dsmiley commented on code in PR #3643:
URL: https://github.com/apache/solr/pull/3643#discussion_r2353381432
##########
solr/core/src/java/org/apache/solr/cloud/ZkController.java:
##########
@@ -392,13 +392,19 @@ public ZkController(
// this must happen after zkStateReader has initialized the cluster props
this.baseURL = URLUtil.getBaseUrlForNodeName(this.nodeName,
urlSchemeFromClusterProp);
- // Now that zkStateReader is available, read OVERSEER_ENABLED.
+ checkForExistingEphemeralNode(); // removes our live node if present
+
+ // Now that zkStateReader is available, read OVERSEER_ENABLED & live nodes.
+ final Optional<SolrVersion> lowestNodeVersion =
zkStateReader.fetchLowestSolrVersion();
+ boolean defaultOverseerEnabled =
Review Comment:
Woah, great catch! Code review FTW. Damn; we need multi-version Solr
testing. Until then, at least manually -- which I should do using maybe Docker.
As for a solution, perhaps a Solr node observing an older version from
lowestNodeVersion at startup (here) should explicitly set the
`overseerEnabled`=false cluster property so that the choice is made durably
without flipping accidentally. This means someone doing a rolling upgrade will
have a cluster that is with an Overseer henceforth. But someone creating a
fresh cluster or who upgrades via entire restart will get distributed mode
(henceforth). I'm glad I added the cluster downgrade prevention check as well.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscr...@solr.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@solr.apache.org
For additional commands, e-mail: issues-h...@solr.apache.org
