Closed csoulios closed 4 months ago
Pinging @elastic/es-data-management (Team:Data Management)
This one is failing because the disk health is UNKNOWN. This is almost certainly because this indicator depends on the health node (it's the first one to do so), and the health node probably just hasn't received any disk data yet since the cluster just came up.
The old cluster health API supports a wait_for_status
parameter, and a lot of rest tests have wait_for_status: green
. Maybe we ought to do that here? Either that or just support either green
or unknown
. Either way, I think it's a test problem and not a bug.
CI Link
https://gradle-enterprise.elastic.co/s/sh2jvmtofx3fq
Repro line
Does it reproduce?
Yes, reproduced locally
Applicable branches
8.4
Failure history
https://gradle-enterprise.elastic.co/scans/tests?search.startTimeMax=1663753945428&search.startTimeMin=1661979600000&search.timeZoneId=Europe/Athens&tests.container=org.elasticsearch.backwards.MixedClusterClientYamlTestSuiteIT&tests.test=test%20%7Bp0%3Dhealth/10_basic/cluster%20health%20basic%20test%7D#
Failure excerpt