[ 
https://issues.apache.org/jira/browse/USERGRID-739?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Shawn Feldman updated USERGRID-739:
-----------------------------------
    Description: 
When allocating an index alias to one of the physical indexes a bucketing 
strategy should be used with ranges based off the number of indexes in 
elasticsearch.

For example, if there are 10 physical indexes then there would be 10 buckets 
based on a hash of the Application ID which would result in the assignment of 
an app alias to a physical index based on the range in which it falls.

#Cluster Naming
clusterName = config{usergrid.cluster_name}
keyspaceName = config{cassandra.keyspace.application}
managementName = config{elasticsearch.managment_index}
indexRoot = {clusterName}_{keyspaceName}
managementIndexName = {indexRoot}_{managementName}
managementAliasName = {indexRoot}_{managementName}_read_alias || 
{indexRoot}_{managementName}_read_alias
applicationIndexName = {indexRoot}_applications_{bucketId}
applicationAliasName = {indexRoot}_{appId}_read_alias || 
{indexRoot}_{appId}_write_alias

  was:
When allocating an index alias to one of the physical indexes a bucketing 
strategy should be used with ranges based off the number of indexes in 
elasticsearch.

For example, if there are 10 physical indexes then there would be 10 buckets 
based on a hash of the Application ID which would result in the assignment of 
an app alias to a physical index based on the range in which it falls.


> Implement Bucketing Algorithm for assignment of Application Index Aliases to 
> Physical Indexes
> ---------------------------------------------------------------------------------------------
>
>                 Key: USERGRID-739
>                 URL: https://issues.apache.org/jira/browse/USERGRID-739
>             Project: Usergrid
>          Issue Type: Story
>            Reporter: Jeffrey 
>            Assignee: Shawn Feldman
>
> When allocating an index alias to one of the physical indexes a bucketing 
> strategy should be used with ranges based off the number of indexes in 
> elasticsearch.
> For example, if there are 10 physical indexes then there would be 10 buckets 
> based on a hash of the Application ID which would result in the assignment of 
> an app alias to a physical index based on the range in which it falls.
> #Cluster Naming
> clusterName = config{usergrid.cluster_name}
> keyspaceName = config{cassandra.keyspace.application}
> managementName = config{elasticsearch.managment_index}
> indexRoot = {clusterName}_{keyspaceName}
> managementIndexName = {indexRoot}_{managementName}
> managementAliasName = {indexRoot}_{managementName}_read_alias || 
> {indexRoot}_{managementName}_read_alias
> applicationIndexName = {indexRoot}_applications_{bucketId}
> applicationAliasName = {indexRoot}_{appId}_read_alias || 
> {indexRoot}_{appId}_write_alias



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to