[jira] [Created] (TEZ-3238) Shuffle service name should be configureable and should not be hardcoded to ‘mapreduce_shuffle’
Rob Leidle created TEZ-3238: --- Summary: Shuffle service name should be configureable and should not be hardcoded to ‘mapreduce_shuffle’ Key: TEZ-3238 URL: https://issues.apache.org/jira/browse/TEZ-3238 Project: Apache Tez Issue Type: Improvement Affects Versions: 0.8.2 Reporter: Rob Leidle In YARN there are no guarantees that a shuffle service with a specific name exists. The setting ' yarn.nodemanager.aux-services’ can be filled with a list of names that can be defined by the Hadoop cluster administrator. It is merely by convention that many clusters have their MapReduce shuffle service named ‘mapreduce_shuffle’. Tez is hard-coded to use a shuffle service named ‘mapreduce_shuffle’: https://github.com/apache/tez/blob/TEZ-8/tez-api/src/main/java/org/apache/tez/dag/api/TezConstants.java#L75. This name should be configureable. Also, this is a hidden dependency of Tez, we should add that the shuffle service is a dependency to the Tez documentation. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
One to one edges and local fetch
I was under the assumption that we optimized 1-1 edge scheduling to reuse containers or run as much as possible on same node and do local fetch. But does not seem to be the case. Is there a jira for this already? Could not find any. Regards, Rohini
[jira] [Created] (TEZ-3237) Corrupted shuffle transfers to disk are not detected during transfer
Jason Lowe created TEZ-3237: --- Summary: Corrupted shuffle transfers to disk are not detected during transfer Key: TEZ-3237 URL: https://issues.apache.org/jira/browse/TEZ-3237 Project: Apache Tez Issue Type: Bug Affects Versions: 0.7.0 Reporter: Jason Lowe When a shuffle transfer is larger than the single transfer limit it gets written straight to disk during the transfer. Unfortunately there are no checksum validations performed during that transfer, so if the data is corrupted at the source or during transmit it goes undetected. Only later when the task tries to consume the transferred data is the error detected, but at that point it's too late to blame the source task for the error. -- This message was sent by Atlassian JIRA (v6.3.4#6332)