[
https://issues.apache.org/jira/browse/NIFI-4043?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16042729#comment-16042729
]
Bryan Bende commented on NIFI-4043:
-----------------------------------
[~ijokarumawak] If you get a chance I would like to get your feedback on this
PR since you did the original work on AtomicDistributedMapCacheClient...
I refactored the wait/notify protocol a little bit and moved towards an atomic
replace based on previous value, rather than a revision number.
The reason being that the revision number is easy to implement with our own
DMCS because we control everything, but with an external cache like Redis,
unless it provides a revision number then you have to use another entry in the
cache to maintain it. I think the replace-if-prev-value-still-equal should
achieve the same thing, but want to see if you know of any reason it wouldn't
be a good idea. Thanks!
> Implement a DistributedMapCacheClient backed by Redis
> -----------------------------------------------------
>
> Key: NIFI-4043
> URL: https://issues.apache.org/jira/browse/NIFI-4043
> Project: Apache NiFi
> Issue Type: Improvement
> Reporter: Bryan Bende
> Assignee: Bryan Bende
> Priority: Minor
>
> We should provide other caching options besides the
> DistributedMapCacheClient/DistributedMapCacheServer.
> Redis is a popular caching solution and we could implement a
> RedisDistributedMapCacheClient that would let us use all the existing
> processors against Redis.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
