[IMPALA-3043] SimpleScheduler does not handle hosts with multiple IP addresses correctly

• Type: Bug
• Status: Closed
• Priority: Minor
• Resolution: Won't Fix
• Affects Version/s: Impala 2.0
• Component/s: Backend
• Target Version: Impala 2.9.0
• Assignee: Lars Volker
• Reporter: Lars Volker
• Created: 19/Feb/16 23:30
• Updated: 26/Jul/17 23:27
• Resolved: 26/Jul/17 23:27

Description

SimpleScheduler in the backend calls HostnameToIpAddrs() to resolve a host to a list of its IP addresses and then picks the first non-localhost one to identify the host. The underlying call to getaddrinfo() might return multiple IP addresses to the same host in a different order on subsequent calls. The relevant RFC3484 only defines a partial order. A small test program confirmed that the order is not static. As a fix we can sort the result of HostnameToIpAddrs() before picking one.