You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Andrey Zagrebin (Jira)" <ji...@apache.org> on 2020/02/06 16:17:00 UTC
[jira] [Commented] (FLINK-15942) Improve logging of infinite
resource profile
[ https://issues.apache.org/jira/browse/FLINK-15942?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17031713#comment-17031713 ]
Andrey Zagrebin commented on FLINK-15942:
-----------------------------------------
As discussed offline, one way to model infinity for resource is to use its Java type max value, like MemorySize#MAX_VALUE
What I am not sure of, is that this max value is still a real value which can suddenly pop up as a valid result of some other computations.
It makes sense to fail any attempts to do anything with the infinity except what is needed. Now, from what I see, only the divide operation needs to produce infinity
so we could enable it only there. For MemorySize, some negative number could be used to denote the infinity or maybe even a boolean private field.
All operations over the infinity should now fail with it, except the division.
Double sized resources also have a special Double#POSITIVE_INFINITY
> Improve logging of infinite resource profile
> --------------------------------------------
>
> Key: FLINK-15942
> URL: https://issues.apache.org/jira/browse/FLINK-15942
> Project: Flink
> Issue Type: Improvement
> Components: Runtime / Configuration, Runtime / Task
> Affects Versions: 1.10.0
> Reporter: Andrey Zagrebin
> Priority: Major
>
> After we set task memory and CPU to infinity in FLINK-15763, it spoiled the logs:
> {code:java}
> 00:23:49,442 INFO org.apache.flink.runtime.taskexecutor.slot.TaskSlotTableImpl - Free slot TaskSlot(index:0, state:ACTIVE, resource profile: ResourceProfile{cpuCores=44942328371557892500000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000.0000000000000000, taskHeapMemory=2097152.000tb (2305843009213693951 bytes), taskOffHeapMemory=2097152.000tb (2305843009213693951 bytes), managedMemory=20.000mb (20971520 bytes), networkMemory=16.000mb (16777216 bytes)}, allocationId: 349dacfbf1ac4d0b44a2d11e1976d264, jobId: 689a0cf24b40f16b6f45157f78754c46).
> {code}
> We should treat the infinity as a special case and print it accordingly
--
This message was sent by Atlassian Jira
(v8.3.4#803005)