Urgent notice from Elastic Support:
Recently it was discovered that in certain configurations, Elasticsearch version 7.17.19, 8.13.0/8.13.1 can exhibit issues where one or more nodes could abruptly stop, with potential for data corruption. Included below is a description of the issue, impacted versions and potential workarounds. We truly apologize for the inconvenience. If you have any further questions or concerns please submit a ticket through the Elastic Support Hub or email to support@elastic.co
Impacted Version
Elasticsearch version 7.17.19 , versions 8.13.0/8.13.1 - with JDK 22
Cause
- The issue is being investigated in this Elasticsearch issue.
- The same issue is also linked to this JDK issue which is currently under investigation.
Workaround
- For self-managed environments: users can install and use JDK version 21.0.2 instead (see Java (JVM) Version)
- For ESS: an incident is currently opened: ESS (Public) Status - Elasticsearch Instance Disruption on Elasticsearch 8.13 (no workaround)
- For ECE/ECK: no workaround
- For further assistance, reach out to Elastic Support
Resolution
A fix is currently being reviewed which includes the downgrade to JDK version 21.0.2, see Downgrade the bundled JDK to JDK 21.0.2. The fix is planned to be released in versions 7.17.20, and 8.13.2.
Please review the attached Knowledge Base Document for more information, we will keep it up-to-date as we find more: https://support.elastic.co/knowledge/d83cf745