Index.translog.durability async

Elastic Stack Elasticsearch
1

In case an index.translog.durability is set to async and process is killed, but OS is up and running.
Will there be any data loss?
From what I understand, only in case of OS reboot/ power loss/ hardware issues, there will be a data loss.
Am I right?

Thanks

2

If you change the translog durability setting to async yes you can have data loss in case of hardware issues but I'd guess that it can also happen in case of hard kill of the jvm..
Why do you want to change it? Speed?

3

Research purposes only, thanks

4

By "Kill of the jvm" you mean to kill hard the ES process?

5

Yeah. Something like a kill -9 (SIGKILL). Not sure how the JVM behaves in that case.

6

@dadoonet - Another question, if I may:

Can this configuration (async), cause translog corruption? And ES will have a problem recovering?
Or can only cause loss of data that was not fsynced?

7

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.