You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@servicecomb.apache.org by "humingcheng (Jira)" <ji...@apache.org> on 2020/02/18 06:24:00 UTC
[jira] [Resolved] (SCB-1742) Optimize Aggregator and etcd's service
discovery cache mechanism
[ https://issues.apache.org/jira/browse/SCB-1742?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
humingcheng resolved SCB-1742.
------------------------------
Resolution: Fixed
New mechanism: When Adaptor is not credible, no service discovery consistency guarantee is made.
> Optimize Aggregator and etcd's service discovery cache mechanism
> ----------------------------------------------------------------
>
> Key: SCB-1742
> URL: https://issues.apache.org/jira/browse/SCB-1742
> Project: Apache ServiceComb
> Issue Type: Improvement
> Components: Service-Center
> Reporter: humingcheng
> Assignee: humingcheng
> Priority: Minor
> Time Spent: 20m
> Remaining Estimate: 0h
>
> Scenario:
> If the request revision is not empty and different from the revision of service-center's instance cache, service-center will call the backend directly to get instance data to response. Thus, the client can't get any instance data if service-center's backend is unavailable, even service-center's cache is available.
> Target:
> Optimize the second level cache of instance, in package server.service.cache.
> We only use the cache and not call the backend directly when using Aggregatory adaptor, as Aggregator does not guarantee accuracy when querying the backend.
> For other adaptors, if the backend is unavailable, the cache content is returned instead of a backend error.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)