[
https://issues.apache.org/jira/browse/KAFKA-2587?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14943764#comment-14943764
]
Parth Brahmbhatt commented on KAFKA-2587:
-----------------------------------------
I am fine with changing the code so it always use the notification handler to
update the cache and we only verify with waitUntil. I will update the patch and
post a PR in next hour.
I mentioned "Simultaneously Consistent Cross-Client Views" in reply to
[~ijuma]'s concerns around this happening in production, where I hope most
users will be running a set of servers. I am still not sure if we will be
affected by that in production given we update acls first, and then insert the
seq node to indicate acls have changed. The notification handler only gets
fired as part of a watcher or as part of startup and in both cases I am
"hoping" that if a client sees the insertion of seq node(second operation) and
can read it , they should also see the first operation (updates to acls).
However zookeeper documentation does not really describe this behavior clearly.
I am going to try and confirm with zookeeper committers/community but if you
already know that the above condition holds (if order of operations was
update-1, insert-2 and a client can read insert-2 entry, they must also be able
to read update-1 even if those operations were applied to different paths)
please let me know.
> Transient test failure: `SimpleAclAuthorizerTest`
> -------------------------------------------------
>
> Key: KAFKA-2587
> URL: https://issues.apache.org/jira/browse/KAFKA-2587
> Project: Kafka
> Issue Type: Sub-task
> Reporter: Ismael Juma
> Assignee: Parth Brahmbhatt
> Fix For: 0.9.0.0
>
>
> I've seen `SimpleAclAuthorizerTest ` fail a couple of times since its recent
> introduction. Here's one such build:
> https://builds.apache.org/job/kafka-trunk-git-pr/576/console
> [~parth.brahmbhatt], can you please take a look and see if it's an easy fix?
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
