[jira] [Resolved] (SLING-4371) Create 'cluster clocks in sync' health check

["Stefan Egli (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/SLING-4371?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Stefan Egli resolved SLING-4371.
--------------------------------
    Resolution: Won't Fix

This is not easily possible to implement since it bases on a successful voting. But if an instance' clock is further behind than the configured heartbeat timeout, it would be thrown out of the view permanently - in which case the health-check would not detect this.

Hence a proper health-check for clock-in-sync would require additional communication between the instances - which would resemble more and more NTP - which is not what we should be doing.

Hence marking Won't Fix

> Create 'cluster clocks in sync' health check
> --------------------------------------------
>
>                 Key: SLING-4371
>                 URL: https://issues.apache.org/jira/browse/SLING-4371
>             Project: Sling
>          Issue Type: Sub-task
>          Components: Extensions, Health Check
>            Reporter: Stefan Egli
>             Fix For: Discovery Impl 1.0.14
>
>
> Write a health-check (probably in the discovery.impl bundle itself?) that checks the newly added 'votedAt' timestamp of each cluster instance and makes sure they don't differ too much. If they do, fail the check. This would be a very simple check to make sysadmins aware that the clocks are not in sync - or for that matter, that there is a problem in the repository, ie that there are too large write-delays between instances hooked to the same repository.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)