shounakmk219 commented on code in PR #16571:
URL: https://github.com/apache/pinot/pull/16571#discussion_r2269538877
##########
pinot-controller/src/main/java/org/apache/pinot/controller/helix/core/minion/PinotTaskManager.java:
##########
@@ -238,6 +238,15 @@ public Map<String, String> createTask(String taskType,
String tableName, @Nullab
LOGGER.warn("No ad-hoc task generated for task type: {}", taskType);
continue;
}
+ int maxNumberOfSubTasks = taskGenerator.getMaxNumSubTasks();
Review Comment:
Agree with Manish on informing the user in some way that the system is
throttling the task generation. But as the default limit is `Integer.MAX_VALUE`
I am not that concerned though.
##########
pinot-controller/src/main/java/org/apache/pinot/controller/helix/core/minion/PinotTaskManager.java:
##########
@@ -739,6 +748,21 @@ protected TaskSchedulingInfo
scheduleTask(PinotTaskGenerator taskGenerator, List
List<PinotTaskConfig> presentTaskConfig =
minionInstanceTagToTaskConfigs.computeIfAbsent(minionInstanceTag,
k -> new ArrayList<>());
taskGenerator.generateTasks(List.of(tableConfig), presentTaskConfig);
+ int maxNumberOfSubTasks = taskGenerator.getMaxNumSubTasks();
+ // choose first maxNumberOfSubTasks tasks to schedule from
presentTaskConfig
+ if (presentTaskConfig.size() > maxNumberOfSubTasks) {
+ LOGGER.warn("Number of tasks generated for task type: {} for table:
{} is {}, which is greater than the "
+ + "maximum number of tasks to schedule: {}. Only the first {}
tasks will be scheduled. This is controlled"
+ + " by the cluster config maxAllowedSubTasks which is set
based on controller's performance",
+ taskType, tableName, presentTaskConfig.size(),
maxNumberOfSubTasks, maxNumberOfSubTasks);
+ presentTaskConfig = new ArrayList<>(presentTaskConfig.subList(0,
maxNumberOfSubTasks));
Review Comment:
We can go over `presentTaskConfig` and replace
`MinionConstants.TABLE_MAX_NUM_TASKS_KEY` with the `maxNumberOfSubTasks` value
right?
--
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]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
