Elasticsearch version 6.8.9edit

Known issuesedit

  • Slow loggers can cause Log4j loggers to leak over time. When a new index is created, a new Log4j logger is associated with it. However, when an index is deleted, Log4j keeps an internal reference to its loggers that results in a memory leak #57228 (issue: #56171)

    This issue is fixed in Elasticsearch 6.8.10 and 7.7.1.

Also see Breaking changes in 6.8.

  • Java based - formats with 8 prefix - week based parsing and calculations are using JDK default calendar data provider which is Sunday,1. Sunday is considered first day of a week and it requires only 1 day in a week to for the first week of the year. It can be worked around by using locale which is based on ISO8601 rule (Monday,4) - for instance en-GB This issue is fixed in Elasticsearch 7.7 https://github.com/elastic/elasticsearch/pull/48209

Enhancementsedit

Authorization
  • Remove synthetic role names of API keys as they confuse users #56005
Infra/Core
  • Enable helpful null pointer exceptions #54853
Snapshot/Restore
  • Update the HDFS version used by HDFS Repo #53693

Bug fixesedit

CCR
  • Retry follow task when remote connection queue full #55314
  • Clear recent errors when auto-follow successfully #54997
  • Handle no such remote cluster exception in ccr #53415 (issue: #53225)
CRUD
  • Reduce deprecation warnings from bulk OCC #53911
Engine
  • Ensure no circular reference in translog tragic exception #55959 (issue: #55893)
Features/CAT APIs
  • Cat tasks output should respect time display settings #54536
Infra/Core
  • Avoid StackOverflowError if write circular reference exception #54147 (issue: #53589)
  • Disallow negative TimeValues #53913 (issue: #54041)
Network
  • Fix use of password protected PKCS#8 keys for SSL #55457 (issue: #8)
  • Add support for more named curves #55179 (issue: #55031)
Reindex

Upgradesedit

Engine
  • Upgrade to Lucene 7.7.3 #55897