[ https://issues.apache.org/jira/browse/YARN-2637?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14233530#comment-14233530 ]
Craig Welch commented on YARN-2637: ----------------------------------- Hmmm, [~leftnoteasy] option 1 does have the possible issue you describe, and the issue with possibly starving all other queues if one queue has the am percent set higher than the others I mentioned above. The approach of only enforcing the limit if at least one application is running was the approach I was thinking of if we went with 2 - the other being to not add the new app in when doing the check (so it's only retroactive to what has started), but I like the former better as it will reduce the overage as much as possible. Obviously, either approach has the potential to allow things to exceed the maxampercent if there are a large number of queues, but there are tradeoffs either way, it's probably a smaller risk... I'll see about a patch for approach 2. > maximum-am-resource-percent could be violated when resource of AM is > > minimumAllocation > ---------------------------------------------------------------------------------------- > > Key: YARN-2637 > URL: https://issues.apache.org/jira/browse/YARN-2637 > Project: Hadoop YARN > Issue Type: Bug > Components: resourcemanager > Affects Versions: 2.6.0 > Reporter: Wangda Tan > Assignee: Craig Welch > Priority: Critical > Attachments: YARN-2637.0.patch, YARN-2637.1.patch, > YARN-2637.12.patch, YARN-2637.13.patch, YARN-2637.2.patch, YARN-2637.6.patch, > YARN-2637.7.patch, YARN-2637.9.patch > > > Currently, number of AM in leaf queue will be calculated in following way: > {code} > max_am_resource = queue_max_capacity * maximum_am_resource_percent > #max_am_number = max_am_resource / minimum_allocation > #max_am_number_for_each_user = #max_am_number * userlimit * userlimit_factor > {code} > And when submit new application to RM, it will check if an app can be > activated in following way: > {code} > for (Iterator<FiCaSchedulerApp> i=pendingApplications.iterator(); > i.hasNext(); ) { > FiCaSchedulerApp application = i.next(); > > // Check queue limit > if (getNumActiveApplications() >= getMaximumActiveApplications()) { > break; > } > > // Check user limit > User user = getUser(application.getUser()); > if (user.getActiveApplications() < > getMaximumActiveApplicationsPerUser()) { > user.activateApplication(); > activeApplications.add(application); > i.remove(); > LOG.info("Application " + application.getApplicationId() + > " from user: " + application.getUser() + > " activated in queue: " + getQueueName()); > } > } > {code} > An example is, > If a queue has capacity = 1G, max_am_resource_percent = 0.2, the maximum > resource that AM can use is 200M, assuming minimum_allocation=1M, #am can be > launched is 200, and if user uses 5M for each AM (> minimum_allocation). All > apps can still be activated, and it will occupy all resource of a queue > instead of only a max_am_resource_percent of a queue. -- This message was sent by Atlassian JIRA (v6.3.4#6332)