckXCommand. I don't think the order makes a difference here. Can you give an example of what you mean? The new approach I used should inherently work because it's doing something very similar to how the timeout stuff works, and that doesn't have any problems AFAIK. > "LAST_ONLY" in execution control for coordinator job still runs all the > actions > ------------------------------------------------------------------------------- > > Key: OOZIE-1319 > URL: https://issues.apache.org/jira/browse/OOZIE-1319 > Project: Oozie > Issue Type: Bug > Reporter: Bowen Zhang > Assignee: Robert Kanter > Attachments: OOZIE-1319.patch, OOZIE-1319.patch, OOZIE-1319.patch, > OOZIE-1319.patch, OOZIE-1319.patch, oozie-1319.patch > > > In execute() of CoordJobGetReadyActionsJPAExecutor.java, once we retrieve the > top item from a "LIFO" query result, we do not discard or delete the > remaining items from the result list. As a result, the next time execute() is > invoked, we will be retrieving the next item in line. Consequently, LAST_ONLY > strategy will also execute all ready actions for a given coordinator job, > making it no different than LIFO. -- This message was sent by Atlassian JIRA (v6.2#6252)