clintropolis commented on a change in pull request #6740: Zookeeper loss
URL: https://github.com/apache/incubator-druid/pull/6740#discussion_r253290365
##########
File path: server/src/main/java/org/apache/druid/curator/CuratorModule.java
##########
@@ -127,6 +154,29 @@ public EnsembleProvider
makeEnsembleProvider(CuratorConfig config, ExhibitorConf
return new FixedEnsembleProvider(config.getZkHosts());
}
+ RetryPolicy retryPolicy;
+ if (config.getTerminateDruidProcess()) {
+ final Function<Void, Void> exitFunction = new Function<Void, Void>()
+ {
+ @Override
+ public Void apply(Void aVoid)
+ {
+ log.error("Zookeeper can't be reached, forcefully stopping virtual
machine...");
+ System.exit(1);
Review comment:
Should this stop the lifecycle too? If this is getting hit it means that the
curator killer is _not_ getting hit and so I think this exit would be unclean.
Also, did you figure out why this needs to kill the process too? I see the
discussion here,
https://github.com/apache/incubator-druid/pull/6740#discussion_r243368737 but
it doesn't really come to a conclusion about whether it's needed.
I'm not sure it is needed. Presumably if exhibitor is down the zk processes
managed by it are dead too, and I would expect the other killer will get hit?
Or is this for a startup case where the other killer somehow doesn't get
triggered? Regardless, could you also add a comment with the explanation?
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
