Details
-
Bug
-
Status: Open
-
Major
-
Resolution: Unresolved
-
3.3.3
-
None
-
None
Description
When a nodemanager is reconnected, resourcemanager marks the acquired containers on that node as LOST and which leads to job failure.
2023-04-10 02:57:16,412 INFO org.apache.hadoop.yarn.server.resourcemanager.ResourceTrackerService (IPC Server handler 41 on 8025): Reconnect from the node at: node1
2023-04-10 02:57:16,412 INFO org.apache.hadoop.yarn.server.resourcemanager.ResourceTrackerService (IPC Server handler 41 on 8025): NodeManager from node node1(cmPort: 8041 httpPort: 8042) registered with capability: <memory:122880, vCores:16>, assigned nodeId node1:8041, node labels { CORE }
2023-04-10 02:57:16,413 INFO org.apache.hadoop.yarn.server.resourcemanager.rmcontainer.RMContainerImpl (ResourceManager Event Processor): container_e15_1677844874019_238016_01_000002 Container Transitioned from ACQUIRED to KILLED