[MAPREDUCE-5002] AM could potentially allocate a reduce container to a map attempt

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 2.0.3-alpha, 0.23.7, 2.7.0
• Fix Version/s: 2.8.0, 3.0.0-alpha1
• Component/s: mr-am
• Hadoop Flags: Reviewed
• Assignee: Chang Li
• Reporter: Jason Darrell Lowe
• Created: 13/Feb/13 01:59
• Updated: 30/Aug/16 01:20
• Resolved: 17/Sep/15 18:19

Description

As discussed in MAPREDUCE-4982, after MAPREDUCE-4893 it is theoretically possible for the AM to accidentally assign a reducer container to a map attempt if the AM doesn't find a reduce attempt actively looking for the container (e.g.: the RM accidentally allocated too many reducer containers).

Issue Links

• is duplicated by: MAPREDUCE-5115 MR app master may try to assign a reduce priority container to a map
• relates to: MAPREDUCE-4982 AM hung with one pending map task
• relates to: MAPREDUCE-4893 MR AppMaster can do sub-optimal assignment of containers to map tasks leading to poor node locality