[
https://issues.apache.org/jira/browse/KNOX-2680?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Sandor Molnar updated KNOX-2680:
--------------------------------
Description:
Currently, Knox performs cluster configuration discovery via the CM API for
every descriptor (on start-up or descriptor/provider changed), even if the
descriptors reference the same cluster. This discovery can be expensive and
doing it multiple times only to get the same data yields sub-optimal
performance.
It will be better if Knox maintained a central record of the cluster
configuration details, which could be referenced for every applicable
descriptor, perhaps with some expiration of the validity of that record, such
that multiple interactions with the CM API for the same cluster can be avoided.
Not to mention, that Knox discovers all services found in the referenced
cluster even if those services are not listed in any of the descriptors.
The associated cluster configuration monitoring functionality will also need to
be modified to accommodate this new behavior.
was:
Currently, Knox performs cluster configuration discovery via the CM API for
every descriptor (on start-up or descriptor/provider changed), even if the
descriptors reference the same cluster. This discovery can be expensive and
doing it multiple times only to get the same data yields sub-optimal
performance.
It will be better if Knox maintained a central record of the cluster
configuration details, which could be referenced for every applicable
descriptor, perhaps with some expiration of the validity of that record, such
that multiple interactions with the CM API for the same cluster can be avoided.
The associated cluster configuration monitoring functionality will also need to
be modified to accommodate this new behavior.
> Centralized Source of Cluster Configuration Details for Discovery
> -----------------------------------------------------------------
>
> Key: KNOX-2680
> URL: https://issues.apache.org/jira/browse/KNOX-2680
> Project: Apache Knox
> Issue Type: Improvement
> Components: Server
> Affects Versions: 1.5.0, 1.6.0
> Reporter: Sandor Molnar
> Assignee: Sandor Molnar
> Priority: Critical
> Fix For: 2.0.0
>
>
> Currently, Knox performs cluster configuration discovery via the CM API for
> every descriptor (on start-up or descriptor/provider changed), even if the
> descriptors reference the same cluster. This discovery can be expensive and
> doing it multiple times only to get the same data yields sub-optimal
> performance.
> It will be better if Knox maintained a central record of the cluster
> configuration details, which could be referenced for every applicable
> descriptor, perhaps with some expiration of the validity of that record, such
> that multiple interactions with the CM API for the same cluster can be
> avoided. Not to mention, that Knox discovers all services found in the
> referenced cluster even if those services are not listed in any of the
> descriptors.
> The associated cluster configuration monitoring functionality will also need
> to be modified to accommodate this new behavior.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
