WARN: o.e.t.ThreadPool - absolute clock went backwards…

Hi,

after a fresh installation this log entry shows up on every node once a minute.

absolute clock went backwards by [300ms/300ms] while timer thread was sleeping

I already searched the forum and the Elasticsearch documentation but I could not find any further explanation about this error.

A google search only brings up the Elasticsearch git repository, where this error is declared.

It seems to me, that this warning has something to do with the internal thread management.
But I could not find any hint what actually causes this warning.

NTP is configured and working on all servers.

The environment:
Elasticsearch 7.16.3 / 4 Nodes
Debian 11.2
VMware ESXI 6.0

Maybe someone can provide a hint why this error is showing up.
Thank you for your help!

Kind regards,
Marc

Really just what the message says: Elasticsearch read the system time, slept for a few hundred milliseconds, then read the time again and saw that the clock went backwards. NTP only typically introduces a clock discontinuity when it's unable to keep the clocks in sync any other way, so it sounds like something is wrong with your system's timekeeping.

1 Like

Thank you for your reply. I am going to debug NTP and the system clock. As soon as I find the problem (and the solution), I am going to post it here.

1 Like

Thank you for your help. The Problem was indeed caused by a misconfigured upstream ntp server.

1 Like

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