[
https://issues.apache.org/jira/browse/PHOENIX-7566?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Kadir Ozdemir updated PHOENIX-7566:
-----------------------------------
Description:
The HA Group Store is composed of ZK nodes from two ZK clusters, one from each
HBase cluster. As described in the [design
document|https://docs.google.com/document/d/1U2ULZ1xENDxGhAxdy5U0JYGVO5790VR5AOAJMiXHIPc/edit?usp=sharing]
for PHOENIX-7562,
* HAGroupStoreClient runs on the server side inside the HBase region servers,
listens to the HA Group Store for updates on the HA group records, and also
updates the HA group records.
* HAGroupStoreManager coordinates server-side behavior with HA Group Store and
uses HAGroupStoreClient to get updates from and make updates on HA Group Store.
It also provides an RPC endpoint for DualClusterClient to retrieve HA group
records.
This subtask is to enhance HAGroupStoreManager and HAGroupStore introduced in
PHOENIX-7493 for consistent failover.
was:
The HA Group Store is composed of ZK nodes from two ZK clusters, one from each
HBase cluster. As described in the [design
document|https://docs.google.com/document/d/1Qw1MM7jten_h2oGMWZRC_8Py4K-01S9VyPMXdjE3Gj0/edit?pli=1&tab=t.0#heading=h.5s1sz96x6dg9]
for PHOENIX-7562,
* HAGroupStoreClient runs on the server side inside the HBase region servers,
listens to the HA Group Store for updates on the HA group records, and also
updates the HA group records.
* HAGroupStoreManager coordinates server-side behavior with HA Group Store and
uses HAGroupStoreClient to get updates from and make updates on HA Group Store.
It also provides an RPC endpoint for DualClusterClient to retrieve HA group
records.
This subtask is to enhance HAGroupStoreManager and HAGroupStore introduced in
PHOENIX-7493 for consistent failover.
> Enhance HAGroupStoreManager and HAGroupStoreClient for consistent failover
> --------------------------------------------------------------------------
>
> Key: PHOENIX-7566
> URL: https://issues.apache.org/jira/browse/PHOENIX-7566
> Project: Phoenix
> Issue Type: Sub-task
> Reporter: Kadir Ozdemir
> Assignee: Ritesh
> Priority: Major
>
> The HA Group Store is composed of ZK nodes from two ZK clusters, one from
> each HBase cluster. As described in the [design
> document|https://docs.google.com/document/d/1U2ULZ1xENDxGhAxdy5U0JYGVO5790VR5AOAJMiXHIPc/edit?usp=sharing]
> for PHOENIX-7562,
> * HAGroupStoreClient runs on the server side inside the HBase region
> servers, listens to the HA Group Store for updates on the HA group records,
> and also updates the HA group records.
> * HAGroupStoreManager coordinates server-side behavior with HA Group Store
> and uses HAGroupStoreClient to get updates from and make updates on HA Group
> Store. It also provides an RPC endpoint for DualClusterClient to retrieve HA
> group records.
> This subtask is to enhance HAGroupStoreManager and HAGroupStore introduced in
> PHOENIX-7493 for consistent failover.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
