[
https://issues.apache.org/jira/browse/FLINK-8174?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16278814#comment-16278814
]
ASF GitHub Bot commented on FLINK-8174:
---------------------------------------
Github user EronWright commented on a diff in the pull request:
https://github.com/apache/flink/pull/5114#discussion_r154999162
--- Diff:
flink-mesos/src/main/scala/org/apache/flink/mesos/scheduler/LaunchCoordinator.scala
---
@@ -148,14 +149,17 @@ class LaunchCoordinator(
case Event(offers: ResourceOffers, data: GatherData) =>
val leases = offers.offers().asScala.map(
- new VMLeaseObject(_).asInstanceOf[VirtualMachineLease])
+ new Offer(_).asInstanceOf[VirtualMachineLease])
if(LOG.isInfoEnabled) {
val (cpus, mem) = leases.foldLeft((0.0,0.0)) {
(z,o) => (z._1 + o.cpuCores(), z._2 + o.memoryMB())
}
LOG.info(s"Received offer(s) of $mem MB, $cpus cpus:")
for(l <- leases) {
- LOG.info(s" ${l.getId} from ${l.hostname()} of ${l.memoryMB()}
MB, ${l.cpuCores()} cpus")
+ val reservations =
l.asInstanceOf[Offer].getResources.asScala.map(_.getRole).toSet
--- End diff --
(see above)
> Mesos RM unable to accept offers for unreserved resources
> ---------------------------------------------------------
>
> Key: FLINK-8174
> URL: https://issues.apache.org/jira/browse/FLINK-8174
> Project: Flink
> Issue Type: Bug
> Components: Mesos
> Affects Versions: 1.4.0, 1.3.3
> Reporter: Eron Wright
> Assignee: Eron Wright
> Priority: Blocker
> Fix For: 1.4.0
>
>
> Flink has suffered a regression due to FLINK-7294. Any attempt to accept a
> resource offer that is based on unreserved resources will fail, because Flink
> (as of FLINK-7294) erroneously insists that the resource come from a prior
> reservation.
> Looking at the original issue, the problem may have been misdiagnosed.
> Ideally Flink should work with both reserved and unreserved resources, but
> the latter is a more common situation that is now broken.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
